Stakeholder Involvement in Value Co-Destruction: a Case Of

Home , Cyclone Martin (1999), Fleur East

ASSOCIATION FOR CONSUMER RESEARCH

Labovitz School of Business & Economics, University of Minnesota Duluth, 11 E. Superior Street, Suite 210, Duluth, MN 55802

Stakeholder Involvement in Value Co-Destruction: a Case of Promoting Condoms in India Samanthika Gallage, Nottingham University Business School Sameer Deshpande, Griffith University

This paper explores the reasons for low condom usage in India by elucidating the stakeholder involvement in the value creation and delivery process. Data was collected from non-profit organisations in India and findings suggested that value co-destruction resulted from cultural incongruencies amplified by stakeholders

[to cite]: Samanthika Gallage and Sameer Deshpande (2020) ,"Stakeholder Involvement in Value Co-Destruction: a Case of Promoting Condoms in India", in NA - Advances in Consumer Research Volume 48, eds. Jennifer Argo, Tina M. Lowrey, and Hope Jensen Schau, Duluth, MN : Association for Consumer Research, Pages: 758-757.

[url]: http://www.acrwebsite.org/volumes/2661294/volumes/v48/NA-48

[copyright notice]: This work is copyrighted by The Association for Consumer Research. For permission to copy or use this work in whole or in part, please contact the Copyright Clearance Center at http://www.copyright.com/. Sociomaterial Performativity of the Surveillant Assemblage on Healthcare Platforms Handan Vicdan, Emlyon Business School, France

EXTENDED ABSTRACT technologies (Orlikowski and Scott, 2008; Kjellberg and Helgesson, This conceptual paper proposes a sociomaterial lens for under- 2006) that enact performances (Barad, 2003) at micro/meso/macro standing how surveillance functions and produces outcomes in plat- levels, and how these outcomes reshape existing practices and transform organizations, and draws attention to this performative process form power dynamics among market actors (Cecez-Kecmanovic et of knowledge generation in healthcare. Platform organizations en- al., 2014; Orlikowski and Scott, 2015). able distant and asynchronous monitoring of people as assemblages The emergence of alternative healthcare platforms (e.g., Pa- of data (Zwick and Denegri-Knott, 2009). With this surveillant as- tientsLikeMe, trackmystack, myhealthteams, 23andme, raremark) semblage, purposes and hierarchies of begin to shift from one-sided for medical knowledge generation is a good example to study the power disparities (Foucault, 1979) to more complex surveillance net- different performances and their related sociomaterial outcomes works (Haggerty and Ericson, 2000) because those who were once at micro/meso/macro levels. Surveillant assemblage on these plat- under surveillance, and thereby disciplined, now enact surveillance forms operates via different levels and involvement of multiple ac- themselves. Questions emerge concerning how platform organiza- tors, human (e.g., patients, physicians, caregivers, platform operations constitute a surveillant assemblage that challenge the existing tives, pharmaceuticals, researchers, and state) and non-human (e.g., practices and power relations for medical knowledge constitution. genetic reports, symptom tracking tools, mood maps, doctor visit Conventional theories of surveillance do not suffice to describe the sheets, disease rating scales, disease outcome measures). At the mi- complexity of the contours and effects of surveillance phenomena cro and meso levels, patient practices are configured by surveillance observed in the datafication of everyday life, as the effects of and technologies and exist through material artefacts (e.g., doctor visit responses to surveillance are much more complex than resistance sheets produced on platforms and used in clinical settings, symptom and compliance (Brivot and Gendron, 2011). Hence a need to move trackers to track patient well-being), and their outcomes yield ma- beyond a dichotomous and reactive view of surveillance (Visser et terial (e.g., application of new treatments, patient self-tracked data al., 2017), and focus on how patients, the surveilled, can enact sur- influence his and others’ bodily experiences), and social (tracking veillance is warranted; a perspective often missing in surveillance tools organize relations among patients and new forms of interaction literature (Visser et al., 2017). between patients and doctors in clinical settings) effects. Patients use With the emergence of participatory medicine, Web 2.0 tech- tracking tools to gain a deeper sense of self and others (Moore and nologies transform the ways patients manage their health by tracking Robinson, 2015) - a potential of surveillance and its reflexive nature health data in real-time and enabling big data generation. Hierarchies (Iedema and Rhodes, 2010). Self-tracking is the result of a dynamic of knowledge production become leveled with increased access to sociomaterial arrangement (Bode and Christensen, 2016). Digital medical knowledge (Rier and Indyk, 2006). Questions emerge con- tracking tools predict patient diagnostic/prognostic data as patients cerning how platform organizations constitute a surveillant assem- quantify and visualize their data, and shape their management of blage of humans and technology, and its outcomes, which challenge care and relations in material and social terms (Callon, 2008; Or- the existing practices and power relations for knowledge production. likowski and Scott, 2008). Outcome measures shared with other pa- Sociomateriality is a useful lens for exploring the materiality and tients and healthcare stakeholders make surveillance more relational performativity (Jones, 2014) of surveillant assemblages and their than ever. They improve communication on patient care with diverse outcomes as enacted in the practices and relations of human and stakeholders in clinical settings, influence patient embodied practices nonhuman actors (Callon, 2008; Orlikowski and Scott, 2015). The and organizing relations to others accordingly, prevent hospitaliza- inextricably related material, digital and social outcomes of surveil- tion or wrong treatments, hence producing both social and material lance in knowledge constitution (Barad, 2003; Orlikowski, 2007; outcomes in and beyond the platform. Institutional actors (pharma- Orlikowski and Scott, 2008) yield both material and social effects, ceuticals) also track patient data, recruit patients for clinical trials, which draws attention to the performativity of surveillance technolo- track drug effectiveness and side effects, and partner with patients for gies, a key concept related to sociomateriality (Jones, 2014). improving self-tracking tools. At the macro level, patient practices Sociomaterial lens in information systems and organizational such as directly reporting drug side effects to government agencies research emphasizes the relationality and the constitutive entangle- using surveillance technologies further shape state pharmacovigi- ment of humans and technology (Barad, 2003; Orlikowski, 2007, lance of pharmaceuticals, and patient generated medical research via 2010; Orlikowski and Scott, 2008, 2015) that enact performances surveillance technologies inform scientific medical research. Hence, via their practices (Barad, 2003), which is rarely studied (Gond et al., multiple performativities of the surveillant assemblage demonstrate 2015) yet is critical for studying the constitution of knowledge as a the potential for shaping the practices of its actors, organizations and sociomaterial practice (Barad, 2003; Orlikowski, 2007; Orlikowski their markets (Callon, 2008; Orlikowski and Scott, 2008). and Scott, 2008). With advancement in digital technologies, wear- Sociomaterial performativity of a surveillant assemblage also able devices, and tracking tools, scholars may study the performativ- draws attention to the performative struggles (Callon, 2007) or coun- ity of different sociomaterial arrangements and their consequences terperformativities (Mackenzie, 2007) in the era of datafication. Mi- for a more distinct sociomaterial theorizing and understanding the cro/meso counterperformativities include the exclusion of illiterate change of power dynamics in these arrangements (Cecez-Kecma- patients in the surveillant assemblage. Platforms may also create a novic, 2014). Furthermore, studies in marketing and consumer re- “technology of the self” via self-tracking, which may lead to ma- search focus on performativity by exploring how marketing tools nipulation of data by users. Data (il)literacy is also an important per- and models shape markets (Kjellberg and Helgesson, 2006; Mason formative struggle, as it may obstruct decisions concerning patient et al., 2015) as well as how consumer-enacted performativities (Mar- care and scientific knowledge production, as well as power dynam- tin and Schouten, 2014; Scaroboto, 2015) shape markets. However, ics. Future research should work on to explore the impact of these less attention is given to the constitutive entanglement of humans and counter-performativities in alternative markets.

Advances in Consumer Research 758 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 759 REFERENCES Jones, Matthew (2014), “A Matter of Life and Death: Exploring Barad, Karen (2003), “Posthumanist Performativity: Toward an Conceptualizations of Sociomateriality in the Context of Understanding of How Matter Comes to Matter,” Journal of Critical Care,” MIS Quarterly, 38 (3), 895-925. Women in Culture and Society, 28(3), 801-831. Kjellberg, Hans and Claes-Fredrik Helgesson (2006), “Multiple Barad, Karen (2007), Meeting the Universe Halfway: Quantum Versions of Markets: Multiplicity and Performativity in Physics and the Entanglement of Matter and Meaning. Market Practice,” Industrial Marketing Management, 35 (7), Durham, NC: Duke University Press. 839–55. Bode, Matthias and Dorthe B. Kristensen (2016), Digital MacKenzie, Donald (2007), “Is Economics Performative? Option Doppelgander Within: A Study on Self-tracking and the Theory and the Construction of Derivatives Markets,” In Quantified Self-movement,” In Canniford, Robin and Domen Do Economists Make Markets? On the Performativity of Bajde (eds.) Assembling; Consumption: Researching Actors, Economics, (pp. 54–86). Princeton, NJ: Princeton University Networks and Markets. (pp. 119-135). New York, NY: Press. Routledge. Mason, Katy, Hans Kjellberg and Johan Hagberg (2015), Boyd, Danah and Kate Crawford (2012), “Critical Questions for “Exploring the Performativity of Big Data. Povocations for a Cultural, Technological, and Marketing: Theories, Practices, and Devices,” Journal of Marketing Scholarly Phenomenon,” Information, Communication & Management, 31 (1–2), 1–15. Society 15(5), 662–679. Moore, Phoebe and Andrew Robinson (2015), “The Quantified Brivot, Marion and Yves Gendron (2011), “Beyond Panopticism: Self: What Counts in the Neoliberal Workplace,” New Media On the Ramifications of Surveillance in a Contemporary & Society, 18(11), 2774–2792. Professional Setting,” Accounting, Organizations and Society, Orlikowski, Wanda J. (2007), “Sociomaterial Practices: Exploring 36, 135–155. Technology at Work,” Organization Studies, 28(09), Callon, Michel (2008), “Economic Markets and the Rise of 1435–1448. Interactive Agencements: From Prosthetic Agencies to Orlikowski, Wanda J. and Susan V. Scott (2008), “Sociomateriality: Habilitated Agencies,” In Pinch T and Swedberg R (eds) Challenging the Separation of Technology, Work and Living in a Material World: Economic Sociology Meets Organization,” The Academy of Management Annals, 2(1), Science and Technology Studies. Cambridge: MIT Press, pp. 433−474. 29–56. Orlikowski, Wanda J. and Susan V. Scott (2015), “The Algorithm Cecez-Kecmanovic, Dubravka, Robert D. Galliers, Ola and the Crowd. Considering the Materiality of Service Henfridsson, Sue Newell, and Richard Vidgen. (2014), “The Innovation,” MIS Quarterly, 39(1), 201−216. Sociomateriality of Information Systems: Current Status, Rier, David A. and Debbie Indyk (2006), “The Rationale of Future Directions,” MIS Quarterly, 38 (3), 809–830. Interorganizational Linkages to Connect Multiple Sites of Foucault, Michel (1979), Discipline and punish: The birth of the Expertise, Knowledge Production, and Knowledge Transfer,” prison (A. Sheridan Trans.) New York: Vintage Books. Social Work in Health Care, 42(3−4), 8−27. Gond, Jean Pascal, Laure Cabantous, Nancy Harding and Mark Scaraboto, Daiane (2015), “Selling, Sharing, and Everything in Learmonth (2015), “What Do We Mean by Performativity in Between: The Hybrid Economies of Collaborative Networks,” Organization and Management Studies? The Uses and Abuses Journal of Consumer Research, 42, 152−176. of Performativity,” International Journal of Management Visser, Laura M., Inge Bleijenbergh, Yvonne Benschop and Allard Reviews, 18(4), 440-463. van Riel (2017), “Prying Eyes: A Dramaturgical Approach to Haggerty, Kevin D. and Richard V. Ericson (2000), “The Professional Surveillance,” Journal of Management Studies. Surveillant Assemblage,” British Journal of Sociology, 51(4), doi: 10.1111/joms.12283 605−622. Zwick, Detlew and Janice Denegri-Knott (2009), “Manufacturing Iedema, Rick and Carl Rhodes (2010), “The Undecided Space of Customers: The Database as New Means of Production,” Ethics in Organizational Surveillance,” Organization Studies, Journal of Consumer Culture, 9(2), 221−247. 31(2), 199−217. Your Need Doesn’t Appeal to Me: How Social Class Shapes Charitable Giving Across Causes Yan Vieites, FGV-EBAPE, Brazil Rafael Goldszmidt, FGV-EBAPE, Brazil Eduardo B. Andrade, FGV-EBAPE, Brazil

EXTENDED ABSTRACT ings from study 1 are indeed explained by a sensitivity gap, then Two recent tragedies received worldwide attention: the Cyclone presenting the causes in a way that makes the degree of necessity Idai and the Notre-Dame Cathedral fire. While the former ranks as emotionally/viscerally salient to all participants would make both one of the deadliest tropical cyclones on record, the latter, though higher- and lower-class consumers more likely to donate to the most tragic for its cultural significance, did not affect people’s access to pressing cause. That was indeed the case. When assigned to the neu- pressing needs. Yet, members of the elite donated approximately ten tral condition, higher-class participants donated more to culture than times more to the rebuilding of the Notre-Dame Cathedral than to food, whereas the opposite held true for their lower-class counter- the victims of the cyclone (Martin 2019; Walsh 2019). Although cel- parts. However, when assigned to the visceral condition, both groups ebrated by some, the enormous amounts donated to the Notre-Dame became more charitable to the food-related cause. Cathedral have also triggered a massive public backlash (Sullivan Study 3 replicated the procedure of study 2 in a hypothetical 2019). At the center of the criticism was the question: which social scenario and asked participants to indicate not only the social cause causes are worthy of philanthropy for the rich? In this research, we they would contribute more money, but also the reason why they shall demonstrate how a basic psychological process (i.e., sensitivity chose to donate more to one cause or the other (degree of need, af- to need) may help explain this intriguing question. finity, or other motive). Beyond replicating findings from study 2, Individual experiences have been shown to largely shape dona- results also showed that the effect of the viscerality manipulation on tion preferences (Radley and Kennedy 1995). Small and Simonsohn patterns of giving was explained by the increase in the sensitivity to (2008), for instance, showed that past experiences with a friend’s need among the wealthy in the visceral (vs. neutral) condition. misfortune increases sympathy towards other victims from the same Study 4 relies on a distinct approach to examine the mechanism. plight. However, since social class often shapes personal experiences Participants were offered the opportunity to donate to one of two in a relatively homogeneous way (Kraus et al. 2012), it may well causes in a hypothetical scenario. For half of the participants, the mold people’s relative sensitivity to particular sets of misfortunes tradeoff hinged upon shelter and sports, whereas for the other half and thereby influence their prosocial choices. the options were safety and sports. A pretest showed that lower-class Given that lower-class individuals are embedded in contexts of consumers experience much higher scarcity in the access to shelter generalized scarcity (Piff et al. 2012), we reasoned that they would than their higher-class counterparts, but both groups share similar present a greater sensitivity to the degree of need of social causes scarcity experiences in terms of safety. Since our rationale predicts relative to their higher-class counterparts. Implicit in this reasoning that higher scarcity experiences prompt a higher sensitivity to need, is the idea that going from scarcity to abundance changes not only we reasoned that lower-class participants would donate more to the the amount of resources one has to give, but also the dimensions most pressing cause than their higher-class peers when the pair of consumers are more sensitive to during the donation process. Among available causes were shelter and sports but not when the pair of the poor, scarcity-based experiences should increase the donor’s sen- causes were safety and sports. This prediction was fully supported. sitivity to need. As a result, they should be more inclined to donate Our final study adopts a framing strategy to provide further evi- to more pressing causes (e.g., helping the homeless find shelter) than dence for the proposed mechanism. Participants were offered the op- to relatively less pressing ones (e.g., promoting cultural activities). portunity to donate to a shelter-related cause and/or a sports-related Among the wealthy, abundance-based experiences should reduce the cause in a hypothetical scenario. Whereas for half of the participants, donor’s sensitivity to need and enhance the weight of other factors the shelter-related appeal emphasized the benefits inherently con- (e.g., social signaling; Sargeant and Woodliffe 2007). We test this nected to having access to adequate shelter (e.g., well-being), for the possibility across 5 field studies with residents from extremely poor other half the appeal centered on the idea that reducing homelessness and wealthy areas of Brazil. might help promote safety in the city. As expected, emphasizing the In study 1, participants received five R$2.00 bills in return for safety-related benefits of helping homeless people increased sensi- their participation. Upon completion of a filler questionnaire, respon- tivity to need, and thereby donation amounts, among the higher-class dents learned that research assistants were collecting resources for more than among the lower-class. two campaigns, one to help the homeless find shelter (a pretested This research offers two main contributions. First, we move be- pressing need) and the other to help promote cultural activities (a yond the question of who behaves more prosocially to assess when pretested non-pressing need). Participants were then given two enve- and why members of contrasting socioeconomic backgrounds act the lopes containing a pictorial depiction of each cause and instructed to most on behalf of others. Second, we extend the finding that personal deposit the money they were willing to donate (including any) in the experiences shape sympathy towards specific causes by showing designated envelope. As expected, higher-class consumers donated how it can reflect a group-based phenomenon. less than their lower-class counterparts to a cause concerned with the homeless but donated more when the cause concerned cultural REFERENCES activities. Kraus, Michael W., Paul K. Piff, Rodolfo Mendoza-Denton, Study 2 uses another combination of causes (food and culture) Michelle L. Rheinschmidt, and Dacher Keltner (2012), and provides initial process evidence. This study relies on the same “Social Class, Solipsism, and Contextualism: How the Rich procedure as study 1 except that half of the participants received an Are Different from the Poor,” Psychological Review, 119(3), envelope containing a neutral pictorial appeal, whereas the other half 546–72. received an envelope containing a visceral graphic appeal. If find-

Martin, Iain (2019), “French Billionaires Pledge $670 Million to Small, Deborah A. and Uri Simonsohn (2008), “Friends of Victims: Restore Notre Dame,” Forbes, https://www.forbes.com/sites/ Personal Experience and Prosocial Behavior,” Journal of iainmartin/2019/04/16/french-billionaires-pledge-340-million- Consumer Research, 35(3), 532–42. to-restore-notre-dame/#1f769b156087. Sullivan, Paul (2019), “Notre-Dame Donation Backlash Raises Piff, Paul K., Daniel M. Stancato, Andres G. Martinez, Michael Debate: What’s Worthy of Philanthropy?,” The New York W. Kraus, and Dacher Keltner (2012), “Class, Chaos, and Times, https://www.nytimes.com/2019/04/26/your-money/ the Construction of Community,” Journal of Personality and notre-dame-donation-backlash-philanthropy.html. Social Psychology, 103(6), 949–62. Walsh, Declan (2019), “Mozambique Cyclone Wiped Out the Radley, Alan and Marie Kennedy (1995), “Charitable Giving Nation’s Food. Now Aid Is in Short Supply,” The New York by Individuals: A Study of Attitudes and Practice,” Human Times, https://www.nytimes.com/2019/04/20/world/africa/ Relations, 48(6), 685–709. mozambique-cyclone.html. Sargeant, Adrian and Lucy Woodliffe (2007), “Gift Giving: An Interdisciplinary Review,” International Journal of Nonprofit and Voluntary Sector Marketing. Social Media as an Empowering Tool for Mothers: Evidence from India and the United States Akshaya Vijayalakshmi, Indian Institute of Management Ahmedabad, India Meng-Hsien (Jenny) Lin, California State University Monterey Bay, USA

EXTENDED ABSTRACT expenditure was captured using the question- “Please consider the A mother’s autonomous status has a substantial influence on following educational expenses: tuition, private tutoring, books, and a child’s educational attainment (Korupp, Ganzeboom and Lippe other school materials, test prep services, transportation to school” 2002), survival (Fantahun et al. 2007), and health condition (Senara- on a scale with 1 = < $250 to 9 = $2000+ with increments of $250. th and Gunawardena 2009). However, scant research has been done An individual’s social needs, recognition needs, and cognitive needs to investigate the impact of mothers on their children’s educational were also captured to understand the motivation of seeking social spending. Research in economics and psychology addressing this media (Leung 2009) and children’s grades in school. Out of the 182 topic suggests that a mother’s autonomous status is a critical determi- participants, 98% of the mothers owned a smartphone, and all par- nant of investment made for their children’s future (Mencher 1988). ticipants had access to some form of social media. Ninety percent of In this paper, we (a) consider the impact of a mother’s level of em- the participants used social media regularly, while others used social powerment on educational issues, (b) explain how access and use of media occasionally. social media influences a mother’s empowerment and thus children’s Our CFA model had a good fit (χ2 (40) = 66.66, p < .05; CFI = outcomes (Figure 1) and (c) test our hypotheses with mothers in both .94; RMSEA = .06, NFI = .86). The results suggest that the mother’s India and the United States. active social media use had a significant effect on the latent variable, intrapersonal empowerment (.26; p < .05), but not interactional Study 1- Mother’s Empowerment In India empowerment, as we had expected. Our results show that passive Educational spending was measured by summing up the expen- social media use had a significant and positive effect on the latent ditures on education (₹) for the oldest child between the ages of 11 variable, interactional empowerment (.17; p < .05). Increased intra- and 17. The expenses include tuition, school fees, books, transporta- personal empowerment reduced their child’s school grades (.73; p tion to school, other school materials, and test preparatory services. < .05; reverse-scored) while interactional empowerment increased According to Zimmerman (1995), to fully understand empowerment, educational expenditures (1.81; p < .05). We also find that recogni- both forms of empowerment, intrapersonal empowerment, and in- tion (.28; p < .05) and social needs (.36; p < .05) significantly impact teractional components, should be considered. We used a developed active social media use while cognitive needs (.36; p < .05) leads to scale to capture this (Liu 2016). We recorded mother’s use of social passive media use. media on Yes, Sometimes or No. Other relevant background control variables were captured, including the self (wife) and husband’s in- Discussion come, the number of children, work status of the mother, and wife By having a mere presence on social media (via passive use) and husband’s education level. helps women develop interpersonal relations and propel collective The average age of the 301 respondents was 38.5 years (SD = action (related to interactional empowerment). These outcomes posi- 6.23). Majority of the children (78.7%) were enrolled in a private tively influence children’s future since with increase in interactional school. Forty-two percent of the surveyed women stated that What- empowerment, more expenditures are spent related to children’s sApp was their primary social media option. About 70% of the What- education. Also, it is worth noting that cognitive needs drive passive sApp users are a part of their child’s school group or other educa- use of social media, possibly relying on social media as a critical tional related online communities, and 34% of these women reported source of information. However, mothers relying on active use of so- either “always” or “mostly” use WhatsApp for this purpose. cial media without the relational bonds (created through interactional A confirmatory factor analysis (CFA) was conducted with a empowerment), may negatively result in (a) reduced expenditures on bootstrap resampling procedure with 2000 samples and 95% bias- children’s education in India, and (b) their children performing more corrected confidence interval. Our model presented a good fit(χ2 poorly in school (lower school grades), which was revealed in both (26) = 53.11, p < .05; CFI = .98; RMSEA = .06, NFI = .96; Hu and in India and US. Our paper is informed by traditional marketing theo- Bentler 1999; Jöreskog and Sörbom 1993). The results suggest that ries of consumer and media socialization and psychological theories less social media use had a significantly reduced effect on the latent of empowerment and places the family, especially the mother and her variable, interactional empowerment (-.48; p < .05), but not intrap- media use, central to a child’s welfare. ersonal empowerment. Children’s school grades decreased (.96; p < .05; reverse-scored) as their mother’s intrapersonal empowerment REFERENCES levels increased, and they also spent less on educational expenditures Fantahun, Mesganaw, Yemane Berhane, Stig Wall, Peter Byass, and (-.99; p < .05). Meanwhile, higher levels of interactional empower- Ulf Högberg (2007), “Women’s Involvement in Household ment significantly increased educational expenditures (.21; p < .05) Decision-making and Strengthening Social Capital—Crucial while not significantly affecting their child’s school grades. There Factors for Child Survival in Ethiopia,” Acta Pædiatrica, 96, were no indirect effects of media use. The control variable school 582-9. type had a significant effect on expenditure, while the gender of the Hu, L. T., and Bentler, P. M. (1999). Cutoff criteria for fit child had a significant impact on grades. indexes in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Study 2- Mother’s Empowerment In The United States Modeling: A Multidisciplinary Journal 6(1), 1–55. doi: To understand how the use of social media impacts empower- 10.1080/10705519909540118 ment in the US, we captured active or passive usage of Facebook, Jöreskog, K. G., and Sörbom, D. (1993). LISREL VIII: User’s Instagram, Twitter or other social media sites (Li 2016). Educational reference guide. Mooresville, IN: Scientific Software.

Korupp, Sylvia E., Harry B. Ganzeboom, and Tanja Van Der Mencher, Joan (1988), “Women’s work and poverty: Women’s Lippe (2002), “Do Mothers Matter? A Comparison of Models contribution to household maintenance in South India,” A of the Influence of Mothers’ and Fathers’ Educational and home divided: Women and income in the Third World, 99-119. Occupational Status on Children’s Educational Attainment,” Senarath, Upul and Nalika Sepali Gunawardena (2009), “Women’s Quality and Quantity, 36 (1), 17-42. Autonomy in Decision Making for Health Care in South Asia,” Leung, Louis (2009), “User-generated Content on the Internet: Asia Pacific Journal of Public Health, 21 (2), 137–43. An Examination of Gratifications, Civic Engagement and Zimmerman, Marc A. (1995), “Psychological Empowerment: Psychological Empowerment,” New Media & Society, 11 (8), Issues and Illustrations,” American Journal of Community 1327-47. Psychology, 23 (5), 581-99. Li, Zongchao (2016), “Psychological Empowerment on Social Media: Who are the Empowered Users?” Public Relations Review, 42 (1), 49-59. The Interplay of Religious Discrimination/Segregation with Consumption and Production Activities Akshaya Vijayalakshmi, Indian Institute of Management Ahmedabad, Ahmedabad, India Nitisha Tomar, University of Wisconsin-Madison, Madison, USA Ankur Kapoor, Indian Institute of Management Udaipur, Udaipur, India

EXTENDED ABSTRACT To redress their dominated positions, stigmatized Muslim pro- This research focuses on the interplay of consumption and ducers rely on neoliberal logics of entrepreneurship, individual ef- production practices with religious discrimination/segregation. Al- fort, and market-mediated solutions for overcoming discrimination. though works of Izberk-Bilgin (2012) and Sandikci and Ger (2010) “Forced entrepreneurs” (in the words of a participant) emerge as they consider religion as an ideological tool for consumers’ identity proj- develop alternative markets for schools, supermarkets, and real es- ects, we analyze a context where a pernicious, discriminating struc- tate, among others. Participants worked in their ‘interest’ to improve tural order prevents the use of religion as an ideological tool. Instead, their field positions by relying on ‘market doxa’ to redress the infe- the neoliberal market discourses are implicated, which, ironically rior position imposed through ‘religious doxa’. Thus markets, and leads to reproduction of discrimination. We situate this work in prior not religion, provide necessary ideological and material resources to research on marketplace exclusion (Scaraboto and Fischer 2012), subvert structural discrimination. redlining (D’Rozario and Williams 2005), spatial exclusion (Castil- We find that even if our informants had the economic means to hos 2019), and stigmatization (Sandikci and Ger 2009). Our central out-migrate, they found it difficult, if not impossible. Barriers to mo- research questions are: (a) How does religious discrimination shape bility were either institutionally sanctioned or culturally legitimized consumption and production practices of the discriminated? (b) How (prohibition of meat consumption in many residential localities). A do these consumption and production practices, in turn, influence participant running a food business could not expand his business segregation and discrimination? as he was denied a larger commercial space in Hindu majority area. The study is set in Ahmedabad (India), which has history of Likewise, participants mention about not being granted loans eas- several interreligious conflicts, and majority of its population lives ily because they live in ‘negative areas’, along with issues in school in religiously homogeneous neighborhoods (Field et al. 2009). As admissions. Bourdieu (1986) argues that the value of one form of the context deals with aspects of distinction (between Hindus and capital lies in its ease of conversion to other forms of capital. We find Muslims) and spatial segregation, Bourdieu’s ‘thinking tools,’ such that location (physical capital) inhibits the conversion of economic as capital, doxa, interest, and symbolic violence, and Wacquant’s capital into symbolic capital. This non-conversion of capital ‘freezes’ analyses of territorial stigmatization, are useful in analyzing the data the position of the dominated in the field, thus aggravating distinc- (Grenfell 2014; Wacquant, Slater and Pereira 2014). tion and symbolic violence. Data. Using QGIS application, public/private amenities were Success stories of a few entrepreneurs provided inspiration to mapped onto each of 48 wards of Ahmedabad (Figure 1). This map- rely on one’s effort to overcome discrimination. This valorization ping helped us analyze the peripheralization that extends beyond the and internalization of market logic (individuality, self-reliance) made well-researched residential segregation - to that of access to main- participants feel responsible for their condition. Thus, market logic stream marketplaces. We also interviewed 16 Muslims who have promoted the consideration of structural and arbitrary as natural and been living in Ahmedabad since the 2000s. Participants provided in- self-inflicted, thereby precluding any collective action against sys- depth details on the marketplace discrimination experienced in their temic disadvantages. In doing so, market logic becomes complicit everyday lives - segregated housing, schooling, employment, bank in reproducing discrimination by replacing collective efforts towards loans, and public facilities. Furthermore, the authors gained a richer structural change with individual efforts towards personal change. understanding by participating in several inter-community events, The struggle of the participants to redress the stigmatization re- thus traveling to otherwise invisible parts of the city. The analyses sulted in entrepreneurial initiatives enveloped in an “identity dance” followed a hermeneutic approach (Thompson 1997) by identifying (Weinberger 2015) of selective fore-/back-grounding of religious themes within and between interviews and connecting these themes markers. Some Muslim entrepreneurs navigated inter-communal with theory. boundaries by keeping non-Muslim names for their businesses (e.g., Findings. Marketing is suggested to favor the in-group through Ruby/Best-Mart) to “pass for normal” (Goffman 1963). Unlike marketing mix, store location, delivery modes, etc. (Bennett et al. Crockett and Wallendorf (2004), some entrepreneurs assigned Mus- 2016). The GIS map and interviews illuminate redlining (paucity of lim names to their businesses (e.g. Bismillah) to reinforce collective retail stores, restaurants, and banks) in Muslim majority areas, lead- Muslim identity, but only when catering to a Muslim clientele. ing to significant differences in amenities between Hindu and Mus- Discussion. Our work contributes to theory in the following lim majority areas. Participants also mentioned being underserved ways. First, we uncover how lack of legitimate physical capital by public or private schools and paying higher prices for electricity. inhibits the conversion of economic capital into symbolic capital. Markets become a tool for expressing the religious doxa and repro- Thus, we address the limited attention Bourdieu (1986) accords to ducing field positions through spatial segregation (Bourdieu 1986, physical space. Second, we uncover the dual role of markets, in the Wacquant et al. 2014). Such structuring influences production/conway they help survive discrimination by enabling means of living but sumption opportunities of the dominated. Interestingly, such a pro- foreclose opportunities of collective action and structural changes, cess not only leads to inferior outcomes for Muslim consumers, but through individualization and responsibilization. By foreclosing col- also hurts businesses that accommodate inefficiency/losses to main- lective action, the market’s neoliberal ethos appears to reproduce dis- tain social legitimacy (by not serving Muslim dominant areas). Thus, tinction and segregation. markets forego ideals of profit maximization and customer satisfac- Our work also informs contexts wherein spatial segregation is tion to align with discriminating structural forces. institutionally or culturally legitimized (refugee camps, gated communities, redlining, etc.). Often, it is believed that market logic will

Advances in Consumer Research 764 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 765 eventually resolve underlying segregation/discrimination. We found Izberk-Bilgin, E., (2012). Infidel brands: unveiling alternative that without efforts through other collective means, the market may meanings of global brands at the nexus of globalization, aggravate the conditions it is meant to alleviate. We attempt to un- consumer culture, and Islamism. Journal of Consumer cover the complex interplay between marketplace activities of in- Research, 39(4), 663-687. dividuals and structural aspects of discrimination/segregation, thus Saatcioglu, B. and Ozanne, J.L. (2013). Moral habitus and status informing theory and public discourse on these issues. negotiation in a marginalized working-class neighborhood. Journal of Consumer Research, 40(4), 692-710. REFERENCES Sandikci, Ö., & Ger, G. (2009). Veiling in style: how does a Bourdieu, P. (1986). The forms of capital. In J. G. Richardson stigmatized practice become fashionable?. Journal of (Ed.), Handbook of theory and research for the sociology of Consumer Research, 37(1), 15-36. education (241–258). New York, NY: Greenwood Press. Scaraboto, D., & Fischer, E. (2012). Frustrated fatshionistas: An Castilhos, R.B. (2019). Branded places and marketplace exclusion. institutional theory perspective on consumer quests for greater Consumption Markets & Culture, 22(5-6), 582-597. choice in mainstream markets. Journal of Consumer Research, Crockett, D. and Wallendorf, M. (2004). The role of normative 39(6), 1234-1257. political ideology in consumer behavior. Journal of Consumer Wacquant, L., Slater, T. and Pereira, V.B. (2014). Territorial Research, 31(3), 511-528. stigmatization in action. Environment and Planning A, 46(6), D’Rozario, D. and Williams, J.D. (2005). Retail redlining: 1270-1280. Definition, theory, typology, and measurement.Journal of Weinberger, M.F., (2015). Dominant consumption rituals and Macromarketing, 25(2), 175-186. intragroup boundary work: How non-celebrants manage Field, E., Levinson, M., Pande, R., & Visaria, S. (2008). conflicting relational and identity goals.Journal of Consumer Segregation, rent control, and riots: The economics of Research, 42(3), 378-400. religious conflict in an Indian city. American Economic Weinberger, M.F., & Crockett, D. (2018). Critical reflections on Review, 98(2), 505-10. consumer identity. In E. J. Arnould & C. J. Thompson (Eds.), Grenfell, M.J. (2014). Pierre Bourdieu: Key Concepts. Routledge. Consumer culture theory (62-84). UK: SAGE Publications. Frequent Consumption of the Sharing Services Racheal Louis Vincent, Sunway University, Malaysia Sanjaya Singh Gaur, New York University, USA

EXTENDED ABSTRACT use (PDOU) and their experience of using shared services. There- fore, to address these research gaps, our study is guided by four re- Introduction search questions. These questions are: What influences people to fre- The concept of the sharing economy is more than just a buzz- quently consume shared services? How the perception of difficulty word. By 2025, the sharing economy is expected to be valued at $335 in using the shared services affects frequent usage behavior? What is billion (Kindergan 2015). The term ‘shared services’ in our study the role of past experience in consumption of shared services? How is with regards to the services provided in the sharing economy social norms affect intention to participate or not to participate in the such as ride-sharing (e.g., Uber and Lyft), accommodation sharing sharing economy? (e.g., Airbnb and Couchsurfing), task-sharing (e.g., Taskrabbit and Airtasker), closet-sharing (e.g., Rent the Runway and Style Lend) Conceptual Framework and many more. These types of services are labelled as ‘shared ser- According to Ajzen and Fishbein (2005), human behavior is vices’ because they provide access without ownership (Eckhardt et guided by attitudes and therefore, attitudinal theories are needed to al. 2019, Kumar, Lahiri, and Dogan 2018). understand and explain individual’s consumption of shared services. With the support of digital and mobile technology, these shared Our research framework draws advances from attitudinal theories services apps (mobile applications) offer consumers an easier access explaining individuals’ frequent usage behavior together with extant to a variety of on-demand goods, services and underutilized assets knowledge of shared consumption. We grounded our research frame- for monetary gains. We know that consumers will repeat their ser- work on Theory of Trying (TT) by Bagozzi and Warshaw (1990) and vice usage if they find that the company provides some sort of value Extended Model of Goal-Directed Behavior (EMGB) by Perugini (intrinsic and/or extrinsic) to them. For example, Uber reported their and Corner (2000) together with frequent usage behavior, perceived “Monthly Active Platform Consumers” (MAPCs) of 91 million in difficulty of use (PDOU) and experience of using shared services to the fourth quarter of 2018. These are users who were found to take explain individual’s shared services behavior. Applying one theory at least one ride on Uber or buy at least one meal on Uber Eats (Sa- to the domain of the other has potential to generate novel insights. linas and Feiner 2019). For explaining individual’s frequent usage of According to Mayer and Sparrowe (2013) in order for integration such shared services, we posit that the difficulty (or ease) of use and to be successful, it is important for a clearly articulated link be- the overall experience of using shared services plays a major role in tween the theory and the new domain to be present as it allows for influencing their usage behavior. productive dialogue. We propose an integration of both mentioned To date, little is known about individual’s shared services fre- theories as they complement each other. TT’s three determinants: quent usage behavior. The focus of the extant literature seems to be attitude toward success, attitude toward failure and attitude toward more inclined towards the motivations and barriers for using shared the process also known as the pre-factual attitudes allow a better services and overlooks the importance of understanding individual’s understanding of individual’s shared services frequent user behav- frequent usage behavior of these services. Past scholarly work done ior when integrated with EMGB. Individuals using shared service on understanding of human behavior seem to emphasize on the may experience impediments and that could drastically affect their intention-behavior relationship (Ajzen 1991, Godin and Kok 1996, shared services usage frequency. On the other hand, EMGB is said Sheeran and Orbell 1998, 2000b, Warshaw and Davis 1985, Sheeran to have outperformed the TPB and has shown substantial predictive and Orbell 2000a) and have yet to consider actual frequent usage be- power (Perugini and Corner 2000, Lee et al. 2012). To the best of havior. The more frequent and recent an individual’s usage behavior our knowledge, this is a first empirical study to integrate two attitu- (Bagozzi and Dholakia 2002), the more likely he/she will repeat their dinal theories: Theory of Trying (TT) and Extended Model of Goal- usage behavior again in the future. Directed Behavior (EMGB) to further explain individual’s frequent With regards to adoption of technology, many studies have pri- usage of shared services. marily focused on perceived ease of use (PEOU) (Fagan, Neill, and Wooldridge 2008, Luarn and Lin 2005, Saadé and Bahli 2005, Ven- Method katesh 1999, Gefen, Karahanna, and Straub 2003, Hansen, Saridakis, Data for the study were collected using a mall-intercept survey and Benson 2018) and limited studies have considered perceived using structured questionnaire. Measures for all the variables used difficulty of use (PDOU) in terms of understanding consumer’s us- in this study were drawn from well-established literature and were age behavior (Trafimow et al. 2002, Barnard et al. 2013, Venkatesh modified to suit the present study. Out of 500 respondents partici- 2000). However, in actual reality, it is the perceived difficulty of use pated in the mall-intercept survey, 393 responses were found to be (PDOU) that influences frequent usage of the adopted services. This usable. Partial Least Squares based Structural Equation Modelling is because we postulate that it is much easier for individuals to ex- (PLS-SEM) was used to test all the hypotheses in the conceptual press their difficulty in using shared services rather than the ease in framework. The SmartPLS 3.0 software was used for the testing using these services. Besides, it was found that consumers showed of both measurement model and structural model. PLS based SEM higher intentions to purchase when they had favorable past purchas- technique was used mainly because it provides a useful estimate ing experiences (Weisberg, Te’eni, and Arman 2011). Thus, we also for a complex framework with many constructs and many indica- examine the role of the experience in making consumers use the tors (Chin 1998, Hair et al. 2011, Ringle, Sarstedt, and Straub 2012). same services frequently. Moreover, Hair et al. (2011), Hair et al. (2014), Henseler, Ringle, and Based on the above backdrop, the aim of our research is to un- Sinkovics (2009), highlighted that PLS-SEM has been widely used derstand individual’s consumption of shared services particularly in marketing research and also been used in many top-tier and other their frequent usage behavior, and the role of perceived difficulty of leading journal publications.

Advances in Consumer Research 766 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 767 Discussion REFERENCES Our findings offer new insights for understanding individual’s Ajzen, I. 1991. “The Theory of Planned Behavior.” Organizational consumption of shared services particularly their frequent usage be- Behavior and Human Decision Processes 50:179-211. havior, the role of perceived difficulty of use (PDOU) and experi- Ajzen, Icek , and Martin Fishbein. 2005. The Influence of Attitudes ence of using shared services. Consequently, our research augments on Behavior. In The Handbook of Attitudes. the Theory of Trying (TT) and Extended Model of Goal-Directed Bagozzi, Richard P, and Utpal M Dholakia. 2002. “Intentional Behavior (EMGB) in several ways. Empirical findings of the study social action in virtual communities.” Journal of Interactive prove to show the proposed framework of the study developed by in- Marketing 16 (2):2-21. tegration of theories was successful in explaining individual’s shared Bagozzi, Richard P, and Paul R Warshaw. 1990. “Trying to services frequent behavior. Consume.” Journal of Consumer Research 17 (2):127-140. Pre-factual attitudes towards success (feelings towards success, Barnard, Yvonne, Mike D. Bradley, Frances Hodgson, and Ashley expectation of success and evaluation of success), attitude towards D. Lloyd. 2013. “Learning To Use New Technologies By failure (feelings towards failure), attitude towards the process (feel- Older Adults: Perceived Difficulties, Experimentation ings towards process) were found to influence individual’s attitude Behaviour and Usability.” Computers in Human Behavior 29 towards using shared services. Similarly, we found attitude towards (4):1715-1724. use, subjective norms and group norms significantly influence in- Chin, W. W. 1998. “The Partial Least Squares Approach to dividual’s desire for using shared services. Desire to use shared Structural Equation Modeling.” In Modern Methods for services and perceived behavioural control were found to influence Business Research, edited by G. A Marcoulides, 295-336. individual’s intention for using shared services. Besides that, we also Lawrence Erlbaum Associates. found intention to use and subjective norms significantly influence Eckhardt, Giana M., Mark B. Houston, Baojun Jiang, Cait individual’s frequent usage of shared services. Perceived difficulty Lamberton, Aric Rindfleisch, and Georgios Zervas. 2019. of use (PDOU) on the other hand, was found to be negatively associ- “Marketing in the Sharing Economy.” Journal of Marketing ated with individual’s frequent usage behaviour of shared services 83 (5):5-27. and experience of using shared services was found to be negatively Fagan, Mary Helen , Stern Neill, and Barbara Ross Wooldridge. associated with PDOU. 2008. “Exploring the Intention to Use Computers: An Empirical Investigation of the Role of Intrinsic Motivation, Conclusions Extrinsic Motivation, and Perceived Ease of Use.” Journal of As mentioned earlier, the primary objective of this study was to Computer Information Systems 48 (3):31-37. understand individual’s consumption of shared services particularly Gefen, David , Elena Karahanna, and Detmar W. Straub. 2003. their frequent usage behavior, and the role of perceived difficulty of “Trust and TAM in Online Shopping: An Integrated Model.” use (PDOU) and experience of using shared services. Our findings MIS Quarterly 27 (1):51-90. show that individual’s pre-factual thought process influences their Godin, G, and G Kok. 1996. “The Theory of Planned Behavior: attitudes towards using the sharing services. How individual’s feel, A Review of Its Applications to Health-Related Behaviors.” expect, evaluate success or failure and their overall thought process American Journal of Health Promotion 11 (2):87-98. of using shared services highly depends on the sharing service pro- Hair, Joe F. Jr, Marko Sarstedt, Lucas Hopkins, and Volker G viders’. Moreover, perceived difficulty of use (PDOU) and experi- Kuppelwieser. 2014. “Partial Least Squares Structural ence of using shared services were found to play an important role in Equation Modeling (PLS-SEM).” European Business Review terms of influencing individual’s frequent consumption of the shar- 26 (2):106-121. ing services. Hair, Joe F. Jr, Marko Sarstedt, Christian M Ringle, and Jeannette There are three main contributions of our study from the per- A Mena. 2011. “An Assessment of The Use of Partial spective of academia. The first contribution is in terms of the expla- Least Squares Structural Equation Modeling in Marketing nation for the frequent consumption (usage) of shared services. Sec- Research.” Journal of the Academy of Marketing Science 40 ond contribution is in terms of the inclusion of perceived difficulty of (3):414-433. use (PDOU) variable in the existing behavioral models which have Hansen, Jared M. , George Saridakis, and Vladlena Benson. 2018. primarily considered perceived ease of use (PEOU). The third contri- “Risk, Trust, and the Interaction of Perceived Ease of Use and bution comes from explaining the role of experience of using shared Behavioral Control in Predicting Consumers’ Use of Social services in the frequent usage behavior. Media for Transactions.” Computers in Human Behavior This study also provides several practical implications for the 80:197-206. key stakeholder of the shared economy, such as shared service pro- Henseler, Jörg, Christian M Ringle, and Rudolf R Sinkovics. viders (operators). It is imperative for the sharing service providers 2009. “The Use of Partial Least Squares Path Modeling in to create a positive brand experience and minimize service failures. International Marketing.” Advances in International Marketing This is vital because experience of using shared services not only 20:277–319. affects individual’s perceived difficulty of use (PDOU) but also their Kindergan, Ashley. 2015. Credit Suisse: By 2025, Companies could frequent usage behavior. Rake In $335 Billion a Year From People ‘Sharing’. Accessed It is important for the sharing service providers to educate con- October 7, 2016. sumers on using the features of their mobile apps such as by includ- Kumar, V., Avishek Lahiri, and Orhan Bahadir Dogan. 2018. “A ing step-by step video of use in the app itself especially for the older Strategic Framework for A Profitable Business Model in The consumers who often have anxiety towards technology adoption. Sharing Economy.” Industrial Marketing Management 69:147- The findings of our study are indeed valuable not only for future 160. scholars who are keen to study similar technology adoption but also provide useful guidelines for the sharing economy practitioners. 768 / Frequent Consumption of the Sharing Services

Lee, H Song, L Bendle, M Kim, and H Han. 2012. “The Impact of Sheeran, P, and S Orbell. 2000a. “Self-Schemas and The Theory of Non-Pharmaceutical Interventions For 2009 H1N1 Influenza Planned Behaviour.” European Journal of Social Psychology on Travel Intentions: A Model of Goal-Directed Behavior.” 30 (4):533-550. Tourism Management 33 (1):89-99. Sheeran, P, and S Orbell. 2000b. “Using Implementation Intentions Luarn, Pin, and Hsin-Hui Lin. 2005. “Toward an Understanding to Increase Attendance for Cervical Cancer Screening.” Of The Behavioral Intention To Use Mobile Banking.” Health Psychology Review 19 (3):283-298. Computers in Human Behavior 21 (6):873-891. Trafimow, David, Paschal Sheeran, Mark Conner, and Krystina A. Mayer, Kyle J., and Raymond T. Sparrowe. 2013. “Integrating Finlay. 2002. “Evidence That Perceived Behavioural Control Theories in AMJ Articles.” Academy of Management Journal is a Multidimensional Construct: Perceived Control and 56 (4):917-922. Perceived Difficulty.” British Journal of Social Psychology 41 Perugini, M, and M Corner. 2000. “Predicting and Understanding (1):101-121. Behavioral Volitions: The Interplay Between Goals and Venkatesh, V. 2000. “Determinants of Perceived Ease of Use: Behaviors.” European Journal of Social Psychology 30:705- Integrating Control, Intrinsic Motivation, And Emotion into 731. the Technology Acceptance Model.” Information Systems Ringle, C. M, M Sarstedt, and D Straub. 2012. “A Critical Look at Research 11 (4):342-365. the Use of PLS-SEM.” MIS Quarterly 36 (1):iii-xiv. Venkatesh, Viswanath. 1999. “Creation of Favorable User Saadé, Raafat, and Bouchaib Bahli. 2005. “The Impact of Cognitive Perceptions: Exploring The Role of Intrinsic Motivation.” Absorption on Perceived Usefulness and Perceived Ease of MIS Quarterly 23 (2):239-260. Use in On-Line Learning: An Extension of The Technology Warshaw, P, and F Davis. 1985. “Disentangling Behavioral Acceptance Model.” Information & Management 42 (2):317- Intention and Behavioral Expectation.” Journal of 327. Experimental Social Psychology 21 (3):213-228. Salinas, Sara , and Lauren Feiner. 2019. Uber Releases Its Long- Weisberg, Jacob, Dov Te’eni, and Limor Arman. 2011. “Past Awaited Ipo Filing. Accessed January 13, 2020. Purchase and Intention to Purchase in E‐Commerce.” Internet Sheeran, P, and S Orbell. 1998. “Do Intentions Predict Condom Research 21 (1):82-96. Use? Meta-Analysis and Examination of Six Moderator Variables.” British Journal of Social Psychology, 37 (2):231- 250. Deriving Mental Energy from Task Completion Xiang Wang, University of Florida, USA Chris Janiszewski, University of Florida, USA Yanmei Zheng, University of Hawaii, USA Juliano Laran, University of Miami, USA Wonseok Jang, Sungkyunkwan University, South Korea

EXTENDED ABSTRACT (p=.88; interaction: p=.011). Felt achievement and competence did People often complain that they have plenty to do, but not not differ across conditions. enough time to do it. Yet, time is not the only resource that limits Study 2 showed behavioral implications of energy replenish- daily productivity. People also have limited mental energy. Approxi- ment – persistence in a subsequent task. Participants completed the mately 20% of Western adults experience mental fatigue on a regu- same deal-finding task, except that task completion was manipulated lar basis (Lerdal et al., 2005; Pawlikowska et al., 1994). Despite the between-subjects by whether or not telling participants that the deal- importance of mental energy, and the fact that many daily tasks are finding task was completed after they finishedth the5 deal-finding naturally depleting, consumer research on how to replenish mental task. Then all participants were directed to a book-evaluation task energy is limited. where they could choose to quit at any time. Task persistence was Mental energy is defined asmomentary feelings of alertness that measured by the number of books participants evaluated before they supports the ability to remain focused on and complete ongoing cog- quit. Participants who were informed of task completion evaluated nitive tasks (O’Connor 2006; Ryan and Frederick 1997). Existing more books than those who were not when reward value was high research investigates the exogenous source of energy replenishment (p=.031) but not when reward value was low (p=.31; interaction: (i.e., a source that is independent of the energy-consuming task) such p=.025). as taking a rest (Helton and Russell 2017) and engaging in intrinsi- Study 3 tested moderation by task effort. The task involved cally interesting activities (e.g., leisure activities) (Reis et al., 2000). learning good sources of six micro-nutrients among five groups of We suggest an endogenous source of energy replenishment – the food. Within each group, participants saw six food items and their completion of a goal-directed task – and propose that energy replen- nutrition facts and needed to identify items that were good sources of ishment at task completion depends on the value of task rewards. each nutrient. Nutrition facts of different foods were listed in % daily People’s energy level increases compared to pre-completion levels value per serving that can be directly compared (low-effort) orin dif- when completing high-reward tasks, but not low-reward tasks. We ferent serving sizes that required additional calculation (high-effort). argue that task completion is the time when people obtain and reflect The food was either easy to find in grocery stores (e.g., watermelon, upon task rewards (Gollwitzer 2012; Heckhausen and Gollwitzer blueberries; high-reward) or not (e.g., Kiwano, Loquat; low-reward). 1987), which creates mental energy. As suppression of one’s needs, In the high-effort condition, mental energy was higher post-comple- or regulation of the behavior that meets one’s needs, consumes mention than pre-completion (i.e., before working on the 5th group) in the tal energy (Baumeister et al. 1998; Vohs and Faber 2007), it is pos- high-reward task (p=.001) but not in the low-reward task (p=.914; sible that satisfying one’s needs (e.g., rewards) boosts mental energy. interaction: p=.016). In the low effort condition, however, mental We further suggest two moderators of the effect of reward value energy was replenished regardless of reward value (main effect of on energy replenishment: task effort and task focus (reward vs. mere completion: p=.043; interaction: p=.561). Study 3 also ruled out al- completion). Reward is a consequence of cognitive control during ternative explanations of positive affect, felt achievement, and com- goal pursuit, which is subjectively costly and influences people’s re- petence. sponses toward the rewards (Botvinick and Braver 2015; Inzlicht et Study 4 tested moderation by task focus. Participants were al. 2018). When task effort is high, task completion should generate asked to 5 hand-written ship logs into digital texts. Reward value was mental energy only when reward value is high because the invest- manipulated by framing the task as helping scientists input important ment of energy was worthwhile. When effort is low, energy replen- historical data (high-reward) or as digitizing logs that were not very ishment should be less sensitive to reward value and will occur at the useful (low-reward). Task focus was manipulated by adding visual completion of both high-reward and low-reward tasks because both cues that emphasized task completion (an arrow pointing to an icon tasks were worthwhile given the minimal investment of effort. signaling the end of the task), or not adding such cues, so that task re- Finally, we propose that there are two sources of the reward ward was more salient (similar to prior studies). In the reward-focus experienced at completion: the act of completing the task itself and condition, mental energy was higher post-completion than pre-com- the rewards associated with completing the task. If this is true, men- pletion in the high-reward task (p=.011) but not in the low-reward tal energy replenishment should be a function of reward value when task (p=.511; interaction: p=.023). However, in the completion-focus people focus on the rewards associated with task completion; when condition, mental energy was replenished regardless of reward value people focus on the mere act of task completion, mental energy (main effect of completion:p =.058; interaction: p=.536). should be replenished as long as the task is completed, regardless of Our findings suggest that mental energy can be replenished in task reward. accordance with the reward experienced at task completion. This Study 1 demonstrated that mental energy is replenished upon energy replenishment strategy could be conceptualized as an evo- completion of a high-reward task, but not a low-reward task. The lutionary advantage, as it is an endogenous source that sustains the task involved finding the best deals (i.e., lowest price) for five prod- completion of a sequence of tasks. These insights suggest the need ucts. Task reward was manipulated by whether or not participants for a broader conceptualization of the generation of mental energy. learned that they would get a $0.1 bonus for each best deal they found. Participants reported higher mental energy post-completion than pre-completion (i.e., before searching for the 5th product) when reward value was high (p<.001) but not when reward value was low

Advances in Consumer Research 769 Volume 48, ©2020 770 / Deriving Mental Energy from Task Completion REFERENCES Lerdal, Anners, Astrid K. Wahl, Tone Rustoen, Berit R. Hanestad, Baumeister, Roy F., Ellen Bratslavsky, Mark Muraven, and Dianne and Torbjorn Moum (2005), “Fatigue in the General M. Tice (1998), “Ego Depletion: Is the Active Self a Limited Population: A Translation and Test of the Psychometric Resource?” Journal of Personality and Social Psychology, 74 Properties of the Norwegian Version of the Fatigue Severity (May), 1252-65. Scale,” Scandinavian Journal of Public Health, 33 (2), 123-30. Botvinick, Matthew, and Todd Braver (2015), “Motivation and O’Connor, Patrick J. (2006), “Mental Energy: Developing a Model Cognitive Control: From Behavior to Neural Mechanism,” for Examining Nutrition-Related Claims,” Nutrition Reviews, Annual Review of Psychology, 66 (January), 83-113. 64 (7), S2. Gollwitzer, Peter (2012), “Mindset Theory of Action Phases.” Pawlikowska T., T. Chalder, S. R. Hirsch, P. Wallace, D. J. Wright, in Handbook of theories of social psychology, P. A. M. Van and S. C. Wessely (1994), “Population based Study of Fatigue Lange, A. W. Kruglanski, & E. T. Higgins (Eds.), Thousand and Psychological Distress,” British Journal of Medicine, 308 Oaks, CA: Sage Publications Ltd, 526-45. (March), 763–6. Heckhausen, Heinz, and Peter M. Gollwitzer (1987), “Thought Reis, Harry T., Kennon M. Sheldon, Shelly L. Gable, Joseph Contents and Cognitive Functioning in Motivational versus Roscoe, and Richard M. Ryan (2000), “Daily Well-Being: Volitional States of Mind,” Motivation and Emotion, 11 (2), The Role of Autonomy, Competence, and Relatedness,” 101-20. Personality and Social Psychology Bulletin, 26 (4), 419-35. Helton, William S. and Paul N. Russell (2017), “Rest is Still Ryan, Richard M. and Christina Frederick (1997), “On Energy, Best: The Role of the Qualitative and Quantitative Load of Personality, and Health: Subjective Vitality as a Dynamic Interruptions on Vigilance,” Human Factors, 59 (February), Reflection of Well-being,”Journal of Personality, 65 91-100. (September), 529-65. Inzlicht, Michael, Amitai Shenhav, and Christopher Y. Olivola Vohs, Kathleen D., and Ronald J. Faber (2007), “Spent Resources: (2018), “The Effort Paradox: Effort is Both Costly and Self-Regulatory Resource Availability Affects Impulse Valued,” Trends in Cognitive Sciences, 22 (4), 337-349. Buying.” Journal of Consumer Research, 33 (March), 537-47. The Looking-Glass Self: Low Self-Concept Clarity Increases Appearance Management Jiaqian (Jane) Wang, Kellogg School of Management, Northwestern University, USA Yiqi Yu, Guanghua School of Management, Peking University, China

EXTENDED ABSTRACT low self-concept clarity predicted greater concerns about physical Recent years have witnessed the rapid expansion of beauty attractiveness (B = .143, SE = .022; p < .001). More importantly, industries. Yet, limited research has shed light on when and why mediation analysis using SPSS PROCESS macro Model 4 revealed consumers are motivated to manage their appearance. The current that the indirect effect of public self-consciousness did not contain research examines how self-concept clarity, a structural feature of zero (95% CI = [.04, .07]), suggesting that public self-consciousness self-concept, shapes appearance management. mediated the association between low self-concept clarity and con- Self-Concept clarity is defined as the possession of “clearly cerns about physical attractiveness. defined, temporally stable, and internally consistent self-views” Next, Study 4 (N = 219; Mturk) causally tested the mediating (Campbell et al. 1996). Maintaining a clear self-concept is an essen- role of public self-consciousness. We manipulated self-concept clar- tial psychological need and can guide how consumers behave. Thus, ity using a typing task which ostensibly assessed typing speed and when lacking self-concept clarity, consumers strive to understand accuracy. Participants typed in six sentences describing high or low who they are and what to do. Because social standards and evalua- self-concept clarity word by word. After completing the manipulations can serve as a reference (Cooley 1902), low self-concept clarity tion check and measures on public self-consciousness, participants increases consumers’ sensitivity to social information. Indeed, when read an introduction to cosmetic procedures and indicated their inter- experiencing low self-concept clarity, consumers are more likely to est in cosmetic procedures (Henderson-King and Henderson-King engage in social comparisons, adopt social standards and ideals, and 2005). Then, they responded to items on private self-consciousness conform to others’ expectations to seek social approval (e.g. Mittal and global self-esteem. As predicted, low self-concept clarity led to

2015). These behaviors manifest the tendency to view the self as em- greater interest in cosmetic procedures (Mlow = 2.47, SD = 1.82 vs. bedded in social contexts, which is termed as public self-conscious- Mhigh = 2.06, SD = 1.54; p = .072, d = .24). Mediation analysis using ness (Fenigstein, Scheier and Buss 1975). It triggers the self-presen- PROCESS macro Model 4 suggested that this effect was mediated by tation motive which drives consumers to build and exhibit a positive heightened public self-consciousness (indirect effect: 95% CI = [.01, self-image to others (Schlenker and Weigold 1990). To the extent .31]) but not global self-esteem (95% CI = [–.12, .05]) or private self- that physical attractiveness is one of the most observable, explicit consciousness (95% CI = [–.12, .08]). factors in social judgment (Freeman and Ambaby 2011), consumers Finally, Study 5 (N = 394; Mturk) employed a process-by-mod- with strong public self-consciousness tend to be highly involved in eration approach to further test the proposed mechanism. Participants appearance management to obtain social approvals. were randomly assigned to one of four conditions in a 2 (self-concept Taken together, we predict that low-concept clarity increases clarity: low vs. high) × 2 (self-presentational value: positive vs. neg- appearance management by heightening public self-consciousness. ative), between-subject design. Participants first completed a writing This effect will be attenuated when appearance management entails task which manipulated self-concept clarity. Then, those in the posi- negative self-presentational value (i.e. brings unfavorable social tive (negative) self-presentational value group read about how Botox judgments). could lead to “more compliments” (“being judged as inauthentic”). Five studies using a survey, archival data, and experiments test- Next, participants indicated their interest in Botox. Results revealed ed these propositions. Study 1 (N = 300; Mturk) showed that dispo- a significant main effect of self-concept clarity (F(1, 390) = 4.35, p 2 sitional low self-concept clarity was associated with more frequent = .038, ηp = .011), no main effect of self-presentational value (F(1, 2 appearance management behaviors (e.g. tanning; B = .117, p < .001) 390) = .95, p = .329, ηp =.002), and a significant interaction (F(1, 2 and stronger preferences for three appearance-improving products 390) = 5.84, p = .016, ηp = .015). A planned contrast suggested that (ps < .001). when Botox led to positive social evaluations, participants in the low

Study 2 (N = 113; Chinese college students) used a two-condi- self-concept clarity condition showed greater interest in Botox (Mlow tion (self-concept clarity: low vs. high), between-subject design to = 2.62, SD = 1.98 vs. Mhigh = 1.85, SD = 1.61; p = .002). However, test the causal effect of self-concept clarity on interest in getting in- when negative social evaluations were highlighted, self-concept clar- formation regarding appearance management. We also explored the ity had no effect (Mlow = 2.04, SD = 1.61 vs. Mhigh = 2.09, SD = 1.62; role of global self-esteem. Results showed that low self-concept clar- p = .810). ity increased interest in learning more about appearance management Theoretically, the current work extends research on self-concept, public self-consciousness, and appearance management. Practi- (Mlow = 6.00, SD = 1.02 vs. Mhigh = 5.46, SD = 1.22; F(1,112) = 6.43, cally, our findings suggest that beauty industries could benefit from p = .013, d = .48) without affecting self-esteem (Mlow = 4.58, SD = combating stereotypes and stigma associated with beauty striving. 1.01 vs. Mhigh = 4.74, SD = .97; F(1,112) = .73, p = .394, d = .16). Also, the effect of self-concept clarity on appearance management Moreover, to tackle consumers’ beauty obsession, it is vital to elimi- held when controlling for global self-esteem (p = .011), suggesting nate the overemphasis on physical attractiveness and promote more that self-esteem was not a likely mediator. inclusive and diverse beauty in society. We then investigated the underlying mechanisms of the effect. Using a secondary data set (the Midlife in the United States survey), REFERENCES Study 3 examined the link between low self-concept clarity, public Campbell, Jennifer D., Paul D. Trapnell, Steven J. Heine, Ilana self-consciousness, and appearance management. Regression results M. Katz, Loraine F. Lavallee, and Darrin R. Lehman (1996), showed that dispositional low self-concept clarity was associated “Self–Concept Clarity: Measurement, Personality Correlates, with greater public self-consciousness (N = 2651; B = .608, SE = and Cultural Boundaries,” Journal of Personality and Social .033; p < .001). Among female respondents who indicated how much Psychology, 70 (1), 141–56. they worried about being less attractive (N = 1413), we found that

Cooley, Charles Horton (1902), Human Nature and the Social Henderson-King, D., & Henderson-King, E. (2005). Acceptance Order, NY: Scribner. of cosmetic surgery: Scale development and validation. Body Fenigstein, Allan, Michael F. Scheier, and Arnold H. Buss (1975), image, 2(2), 137-149. “Public and Private Self–Consciousness: Assessment and Mittal, Banwari (2015), “Self–Concept Clarity: Exploring Its Role Theory,” Journal of Consulting and Clinical Psychology, 43 in Consumer Behavior,” Journal of Economic Psychology, 46, (4), 522–27. 98–110. Freeman, Jonathan B. and Nalini Ambady (2011), “A Dynamic Schlenker, Barry R. and Michael F. Weigold (1990), “Self– Interactive Theory of Person Construal,” Psychological Consciousness and Self–Presentation: Being Autonomous Review, 118 (2), 247–79. versus Appearing Autonomous,” Journal of Personality and Social Psychology, 59 (4), 820–28. Loss Aversion: The Impact of Uncertainty on Experiential Product Preference Lili Wang, Zhejiang University, China Xuan Zhang, Zhejiang University, China Wei Lyu, Shanghai Jiao Tong University, China EXTEND ABSTRACT To further confirm this effect is driven by uncertainty’ aversion Individuals experience uncertain at a high frequency for all effect rather than certainty’s complementary effect, the next study ages. It seems that people make almost all of the key decisions with added a pure control condition to further examine these two competi- various kinds of uncertainty, such as investment and life planning. tive accounts and tested the mediating role of psychological owner- Uncertainty has a great influence on consumers’ emotion, judgement, ship. Participants (N = 308) were randomly assigned to one of the decision making and behavior. How does uncertainty influence con- three certainty manipulation conditions and then made a choice from sumers’ product preference, particularly for experiential vs. material a laptop and a travel. After that, they completed the psychological preference? ownership measure. Results showed that participants in uncertainty Uncertainty increases reliance on emotional attachment when condition (M = 44.33%) were marginally less likely to choose travel making decisions (Faraji-Rad and Pham 2017), repetition behavior than participants in certainty-condition (M = 57.43%; b = -. 53 (.29), for uncertain reward (Shen, Hsee and Talloen 2019), resource invest- z = 1.84; p = .066) and in control condition (M = 61.26%, b = .69 ment (Shen, Fishbach and Hsee 2015), and motivations to maintain a (.28), z = 2.43; p = .015). There was no significant difference between stable self-concept (Savary and Dahr 2019). Building on these find- certainty (57.43%) and control condition (61.26%, b = -.16 (.28), z = ings, we propose that uncertainty leads to a loss of psychological 0.57; p = .570). In addition, the impact of uncertainty on experiential ownership by impairing belonging, self-identity and self-efficacy product aversion is driven by relative sense of psychological owner- (Elstak et al., 2015; Pierce et al., 2001). Compared to material prod- ship from experiential vs. material product. We ran two process mod- uct, experiential product provides shorter longevity (Tully, Hersh- el and in support of our hypothesis, relative perceived psychological field, and Meyvis 2015; Weidman and Dunn 2016), lower identity ownership mediated the effect of sense of uncertainty on experiential association (Dommer and Swaminathan 2013; Morewedge and Gib- product preference. lin 2015; Weiss and Johar 2013, 2016) and is unalterable (Rosenz- Study 4 investigated the moderator of perceived psychologi- weig and Gilovich 2012). Thus, to some extent, experiential product cal ownership. Participants (N = 290) completed certainty manipu- would have lower psychological ownership than material product. lation, gift choice and perceived relative psychological ownership. Therefore, for experiential vs. material product preference under un- Results showed a significant interaction between state and relative certainty, there might be two competitive accounts: psychological ownership (b = .33(.17), z = 1.96, p = .051). Floodlight 1) From loss aversion perspective, sense of uncertainty should analysis showed that those perceived experiential product had lower reduce experiential (vs. material) product preference to avoid further psychological ownership compared to material product (relative loss of psychological ownership. psychological ownership score of 2.86 or lower) were more likely 2) For complementary consumption perspective, sense of un- avoid to choose travel as the gift under uncertainty condition. Once certainty might increase material (vs. experiential) product prefer- perceived relative psychological ownership of experiential product ence to restore lost psychological ownership. increases, this negative effect would disappear. Across five studies, we attempt to examine how sense of un- Study 5 directly manipulated psychological ownership in a real certainty influences experiential vs. material product preference and choice context. Students (N = 231) completed certainty manipulation which account driven this effect. task, psychological ownership manipulation task, and then made a Study 1 examined the relationships between participants’ gen- real choice between an umbrella (material product) and motive ticket eral sense of uncertainty and experiential (vs. material) product pref- (experiential product, pretested). Logistic regression showed a mar- erence. Students (N = 260) completed different kinds of insecurity ginally significant interaction between state and relative perceived 2 (Wittkowski 2001) as a proxy of uncertainty perception. Then, expe- psychological ownership (b = 1.05(.54), z = 1.93, p = .054, ηp = .02). riential product preference (Ryan, Paulina and Ravi 2012). Results Decomposing this interaction showed in the control condition, par- showed that general sense of uncertainty was negatively related to ticipants were less likely to choose a film ticket as a bonus when they preference for experiential product (b = -.29 (.09), t(258) = -3.18, p were uncertain (34.55%) than when they were certain (M = 50.85%; = .002). Study 2 tested the main effect of uncertainty on experiential z = -1.75, p = .081). However, in the high psychological ownership (vs. material) product preference. Participants (N = 226) firstly com- condition, participants showed no preference difference for the film pleted the certainty manipulation task (Faraji-Rad and Pham 2017). ticket and the umbrella when they were certain (33.33%) and when Then participants read descriptions of experiential or material grills they were uncertain (M = 42.11%; z = .98, p = .328). (Tully and Sharma 2018) depended on which condition they were in Taken together, these studies provide consistent and convincing and then evaluated the grill. A 2 (state) x 2 (product type) ANOVA evidence that sense of uncertainty leads people to less prefer experi- revealed predicted significant interaction between state and product ential product because experiential product is associated with lower 2 type (F(1, 222) = 4.38, p = .037, ηp = .02). When the product was psychological ownership compared to material product. experiential-oriented, compared to certainty condition (M = 5.56, SD = 1.49), uncertainty lead to a marginally more negative attitude to- REFERENCES ward the outdoor grill (M = 5.10, SD = 1.29, t(222) = 1.89, p = .060, Atasoy, Ozgun, and Carey K. Morewedge (2018), “Digital Goods d = .33 ). However, when the product was material-oriented, product Are Valued Less Than Physical Goods,” Journal of Consumer evaluation did not differ between certainty (M = 5.43, SD = 1.29) and Research, 44(6), 1343-1357. uncertainty condition (M = 5.69, SD = 1.07, t(222) = 1.07, p = .287, Caprariello, Peter A. and Harry T. Reis (2013), “To Do, to Have, d = .22 ). These findings provide direct evidence that uncertainty de- or to Share? Valuing Experiences Over Material Possessions creases experiential product preference rather than increases material Depends on the Involvement of Others,” Journal of product preference. Personality and Social Psychology, 104(2), 199-215.

Carter, Travis J. and Thomas Gilovich (2010), “The Relative Mittal, Chiraag, and Vladas Griskevicius (2014), “Sense of Relativity of Material and Experiential Purchases,” Journal of Control Under Uncertainty Depends on People’s Childhood Personality and Social Psychology, 98(1), 146–159. Environment: A Life History Theory Approach,” Journal of Collins, Nancy L. and Stephen J. Read (1990), “Adult Attachment, Personality and Social Psychology, 107(4), 621-637. Working Models, and Relationship Quality in Dating Mittal, Sarah, Sommer Kapitan, and David H. Silvera (2019), Couples,” Journal of Personality and Social Psychology, 58 “Go big or go home: Risk seeking for experiential choices,” (4), 644–663. Journal of Consumer Behaviour, 18(2), 97-108. Cunningham, Shelia J., David J. Turk, Lynda M. Macdonald, and Morewedge, Carey K. and Colleen E. Giblin (2015), “Explanations C. Neil Macrae (2008), “Yours or Mine? Ownership and of the Endowment Effect: An Integrative Review,” Trends in Memory,” Consciousness and Cognition, 17 (1), 312–318. Cognitive Sciences, 19(6), 339-48. Dommer, Sarah L. and Vanitha Swaminathan (2013), “Explaining Oleson, Kathryn C., Kirsten M. Poehlmann, John H. Yost, Molly the Endowment Effect Through Ownership: The Role of E. Lynch, and Robert M. Arkin (2000), “Subjective Over Identity, Gender, and Self-Threat,” Journal of Consumer Achievement: Individual Differences in Self-Doubt and Research, 39(5), 1034–1050. Concern with Performance,” Journal of Personality, 68 (3), Elstak, Mirdita N. , Mamta Bhatt, Cees B. M Van Riel, et al. (2015), 491–524. “Organizational Identification during a Merger: The Role of Paundra, Joshua, Laurens Rook, Jan Van Dalen, and Wolfgan Self‐Enhancement and Uncertainty Reduction Motives during Ketter (2017), “Preferences for Car Sharing Services: Effects a Major Organizational Change,” Journal of Management of Instrumental Attributes and Psychological Ownership,” Studies, 52(1), 32-62. Journal of Environmental Psychology, 53, 121-130. Faraji-Rad, Ali, and Michel Tuan Pham (2017), “Uncertainty Pierce, Jon L., Tatiana Kostova, and Kurt T. Dirks (2001), “Toward Increases the Reliance on Affect in Decisions,”Journal of A Theory of Psychological Ownership in Organizations,” Consumer Research, 44(1), 1-21. Academy of Management Review, 26, 298–310. Fenigstein, Allen, Michael F. Sheier, and Arnold H. Buss (1975), Reb, Jochen, and Terry Connolly (2007), “Possession, Feelings “Public and Private Self-Consciousness: Assessment and of Ownership and the Endowment Effect,”Judgment and Theory,” Journal of Consulting and Clinical Psychology, Decision Making, 2(2), 107-114. 43(4), 522–27. Rosenzweig, Emily and Thomas Gilovich (2012), “Buyer’s Gallo, Inigo, Lily Jampol, Alberto Rampullo, and Thomas Gilovich Remorse or Missed Opportunity?” Journal of Personality and (2018), “Surprise! The Positive Impact of Uncertainty on the Social Psychology, 102(2), 215–223. Evaluation of Experiential Purchases,” ACR North American Rucker, Derek D. and Adam D. Galinsky (2009), “Conspicuous Advances. Consumption versus Utilitarian Ideals: How Different Levels Gao, Leilei, S. Christian Wheeler, and Baba Shiv (2009), “The of Power Shape Consumer Behavior,” Journal of Experimental ‘Shaken Self’: Product Choices as a Means of Restoring Social Psychology, 45(3), 549–555. Self-View Confidence,” Journal of Consumer Research, 36(1), Savary, Jennifer, and Ravi Dhar (2020), “The Uncertain Self: How 29–38. Self-Concept Structure Affects Subscription Choice,”Journal Gilovich, Thomas, and Inigo Gallo (2020), “Consumers’ Pursuit of of Consumer Research, 46(5), 887-903. Material and Experiential Purchases: A Review”, Consumer Shen, Luxi, Ayelet Fishbach, and Christopher K. Hsee (2015), Psychology Review, 3(1), 20-33. “The Motivating-Uncertainty Effect: Uncertainty Increases Hogg, Michael A. (2007), “Uncertainty–Identity Theory,” Advances Resource Investment in the Process of Reward Pursuit,” in Experimental Social Psychology, 39, 69–126. Journal of Consumer Research, 41(5), 1301-1315. Howell, Ryan T., and Graham Hill (2009), “The Mediators Shen, Luxi, Christopher K. Hsee, and Joachim H. Talloen (2019), of Experiential Purchases: Determining the Impact of “The Fun and Function of Uncertainty: Uncertain Incentives Psychological Needs Satisfaction and Social Comparison,” Reinforce Repetition Decisions,” Journal of Consumer The Journal of Positive Psychology, 4(6), 511-522. Research, 46(1), 69-81. Howell, Ryan T., Paulina Pchelin, and Ravi Iyer (2012), “The Shu, Suzanne B. and Joann Peck (2011), “Psychological Ownership Preference for Experiences Over Possessions: Measurement and Affective Reaction: Emotional Attachment Process and Construct Validation of the Experiential Buying Tendency Variables and the Endowment Effect,”Journal of Consumer Scale,” The Journal of Positive Psychology, 7(1), 57-71. Psychology, 21(4), 439–452. Kagan, Jerome (1972), “Motives and Development,” Journal of Tully, Stephanie M. and Eesha Sharma (2018), “Context-Dependent Personality and Social Psychology, 22(1), 51–66. Drivers of Discretionary Debt Decisions: Explaining Kahneman, Daniel, Jack L. Knetsch, and Richard H. Thaler (1990), Willingness to Borrow for Experiential Purchases,” Journal of “Experimental Tests of the Endowment Effect and the Coase Consumer Research, 44(5), 960-973. Theorem,” Journal of political Economy, 98(6), 1325-1348. Tully, Stephanie M., Hal E. Hershfield, and Tom Meyvis (2015), Laurin, Kristin, Aaron C. Kay, and David A. Moscovitch (2008), “Seeking lasting enjoyment with limited money: Financial “On the Belief in God: Towards an Understanding of the constraints increase preference for material goods over Emotional Substrates of Compensatory Control,” Journal of experiences,” Journal of Consumer Research, 42(1), 59-75. Experimental Social Psychology, 44 (6), 1559–1562. van den Bos, Kees (2009), “Making Sense of Life: The Existential McGregor, Ian, Mark P. Zanna, John G. Holmes, and Steven J. Self Trying to Deal with Personal Uncertainty,” Psychological Spencer (2001), “Compensatory Conviction in the Face of Inquiry, 20(4), 197–217. Personal Uncertainty: Going to Extremes and Being Oneself,” Weiss, Liad and Gita V. Johar (2013), “Egocentric Categorization Journal of Personality and Social Psychology, 80(3), 472–488. and Product Judgment: Seeing Your Traits in What You Own (and Their Opposite in What You Don’t),” Journal of Consumer Research, 40(1), 185–201. Advances in Consumer Research (Volume 48) / 775

——— (2016), “Products as Self-Evaluation Standards: When Wittkowski, Joachim (2001), “The Construction of the Owned and Unowned Products Have Opposite Effects on Self- Multidimensional Orientation toward Dying and Death Judgment,” Journal of Consumer Research, 42(6), 915–930. Inventory,” Death Studies, 25(6), 479–495. Whillans, Ashley V., Aaron C. Weidman, and Elizabeth W. Dunn (2016), “Valuing Time Over Money Is Associated With Greater Happiness,” Social Psychological and Personality Science, 7(3), 213-222. Time Machine to the Future The Potentials and Risks of Using Virtual Reality to Stimulate Saving Behavior Andrea Weihrauch, University of Amsterdam, Netherlands Tobias Schlager, HEC Lausanne, Switzerland

EXTENDED ABSTRACT company and an actor with many years of industry experience, we Every consumer faces the risk to experience (unexpected) nega- designed a simulation of a negative live event, namely the temporar- tive life event (i.e., job loss, divorce). Such negative events often ily loss of employment. We used this negative life event as it 1) has come with financial consequences, and can significantly affect the the strongest negative impact on consumers financial situation (com- financial situation of consumers (McKay and Kempson 2003). The pared to e.g., parenthood, divorce, or bereavement), 2) is a negative best way to (financially) prepare would be to create a “safety buffer”. life event that many consumers experience at some point in their life, However, evidence suggest that many consumers are financially un- 3) can occur in different stages of one’s life. During the simulation, prepared (Macrae et al. 2017; Peetz and Buehler 2009). One of the consumers “walked” into an office, took a seat across from their boss, main reasons for this “lack of preparation” is that they are unable or and were informed that their work contract would be changed, and unmotivated to imagine their future (Gilbert and Ebert 2002). that they would experience a significant decrease in income for a One strategy to help consumers to prepare for negative life six-month period. events is to expose them to simulations of the future. Loewenstein In study 1, we exposed participants (n = 237, 48.1% female, Mage (1996) theorized that a vivid impression of a future situation can sub- = 19.98) to that stimulus in high immersion (VR) and low immersion stantially affect people’s current behavior. Exposing consumers to a video (video). We assessed their self-reported saving motivation, and simulation of a future event should enable them to better imagine its an incentive-conform measure of saving behavior. We also tested an occurrence, and may eventually motivate them to (financially) pre- intervention to provide some evidence of the suggested process of pare for that event. Traditionally, firms used videos to simulate those activated cognitive defenses, and included a negatively-framed ver- events. However, the development of new technologies provides dif- sion of the life event, as well as a positively-framed version. ferent opportunities to immerse people. There are many examples for In study 2 (n = 333, 52.9% female, Mage = 20.18), we replicated VR use to simulate negative life events (i.e., an UK insurance com- the procedure of Study 1 and further included an additional condi- pany exposes consumers to a house fire to encourage the purchase of tion: a negative scenario (both in video and VR) with an intervention an insurance). aimed to remove possible cognitive defenses. As argued, participants Initial evidence suggests that Virtual Reality (VR) can indeed likely use arguments such as “this will never happen to me”. To effectively simulate future events and encourage people to adapt mitigate this, we asked participants after exposure to the negative their behavior. For instance, Hershfield et al. (2011) showed that ex- scenario to provide one or more reasons why such a temporary un- posing people to an age-progressed avatar of their future selves in employment is likely to occur to them. VR increases their willingness to save money for retirement. Our results show that the prevailing notion “the more immer- While such findings shed a positive light on the ability of using sion is always better” might not always be true. For negative life VR simulations to motivate and enable saving behavior, using VR event, less immersive technologies than VR (such as video simula- could have downsides: Not only is creating VR stimuli costly, but tions) can be more efficient to stimulate saving behavior. The un- the high level of immersion such simulations produce could lead to derlying reason is that a too realistic simulation of a negative event some unexpected risks. Increased immersion entails that the experi- triggers cognitive defenses in consumers. Should the use of VR still ence is more confronting. For instance, experiencing an extremely be considered, accompanying the simulation with a task to remove realistic simulation of being laid off, is likely perceived to be painful these cognitive defenses is advised to attenuate the negative effect of and uncomfortable. Imaging a very negative life event might trigger immersion on saving motivation. This seems to be a more effective cognitive defenses (Miceli and Castelfranchi 1998), with the result strategy than framing the negative event as more positive. that consumers underestimate the likelihood of the event or deny This manuscript makes important contributions to 1) the litera- its impact (Parfit 1984; Taylor 1991). Thoughts such as “this would ture examining the potential of VR to change consumer behavior; 2) never happen to me” could in the worst case even be demotivating. research aiming to better (financially) prepare consumers for their This leads us to the following three hypotheses: future; 3) research on positive versus negative information, and 4) literature on cognitive defenses. In sum, it supports the growing lit- Hypothesis 1: Exposure to more immersive simulations (VR vs. erature on the intersection of financial, technological and behavioral video) of negative future events decreases saving sciences. motivation. The insights of our studies should inform both companies and public policy makers, and ultimately consumers. Our results provide Hypothesis 2: Exposure to more immersive simulations (VR vs. concrete guidelines in designing the best instrument to better prepare video) of negative future events triggers cogni- customers for the financial implications of (negative) life events. tive defenses, which in turn lead to a decrease in saving motivation. REFERENCES Hypothesis 3: Framing the negative life event as more positive Gilbert, D. T. and J.E. Ebert (2002), “Decisions and revisions: The can attenuate the negative effect of immersion affective forecasting of changeable outcomes,”Journal of on saving motivation. personality and social psychology, 82(4), 503. Hershfield, H. E. (2011), “Future self‐continuity: How conceptions We conducted a series of stimuli pretests and two laboratory of the future self -transform intertemporal choice,” Annals of experiments to examine whether and how using VR to simulate life the New York Academy of Sciences, 1235(1), 30-43. events affects saving behavior. Together with a professional VR

Hershfield, H. E., D.G. Goldstein, W.F. Sharpe, J. Fox, L. Yeykelis, Miceli, M. and C. Castelfranchi (1998), “How to silence one’s L.L., Carstensen, and J.N. Bailenson (2011), “Increasing conscience: Cognitive defenses against the feeling of guilt,” saving behavior through age-progressed renderings of the Journal for the Theory of Social Behaviour, 28(3), 287-318. future self,” Journal of Marketing Research, 48(SPL), 23-37. Parfit, D. (1984). Reasons and persons. Oxford: Clarendon Press. Loewenstein, G. (1996), “Out of Control: Visceral Influences on Peetz, J. and R. Buehler (2009), “Is there a budget fallacy? The Behavior,” Organizational Behavior and Human Decision role of savings goals in the prediction of personal spending,” Processes, 65(3), 272–92. Personality and Social Psychology Bulletin, 35(12), 1579-91. Macrae, C. N., J.P. Mitchell, M. Golubickis, N.S. Ho, R. Taylor, S.E. (1991), “Asymmetrical effects of positive and Sherlock, R., Parlongo, O.C.M. Simpson, and B.M Christian negative events: the mobilization-minimization hypothesis,” (2017), “Saving for your future self: The role of imaginary Psychological bulletin, 110(1), 67. experiences,” Self and Identity, 16(4), 384-98. McKay, S. and E. Kempson (2003), Savings and life events. Leeds: Corporate Document Services. The Environmental Consequences of People’s Moods: Positive Moods and Disposal Behavior Sarah Whitley, University of Georgia, USA Remi Trudel, Boston University, USA Weston Baxter, Imperial College London, UK

EXTENDED ABSTRACT Norli 2007), and thus we examine the impact of sporting event out- Recycling presents a real way to combat climate change, but comes on a community highly invested in the outcomes of a specific can something as simple as your mood influence your likelihood to sport: college basketball. Using collegiate basketball game results recycle? We suggest that positive moods increase one’s likelihood to as a proxy for mood, we examine the effect of historical basketball recycle with behavioral evidence observed both in the lab and on an game outcomes on observed college campus recycling collections. aggregate level in the community. Every day people are faced with disposal decisions: whether to Method keep, trash, or recycle an object. The US EPA concludes that recycling We collected daily recycling collection amounts (lbs), span- provides a real way to help address climate change. Unfortunately, ning 2012-2016, from four U.S. universities whose men’s basket- much of what ends up in landfills could have been recycled (US EPA ball teams were ranked in the top 30 of the 2016 pre-season NCAA 2012). Prior research has shown a number of cognitive factors that Division 1 Men’s College Basketball Rankings. Looking at the first influence one’s decision to trash versus recycle (product form, Tru- recycling collection date following each historical basketball game del and Argo 2013; identity-relevance, Trudel, Argo, and Meng et al. from 2012-2016, school-specific campus recycling collections were 2016a; proximity, Catlin and Wang 2013), but what about affective regressed on basketball game outcomes (wins vs. losses/ties) to test factors? Research has seldom explored affective influences on sus- for any relationship between mood and recycling behavior. tainable behaviors. Ad/image-induced negative emotions (sadness/ Data and Variables. Each university had a different recycling guilt) from environmental campaigns have been shown to increase collection schedule (e.g., Monday/Thursday, Tuesday/Friday, etc.), environmental concern (Schwartz and Loewenstein 2017; Sevillano, resulting in each collection date (i.e., the day recycling was collected Aragonés, and Schultz 2007), other work has shown that people are from the bin and weighed by the city) representing the amount col- waste averse and consume less resources to avoid negative emotions lected over a varied period of days. To create a dependent variable from wasting (Sun and Trudel 2017), but research has yet to explore comparable across schools at the collection date level, we calculated the influence of incidental affect on disposal decisions. In this work, the average daily recycling amount for each collection date by sim- we examine whether affect, specifically mood, influences consumers’ ply dividing the total weight (lbs) of recycling content collected on disposal decisions: the choice to trash vs. recycle an object. each school-specific collection date by the number of days content Across four studies, we examine if and in what manner mood accumulated in the recycling bin (i.e., # of days since last collection may influence recycling behavior. In our first two studies, using date). This resulted in our dependent variable of average daily recy- historical recycling collection data (i.e., community refuse pickup cling (lbs/day) for each collection date. Our final dataset included amounts), we test for evidence of a relationship between macroeco- 373 collection day observations. nomic events that influence mood (e.g., sporting event outcomes, Historical game outcomes (win, loss, tie) were matched to the weather) and observed/measured recycling behavior. Next, we ex- first recycling collection date immediately following the date of the plore the causal direction of this hypothesized relationship in the lab game. In some cases, more than one basketball game occurred before by observing participants’ decision to recycle versus trash a piece of a single recycling collection date; to account for this, we calculated scrap paper after being induced into a positive, negative, or neutral a “Majority Wins” independent variable characterizing whether the mood state. Lastly, we propose that one of the mechanisms driving majority of games occurring before the collection date were wins the effect of mood on recycling behavior is consumers’ ability to cor- (Majority Wins=1) or losses/neutral (Majority Wins=0). rectly categorize products at the point of disposal decision (i.e., deciding if the object should go into the recycle vs. trash bin), as posi- Results and Discussion tive moods engender more flexible cognitive categorizations (Isen We used a linear mixed effects model (ML estimation) with and Daubman 1986). game wins (Majority Wins) as the fixed effect predictor and both school and season (i.e., 2012-2013) as random effects to analyze STUDY 1 their association with daily average recycling. Basketball game wins Recycling is a highly observable behavior in that it is quantified were associated with a significant increase in average daily recycling on both an individual and aggregate level. Consumers in charge of (b=1,096.14, SE=523.25, t=2.10, p=.037), suggesting a positive cor- their own recycling collections deposit it at a recycling facility and relation between mood and recycling behavior. pay fees according to the weight of the recycled content, while community STUDY 2 collections do this on an aggregate scale, keeping records of We next sought to replicate the results of the prior study us- daily, weekly, and monthly recycled content. Provided this wide-scale ing another historical data analysis employing a different macro- quantification of recycling behavior, we first examined whether it is level proxy for mood: weather. Weather has been shown to influence possible to observe a relationship between mood and observed re- people’s moods, with good weather bringing about positive moods cycling collection amounts. While it is not possible to measure each and inclement weather prompting negative moods (Hirshleifer and individual’s mood when considering wide- scale community collec- Shumway 2003; Schwarz and Clore 1983). Using precipitation as a tion observations, there are certain types of events that can impact proxy for negative mood, we examine the effect of historical precipi- the general mood of a community. Prior research suggests that sport- tation levels on neighborhood recycling collection amounts. Lastly, ing event outcomes significantly impact mood (Edmans, Garcia, and unlike Study 1, we were able to obtain data about trash collection

Advances in Consumer Research 778 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 779 amounts from our source and thus incorporate the effects of trashing (n=74), or neutral (n=69) mood and their disposal behavior of a piece behavior in our analysis of recycling. of scrap paper was recorded. Procedure. Participants sat at a lab computer and were provided Method with a small piece of paper (3x1.5 inch) with a 3-digit code. This One year (4/1/2015-3/31/2016) of daily curbside trash and re- small paper was used because although paper is recyclable, small cycling collection amounts (metric tons) were obtained for 14 neigh- prices of paper have been shown in prior research to often be trashed borhoods in a London suburb, resulting in a total of 2,831 curbside rather than recycled (Trudel & Argo 2013), providing a refuse ob- collection date observations across 203 individual collection dates. ject that is capable of being either recycled or trashed and for which These daily collection amounts were regressed upon daily precipita- people show inherent variance in their disposal behavior. Depending tion (mm) data for the London area for the same dates. on their assigned condition, participants were asked to read a posi- Data and Variables. Since the neighborhoods operated accord- tive, negative, or neutral online news article (Fishbach and Labroo ing to different collection schedules (e.g., Monday/Thursday, Tues- 2007) and then asked to reflect on their own experiences and write day/Friday, etc.), in the same vein as Study 1, a dependent variable a short news-like article detailing one of the saddest, happiest, or a of average daily recycling (metric tons/day) was created to allow typical day in their life, respectively. Upon finishing the mood ma- for comparisons across neighborhoods by collection date. We cre- nipulation task, participants were told to “dispose of their paper on ated an average daily trash (metric tons/day) variable in the same the way out.” Directly outside the lab were two disposal bins, labeled manner. Trash served as a covariate in our analysis to evaluate the “Trash” and “Recycle.” Data was collected at the end of the day by a independent effect of precipitation on recycling absent any effects research assistant who counted the number of pieces of paper in the from trash. Daily precipitation (mm) was averaged over the days trash versus recycling bins. The codes on each disposed paper was contained within each collection period (i.e., period between prior used to link participant disposal behavior with assigned condition. collection date and current collection date – days over which content accumulated in the bin) and served as our independent variable of Results and Discussion average daily precipitation (mm/day). Twenty-two participants left with the paper, leaving 212 observations (59.4% female; Mage=19.84 years). A binary logistic re- Results and Discussion gression (Trash=0, Recycle=1) revealed that mood influenced recy- We used a linear mixed effects model (ML estimation) with av- cling (χ2(2)=9.83, p=.007). Participants recycled significantly more erage daily precipitation as a fixed effect predictor and neighborhood when in a positive mood (44.9%) than in a negative (25.7%; β=.86, as a random effect. We also controlled for average daily trash col- p=.017) or neutral mood (21.7%; β=1.08, p=.004; See Figure). Dis- lected, holiday occurrence, and collection pickup schedule (Monday/ posal behavior between the negative and neutral mood conditions Thursday vs. Tuesday/Friday) by adding them as fixed effects in the did not differ (β=.22, p=.581). These results provide causal support model. Our analysis revealed that increases in precipitation (i.e., a for the positive relationship between mood and recycling and sug- proxy for increased negative mood) was associated with a signifi- gest this effect is driven by a positive mood increasing recycling, cant decrease in the amount of recycled content collected (b=-.001, rather than a negative mood discouraging recycling. SE=.0003, t=-3.36, p=.001). This suggest that negative mood has a negative impact on recycling, supporting the same conclusion from STUDY 4 Study 1 that mood is positively correlated with recycling (i.e., recy- Various affective, motivational, and cognitive mechanisms may cling moves in the same direction as mood). underlie the observed effect of mood on recycling behavior. Since re- While the covariate of average daily trash had a significant pos- cycling and trashing are associated with positive and negative affect, itive relationship with recycling (b=.214, SE=.006, t=34.88, p<.001), respectively (Sun and Trudel 2017), people in positive moods may a separate analysis with trash as the dependent variable revealed that be more likely to recycle in order to maintain their positive mood precipitation had a negative effect on average daily trash collected state (Wegener and Petty 1994). (b=.004, SE=.001, t=-4.75, p<.001). Together, these results suggest Additionally, as recycling is often a more effortful task than that the negative association of precipitation and recycling holds trashing, since positive moods have been shown to increase effort over and above precipitation’s influence on trash collected. (Erez and Isen 2002; Fishbach and Labroo 2007), this suggests peo- Additionally, an analysis using standardized variables con- ple with positive moods may be more likely to recycle. We suggest firmed that the negative effect of precipitation on recycling (β=- that despite these affective and motivational mechanisms, we believe .074) was greater in magnitude than that of precipitation on trash part of the reason positive mood increases one’s likelihood to recycle (β=-.064). is cognitive. The decision to recycle, itself, is a cognitive one: when faced with an article of refuse, a consumer must decide to either STUDY 3 recycle or trash the object. While Studies 1 and 2 provide correlational evidence of the re- However, consumers often erroneously categorize recyclable lationship between mood and recycling behavior in the real-world, in items as trash and throw them in the trash bin rather than the re- this study we use a controlled lab environment to show that mood di- cycling bin (Trudel, Argo, and Meng 2016b). Positive moods lead rectly causes changes in recycling. Additionally, it is not clear if the to more flexible categorizations (Isen and Daubman 1986) and we positive correlation between mood and recycling is due to positive suggest that consumers in positive moods will be more likely to cat- moods increasing recycling or negative moods decreasing recycling. egorize fringe exemplars (e.g., smashed cans, scrap paper; Trudel By adding a neutral mood condition, we can disentangle the effect of and Argo 2013) correctly as something to be recycled rather than positive versus negative moods on recycling. trashed. In this study, we evaluate whether mood influences how Method consumers categorize fringe exemplars of recyclable products (i.e., Participants and design. Undergraduate participants (N=234) scrap paper), as correct categorization should theoretically drive cor- were randomly induced with either a positive (n=69), negative rect disposal. 780 / The Environmental Consequences of People’s Moods: Positive Moods and Disposal Behavior Method Catlin, J. R., & Wang, Y. (2013). Recycling gone bad: when the Participants and design. Undergraduate participants (N=166) option to recycle increases resource consumption. Journal of were randomly induced with either a positive (n=56), negative Consumers Psychology, 23(1), 122-127. (n=55), or neutral (n=55) mood and asked about their categorization Edmans, A., Garcia, D., & Norli, O. (2007). Sports sentiment and of a piece of scrap paper. stock returns. The Journal of Finance, 62(4), 1967-1998. Procedure. Analogous to Study 3, participants sat at a lab com- Erez, A., & Isen, A. M. (2002). The influence of positive affect on puter and were provided with a small piece of paper (3 x 1.5 in) the components of expectancy motivation. Journal of Applied with a 3-digit code before being induced with a positive, negative, or Psychology, 87(6), 1055-1067. neutral mood. The same mood manipulation from Study 3 was used. Fishbach, A., & Labroo, A. A. (2007). Be better or be merry: how Participants were then asked to focus on piece of paper and indicate mood affects self- control.Journal of Personality and Social “how much is the paper like trash?” (1=Not at all like Trash; 5 = Psychology, 93(2), 158-173. Exactly like Trash), a measure used to investigate the categorization Goldstein, N. J., Cialdini, R. B., & Griskevicius, V. (2008). A of recyclable material (Trudel et al. 2016b). room with a viewpoint: using social norms to motivate environmental conservation in hotels. Journal of Consumer Results and Discussion Research, 35(3), 472-482. ANOVA results revealed differences in categorization between Halvorsen, B. (2008). Effects of norms and opportunity cost of time mood conditions (F(2, 163) = 4.78, p = .010). Specifically, partici- on household recycling. Land Economics, 84(3), 501-516. pants in a positive mood (M = 2.59, SD = 1.33) were less likely to Hirshleifer, D., & Shumway, T. (2003). Good day sunshine: stock rate the scrap paper as typical of trash compared to participants in a returns and the weather. The Journal of Finance, 58(3), 1009- negative mood (M = 3.27, SD = 1.18; p = .003) or neutral mood (M= 1032. 3.04, SD = 1.02, p = .048). There was no difference in categorization Isen, A. M., & Daubman, K. A. (1986). The influence of affect on of the scrap paper between participants in the negative and neutral categorization. Journal of Personality and Social Psychology, mood conditions (p = .297). Supporting our claim, these results show 47(6), 1206-1217. that people in positive moods are more integrative in their categori- Jenkins, R.R., Martinez, S.A., Palmer, K., & Podolsky, M. zation and less likely to erroneously categorize a small piece of paper J. (2003). The determinants of household recycling: a as trash. This proper categorization (as not trash) should theoreti- material-specific analysis of recycling program features cally lead to proper subsequent disposal (i.e., recycling), suggesting and unit pricing. Journal of Environmental Economics and that positive moods should drive an increase in one’s likelihood to Management, 45(2), 294-318. recycle. Reschovsky, J. D., & Stone, S. E. (1994). Market incentives to encourage household waste recycling: paying for what you GENERAL DISCUSSION throw away. Journal of Policy Analysis and Management, Overall, we show that mood significantly impacts recycling 13(1), 120-139. behavior, both on an individual and aggregate level, with posi- Salovey, P., & Rosenhan, D. L. (1989). Mood states and prosocial tive moods increasing recycling both among experimental induced behavior. Handbook of Social Psychophysiology, 371-391. moods and naturally occurring mood states from macro-level events. Schwartz, D., & Loewenstein, G. (2017). The chill of the moment: We find that positive moods make consumers more likely to recycle emotions and proenvironmental behavior. Journal of Public a recyclable object (Study 3), and believe this greater likelihood to Policy & Marketing, 36(2), 255-268. properly recycle objects (rather than trash them) underlie the positive Schwarz, N., & Clore, G. L. (1983). Mood, misattribution, and relationship between mood and recycled content observed in Studies judgments of well-being: Informative and directive functions 1 and 2 (i.e., recycled content can only increase by more recycling- of affective states.Journal of Personality and Social capable content being placed in the bin). While affect regulation and Psychology, 45(3), 513-523. motivational mechanisms may certainly play a role in the impact Sevillano, V., Aragonés, J. I., & Schultz, P. W. (2007). Perspective of positive moods on recycling, we suggest that reduced improper taking, environmental concern, and the moderating role of categorization of recyclable products as trash under a positive mood dispositional empathy. Environment and Behavior, 39(5), (Study 4) underlies the observed increase in recycling associated 685-705. with positive mood. Sun, M., & Trudel, R. (2017). The effect of recycling versus This work adds to prior research noting the mixed effects of trashing on consumption: theory and experimental evidence. mood on traditional prosocial behaviors focused on helping oth- Journal of Marketing Research, 54(2), 293-305. ers (e.g., donating, volunteering, helping with a task; Salovey and Trudel, R., & Argo, J. J. (2013). The effect of product size and Rosenhan 1989; Carlson, Charlin, and Miller 1988), by exploring the form distortion on consumer recycling behavior. Journal of impact of mood on a sustainable behavior (i.e., recycling), a decid- Consumer Research, 40(4), 632-643. edly both self- and other-helping behavior. Lastly, in addition to the Trudel, R., Argo, J. J., & Meng, M. D. (2016a). The recycled self: more cognitive (Jenkins et al. 2003; Reschovsky and Stone 1994; consumers’ disposal decisions of identity-linked products. White, MacDonnell, and Dahl 2011), social (Goldstein, Cialdini, and Journal of Consumer Research, 43(2), 246-264. Griskevicius 2008) and convenience-based (Halvorsen 2008) ante- Trudel, R., Argo, J. J., & Meng, M. D. (2016b). Trash or recycle? cedents of disposal decisions, we find that affect, specifically inci- How product distortion leads to categorization error during dental mood influences one’s decision to recycle. disposal. Environment and Behavior, 48(7), 966-985. US Environmental Protection Agency (2012). Municipal Solid REFERENCES Waste in the United States: 2012 Facts and Figures,” https:// Carlson, M., Charlin, V., & Miller, N. (1988). Positive mood archive.epa.gov/epawaste/nonhaz/municipal/web/pdf/2012_ and helping behavior: a test of six hypotheses. Journal of msw_fs.pdf. Personality and Social Psychology, 55(2), 211-229. Advances in Consumer Research (Volume 48) / 781

Wegener, D. T., & Petty, R. E. (1994). Mood management across White, K. M., MacDonnell, R., & Dahl, D.W. (2011). It’s the affective states: the hedonic contingency hypothesis.Journal mind-set that matters: the role of construal level and message of Personality and Social Psychology, 66(6), 1034-048. framing in influencing consumer efficacy and conservation behaviors. Journal of Marketing Research, 48(3), 471-485. Examining the Ongoing Process of ‘Hegemonic Legitimacy’ of Meat: a Discursive Institutional Perspective on Market Maintenance Lucie Wiart, Université de Lille, France Deirdre Shaw, University of Glasgow, Scotland, UK Nil Özçaglar-Toulouse, Université de Lille, France

EXTENDED ABSTRACT We outline our findings across three key time periods repre- In marketing, the study of how markets evolve through different senting points of discursive struggle between opposing ideologies in dynamics or logics has been well developed (Ertimur and Coskun- relation to meat (non)consumption and resulting challenges to legiti- er-Balli, 2015). In most of these studies, institutional perspectives macy maintenance: have been adopted for examining processes of market legitimacy creation, evolution, and maintenance (see Humphreys and LaTour, 1944-1980: Meat as an industrial product 2013; Humphreys, 2010; Kates, 2004; Scaraboto and Fischer, 2012). During this period we observe the creation, evolution and main- However, the process of legitimacy maintenance has been neglected. tenance of meat as a legitimate industrial farm product articulated by Also, while the process of legitimacy across the three areas of mar- commodity and scientific signifiers. Counter-hegemonic resistance is ket creation, evolution and maintenance has been conceptualized, the lacking and economic discourse is dominant, indeed, meat consump- relationships between these processes have not been empirically ex- tion is considered vital to support economic development. amined. Such an omission is important as markets are not static, thus, we would expect movement between and within these concepts. Fur- 1980-2019: Safe to eat? thermore, market maintenance requires ongoing effort over time, The period from 1980 represented a ‘period of dislocation’ suggesting a process of evolution which may require elements of (Laclau and Mouffe, 1985) where two major events, namely, ‘hor- market creation. Therefore, in highlighting the need to examine mar- monal veal’ (1980) and ‘mad cow disease’ (1996) resulted in legiti- ket maintenance we also highlight the need to examine the process of macy threats in relation to the conceptualization of farm animals as legitimacy as a whole to understand the role of market maintenance objectified and industrialized. To maintain market legitimacy we ob- within this process not in isolation from it. serve the creation of a new discourse of ‘Happy Meat’, countering To achieve this we focus on discursive aspects of legitimacy to animal suffering while maintaining the position of meat as a food consider power relations and struggles (Vaara and al, 2006; Vaara source. Accordingly, the subject position of ‘Compassionate Carni- and Tienari, 2008; Vaara and Monin, 2010; Johnson, 2017). Indeed, vore’ is created affording a humane identity to meat consumers. some research understands legitimacy as a process of discursive struggle through which discourses both attempt to either resist or 2000-2019: Climate crisis reproduce and reinforce legitimacy (Laine and Vaara, 2017). We The mainstreaming of discourse around climate change position specifically draw on Critical Discourse Studies (CDS) considering overconsumption of meat as unsustainable and unhealthy. From this discourse as a practice of constructing the social world and mean- we observe the creation of ‘flexitarian’ as a new moral and dominant ings (Fairclough and Wodak, 1997; Jorgensens and Phillips, 2002). subject-position enabling the construction of a responsible identity. From this perspective, the market could be considered as an entirely Here the market created ‘meat substitutes’, offering a ‘fake meat’ discursive formation: it provides concepts, objects, and subject posi- product. In doing so, the concept of farm animal is gradually disar- tions (Fairclough, 1992; Phillips and Hardy, 1997). The three pro- ticulated from the concept of meat, the new meat- substitute market cesses of market creation, evolution and maintenance would neces- completely erases the animal signifier. Through this new market, sarily be linked to the discursive constructions of concepts, objects negative signifiers associated with farm animals and livestock sec- and subject positions. The context of discursive struggles over these tors are, therefore, not co-opted, rather erased. Such a creation, main- three differing constructs will shed light on the market as dynamic tains the position of ‘meat’ as food while shifting what the object across the processes of legitimacy. Such an approach will reveal how of meat is in terms of what it comprises, either as plant based meat market as a discourse is (re)created and evolving and the resultant substitutes or lab meat grown from animal cells. relationship and impact on market maintenance. Our findings reveal that the process of legitimacy is dynamic To examine these discursive processes, we focus on the context across market creation, evolution and maintenance and that to main- of meat consumption, which has a long history of critique (Spencer, tain legitimacy requires engagement with creation and evolution in 2000) right up to present day links between meat consumption and a non-linear circling between concepts. In this process we see dis- climate change (Springmann et al, 2018). Although consumers have cursive struggles forcing market creation and the evolution of mean- been converting to vegan, vegetarian and flexitarian diets in increasing and relevance. Thus, to maintain hegemonic legitimacy necessi- ing numbers, demand for meat is increasing globally (Godfray et al, ties a continuously crafted discourse embedded within cultural and 2018). In this paper we ask how a food source that has been chal- historical contexts, constantly arranging and recreating legitimacy lenged over a significant time period has managed to maintain its through new articulations or reinforced processes of differentiation legitimacy. and chains of opposition. We draw on a ‘discourse-historical approach’ (Wodak, 2001; We also develop a post-structural institutional understanding of Reisgl, 2018; Reisgl and Wodak, 2015) that puts emphasis on histori- individual resistance and agency toward a hegemonic market that cal anchoring and how discourses are subject to diachronic change. could direct future research. Indeed where discourse is studied in We conducted a historical discourse analysis from 1944 to 2019, research on market evolution, it is constructed by researchers as gathering media articles on meat consumption from two French an instrument for actors to impact markets (see e.g., Giesler, 2012; newspapers. The specific perspective of Laclau and Mouffe (1985) Thompson, 2004; Kates, 2004; Press et al, 2014) or conceptualized regarding hegemony and discourses is used to analyze the data and as a reflector of underlying normative change (see Humphreys, for developing theoretical insights.

2010). This ignores the role of discourse as a vehicle for change Jorgensens M. and Phillips, L, (2002), Discourse analysis as in itself and as constitutive of social reality (Larsson, 2019). Here, Theory and Method, Sage Publications, London we find that individual actors convey either dominant or alternative Kates, S. (2004), The Dynamics of Brand Legitimacy: An discourse through performing subject positions and could, therefore, Interpretive Study in the Gay Men’s Community, Journal of be seen as instruments for market maintenance. Consumer Research, 31-2, 455-464 Laclau, E. and Mouffe, C. (1985),Hegemony and Socialist REFERENCES Strategy, Verso, London Ben Slimane, K., Chaney, D., Humphreys, A. and Leca, B. (2019), Reisigl, M. (2018), The Discourse-Historical Approach, in Bringing institutional theory to marketing: Taking stock and Flowerdew, J. and Richardson, J.R, (2018), The Routledge future research directions, Journal of Business Research, 389- Handbook of Critical Discourse Studies, Routledge, London 394 and New York. Ertimur, B., Coskuner-Balli, G. (2015) ‘Navigating the Institutional Reisigl, M. and Wodak, R. (2016). The discourse-historical Logics of Markets: Implications for Strategic Brand approach. In R. Wodak and M. Meyer (eds.), Methods of Management’, Journal of Marketing, 79(2): 40–61. critical discourse studies, 3rd edn, 23–61. London, Thousand Fairclough (N.) & Wodak (R.). 1997 Critical discourse analysis. In: Oaks, CA, New Delhi: Sage. van Dijk (T.). Discourse as Social Interaction. London: Sage. Scaraboto, D. and E. Fischer (2013), Frustrated Fatshionistats: Flowerdew, J. and Richardson, J.R, (2018), The Routledge An Institutional Theory Perspective on Consumer Quests for Handbook of Critical Discourse Studies, Routledge, London Greater Choice in Mainstream Markets, Journal of Consumer and New York. Research, 39-6, 1234-1257 Giesler, M. (2012) ‘How Doppelgänger Brand Images Influence the Spencer, C. (1993) The Heretic’s Feast, A history of vegetarianism, Market Creation Process: Longitudinal Insights from the Rise Fourth Estate, London Springmann, M., Clark, M., Mason- of Botox Cosmetic’, Journal of Marketing, 76(November): D’Croz, D. et al. (2018) Options for keeping the food system 56–68 within environmental limits. Nature 562, 519–525 Godfray, C.J., Aveyard, P. Garnett, T., Hall, W. J., Key, T.J., Vaara, E., Tienari, J., & Laurila, J. (2006), Pulp and paper Lorimer, J., Pierrehumbert, R.T., Scarborough, P., Springmann, fiction: On the discursive legitimation of global industrial M. and Jebb, S.A. (2018), Meat consumption, health and the restructuring. Organization Studies, 27: 789–810. environment, Science, 361: 6399 Vaara, Eero, and Janne Tienari. (2008). “A Discursive Perspective Humphreys, A. (2010). Semiotic Structure and the Legitimation on Legitimation Strategies in Multinational Corporations.” of Consumption Practices: The Case of Casino Gambling, Academy of Management Review 33 (4): 985–993. Journal of Consumer Research, 37(3) Vaara, Eero. (2014), “Struggles over Legitimacy in the Eurozone Humphreys, A. and LaTour, K. (2013). Framing the game: Crisis: Discursive Legitimation Strategies and their Ideological Assessing the impact of cultural representations on consumer Underpinnings.” Discourse & Society 25 (4): 500–518. perceptions of legitimacy, Journal of Consumer Research, Van Dijk, T. (1998). Ideology: A multidisciplinary approach. 40(4), 773-795 London: Sage Johnson, G.D, Thomas, K.D et Grier, S.A (2017), When the burger becomes halal: a critical discourse analysis of privilege and marketplace inclusion, Consumption Markets and Culture Counting Calories: How Calorie Perceptions and Estimates Diverge Kaitlin Woolley, Cornell University, USA Peggy Liu, University of Pittsburgh, USA

EXTENDED ABSTRACT ed both ordinal estimates (“very few” to “a lot of” calories burned) To manage calorie intake and achieve weight-related health or absolute estimates (numeric response). Ordinal estimates were goals, consumers often monitor calories expended and consumed greater for more intense, shorter exercises (e.g., 10-minutes of mod- (Campbell and Warren 2015; Jia et al. 2020). We identify a novel erate biking) than for less intense, longer exercises (e.g., 30-min- factor biasing calorie assessments and subsequent food choice: the utes of slow biking) (F(1, 395) = 52.75, p <.001), which reversed estimation mode consumers use to evaluate calories burned during when forming absolute estimates (F(1, 395) = 5.18, p = .023), despite exercise or contained in a food portion. equivalent objective calories burned across intensity/duration. Stud- We distinguish between two modes consumers use to estimate ies 2-6 examined this prediction within the food domain. caloric consumption and expenditure—forming ordinal calorie es- Study 2 assessed calories in a larger portion of healthier food timates (i.e., “very few” to “very many” calories) versus forming (200-calorie almond snack plate) and a smaller portion of less absolute calorie estimates (i.e., numeric response). We propose that healthy food (100-calorie chocolate covered almond snack plate). calorie estimates systematically diverge as a function of estimation Participants indicated either ordinal estimates (“very few” to “many” mode, with important consequences for consumers’ choices aimed calories) or absolute estimates (numeric response). Ordinal estimates at managing caloric intake. In addition, because both approaches are of almonds (vs. chocolate covered almonds) were significantly lower used interchangeably by researchers (Carels et al. 2006; Chernev, (t(121) = 6.50, p < .001); this pattern attenuated for absolute esti- 2011; Liu et al. 2019), some conclusions based on prior findings may mates (t(120) = .99, p = .325). This divergence affected actual food be moderated by estimation mode. choice: when choosing a snack plate with fewer calories to eat in the lab, more participants chose the larger portion of almonds when Type and Quantity as Inputs to Caloric Judgments forming ordinal (vs. absolute) estimates (59.8% vs. 46.3%; χ2(1, 243) Calorie estimates are a function of both what a stimulus is in = 4.48, p = .034). terms of its “type” (e.g., healthy vs. unhealthy food; easy vs. dif- Study 3 replicated this effect and demonstrated consequences ficult exercise) and its “quantity” (e.g., food portion size; exercise for consumer research by testing for moderation of an existing find- duration). Prior research has found that information on type is “pri- ing in the literature. First, we replicated a previous finding (study 2 mary” in that it is processed before information on quantity (Liu et in Chandon and Wansink 2007) that people forming absolute esti- al. 2019). mates estimate more calories in a Subway 12-inch turkey sub (larger, We propose that consumers differentially rely on type versus healthier) than in a McDonald’s cheeseburger (smaller, less healthy) quantity depending on whether they form ordinal or absolute esti- (t(136) = 1.90, p = .059). However, this pattern reversed when as- mates. First, we theorize that ordinal estimates involve categorical sessing ordinal calorie estimates, wherein people estimated fewer thinking driven by scale compatibility (Tversky et al. 1988; Fischer calories in the Subway’s sandwich (vs. McDonald’s cheeseburger) et al. 1999). That is, ordinal estimates anchor on type (more vs. less (t(139) = -9.27, p < .001). healthy food), and as this information is compatible with the ordinal Studies 4-6 demonstrate evidence for our underlying process. estimate response scale (few vs. many calories), information on type First, study 4 fully crossed type (almonds vs. chocolate covered al- is incorporated into their estimate (i.e., healthier food = fewer calo- monds) and quantity (small vs. large portion), demonstrating that ries; less healthy food = more calories). After spontaneously consid- ordinal estimates are more sensitive to type (F(1, 586) = 14.36, p < ering type, we suggest that ordinal estimates then fail to adjust much .001), whereas absolute estimates are more sensitive to quantity (F(1, for quantity. By contrast, we suggest that absolute estimates involve 586) = 28.12, p < .001). continuous thinking. Specifically, although type information is pri- Study 5 examined moderation within ordinal estimates by mary, this information is not compatible with the absolute estimate quantity primacy. Increasing the primacy of quantity information led response scale (i.e., number of calories), and thus we suggest that those forming ordinal and absolute estimates to both estimate more forming absolute estimates prompts people to consider other infor- calories in a larger portion of healthier food (vs. smaller portion of mation beyond type, leading them to incorporate quantity into their less healthy food). Lastly, study 6 examined moderation within ab- calorie estimate. solute estimates by intuitive (vs. deliberative) mindset (e.g., Wool- ley and Risen 2018). Adopting an intuitive mindset (quick, imme- When Will Ordinal and Absolute Calorie Estimates diate response) led those forming absolute and ordinal estimates to Reverse? both estimate fewer calories in a larger portion of healthier food (vs. If ordinal and absolute calorie estimates are differentially sen- smaller portion of less healthy food). sitive to type versus quantity information, these estimation modes Overall, six studies identified a novel bias in consumers’ calo- should lead to calorie estimate reversals when type and quantity di- rie estimates as a function of whether they made ordinal or absolute verge. That is, estimates should diverge when considering calories estimates. Depending on calorie estimate mode, we document pre- burned by difficult, shorter (vs. easier, longer) workouts, or when dictable divergences between ordinal and absolute calorie estimates considering calories contained in smaller, healthier food portions (vs. when evaluating stimuli that tradeoff on type and quantity. These larger, less healthy portions). We test this theory across six studies findings are of consequence to consumers, as whether consumers use (five pre-registered; see table 1a) by examining whether the effect ordinal versus absolute estimates can affect their food choices, and of estimation mode (ordinal vs. absolute) diverges as a function of for researchers, as the method of evaluating calories matters in terms type/quantity. of how consumers incorporate type and quantity. Study 1 examined divergence in ordinal versus absolute estimates of calories burned for different exercises. Participants indicat-

Advances in Consumer Research 784 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 785 REFERENCES Hsee, Christopher K. (1996), “The Evaluability Hypothesis: Campbell, Margaret C. and Caleb Warren (2015), “The Progress An Explanation for Preference Reversals between Joint Bias in Goal Pursuit: When One Step Forward Seems Larger and Separate Evaluations of Alternatives,” Organizational Than One Step Back,” Journal of Consumer Research, 41 (5), Behavior and Human Decision Processes, 67 (3), 247-57. 1316-31. Jia, Miaolei Liam, Xiuping Li, and Aradhna Krishna (2020), Carels, Robert A., Jessica Harper, and Krista Konrad (2006), “Contraction with Unpacking: When Unpacking Leads to “Qualitative Perceptions and Caloric Estimations of Lower Calorie Budgets,” Journal of Consumer Research, 46 Healthy and Unhealthy Foods by Behavioral Weight Loss (5), 853-70. Participants,” Appetite, 46 (2), 199-206. Kahneman, Daniel (2003), “A Psychological Perspective on CDC (2015), “Adult Obesity Facts,” https://www.cdc.gov/obesity/ Economics,” The American Economic Review, 93 (2), 162-68. data/adult.html. Liu, Peggy J., Kelly L. Haws, Karen Scherr, Joseph P. Redden, Chaiken, Shelly and Yaacov Trope (1999), Dual-Process Theories James R. Bettman, and Gavan J. Fitzsimons (2019), “The in Social Psychology, New York: Guilford Press. Primacy of ‘What’ over ‘How Much’: How Type and Chandon, Pierre and Brian Wansink (2007), “The Biasing Health Quantity Shape Healthiness Perceptions of Food Portions,” Halos of Fast‐Food Restaurant Health Claims: Lower Calorie Management Science, 65 (7), 3353-81. Estimates and Higher Side‐Dish Consumption Intentions,” Olshansky, S. Jay, Douglas J. Passaro, Ronald C. Hershow, Jennifer Journal of Consumer Research, 34 (3), 301-14. Layden, Bruce A. Carnes, Jacob Brody, Leonard Hayflick, Chernev, Alexander (2011), “The Dieter’s Paradox,” Journal of Robert N. Butler, David B. Allison, and David S. Ludwig Consumer Psychology, 21 (2), 178-83. (2005), “A Potential Decline in Life Expectancy in the United Chernev, Alexander and Pierre Chandon (2011), “Calorie States in the 21st Century,” New England Journal of Medicine, Estimation Biases in Consumer Choice,” in Leveraging 352 (11), 1138-45. Consumer Psychology for Effective Health Communication, Sloman, Steven (2014), “Two Systems of Reasoning: An Update,” ed. Rajeev Batra, Punam A. Keller and Victor Strecher, in Dual Process Theories of the Social Mind, ed. Jeffrey W. Armonk, NY: M.E. Sharpe, 104-21. Sherman, Bertram Gawronski and Yaacov Trope, New York: Hayes, Andrew F. and Kristopher J. Preacher (2014), “Statistical Guilford. Mediation Analysis with a Multicategorical Independent Variable,” British Journal of Mathematical and Statistical Psychology, 67 (3), 451-70. Say Cheese or Not? Ruomeng Wu, Western Kentucky University, USA Meng Liu, Independent Researcher Ze Wang, University of Central Florida, USA Ming-Shen Ho, Clemson University, USA

EXTENDED ABSTRACT displayer addressed by her first name had a bigger smile (p = .01) From Kraft’s creation of “the smile noodle” to Herta’s promo- but less often if the displayer addressed by her doctor title showed a tion of the “Herta makes a meal happy” campaign, smile is an es- bigger smile (p = .03). When participants spent less time (1 SD below sential marketing tactic for firms to better engage consumers. Indeed, the mean) viewing the ad, the smile type x rank interaction was not smile is widely believed to activate positive emotions and generate significant (p = .66). Additional analyses also revealed a significant favorable feedback in service and promotion contexts (e.g. Bugen- three-way interaction (processing time x rank x smile) on consum- tal 1986; Cheng, Mukhopadhyay, and Williams 2020; Mueser, Grau, ers’ purchase likelihood (p = .04). Similar patterns of results were Sussman, and Rosen 1984). observed. Smiles are not expressed in a vacuum. Consumers decode emo- Study 3 (N = 166) directly tests the mediation path and extends tional expressions based on a set of emotion decoding rules in the our investigation to emoji characters and brand rank. This study uses communication context. While burgeoning research has explored a 2 (smile type: slight vs. big) x 2 (rank: lower vs. higher) between- how emotional expressions may elicit consumers’ reactions in a subjects design. We operationalize the rank by brand names and con- variety of domains (Barger and Grandey 2006; Small and Verrochi firmed the effectiveness of this manipulation by pretests. Consistent 2009; Wang, Mao, Li, and Liu 2017), little is known about how the with our main proposition, the analysis of liking as a function of rank interplay of emotional expressions, displayer characteristics, and and smile type yielded an interaction effect(F = 10.02, p = .002). For contextual factors collectively impact consumer inference. This re- the lower-rank brand, participants liked the design better if the emoji search addresses this gap by highlighting the important roles of the had a big smile (p = .014). For the higher-rank brand, participants displayer’s social rank and consumers’ decision time in consumers’ liked the design more if the emoji had a slight smile (p = .051). More interpretation of smiles. importantly, a mediation analysis (Model 8; Hayes 2013) directly Study 1 (N = 294) provides initial evidence of the proposed supported our hypothesis about the mediating role of expectations role of social rank in the interpretation of smiles. This study uses a for emotion displays concerning rank and smile type. 2 (smile type: slight vs. big) x 2 (rank: lower vs. higher) between- Study 4 (N = 214) extends our proposed effects to a field setting. subjects design. To manipulate smile type, we showed participants a This study has a 3 (smile type: slight vs. big vs. no smile) x 2 (rank: photo featuring a male employee’s slight or big smile. To manipulate lower vs. higher) between-subjects design. Amazon Mechanic turk- rank, the job position of the employee was either an office assistant ers were approached after they finished a paid task. We then pleaded or a manager. The rank manipulation was confirmed by a pretest with them to participate in another task without extra payment. Following the same subject pool (p = .02). As expected, there was an interac- the same manipulations of smile type and rank in Study 3, we tracked tive effect between smile type and rank (F = 6.93, p = .01), such that the percentage of participants who accepted our request to finish the when the employee was an office assistant (lower rank), participants voluntary survey. In particular, when we introduced ourselves by a liked him more if he had a big rather than slight smile (p = .04). first name along with a big (vs. slight or no) smiley face at the end of When the employee was a manager (higher rank), nevertheless, par- the request description, participants had a higher compliance likeli- ticipants liked him less if he had a bigger smile (p = .06). hood. However, when we introduced ourselves as a Ph.D. researcher, Study 2 (N = 229) replicates the proposed effects in an advertis- a slight smile led to a higher compliance rate (Wald’s χ2 (1) = 6.64, p ing context and extends the investigation to purchase intentions. We = .01 for the interaction). also investigate the moderating role of processing time. Consumers In conclusion, when a big (slight) smile is displayed matching often make decisions under time pressure (Cristol and Sealey 1996; the displayer’s low (high) rank, consumers perceive a higher level of Dhar and Nowlis 1999). Consistent with the heuristic-systematic consistency with social expectations for emotion displays. Therefore, dual-processing model, consumers focus on fewer informational they like the displayer better and are more likely to purchase the as- cues when they have limited time (Dhar and Nowlis 1999). When sociated product or service. As important evidence supporting this the processing time is shorter, we propose that consumers should proposed conscious process, these results hold only when consumers focus on the overall communications and be less likely to process have enough time to decode the emotional expression cues (smiles) peripheral informational cues, such as social expectations for emo- in the communication process. tional displays. This study has a 2 (smile type: slight vs. big) x 2 (rank: lower REFERENCES vs. higher) between-subjects design. Participants were shown one of Barger, Patricia B., and Alicia A. Grandey (2006). Service with a four ads with a female displayer promoting a black tea. The female Smile and Encounter Satisfaction: Emotional Contagion and had either a big or a slight smile (confirmed by a manipulation check) Appraisal Mechanisms. Academy of Management Journal, 49 and was titled as either doctor (pretested to have higher rank) or first (November), 1229-1238. name (pretested to have lower rank). We also measured processing Bugental, Daphne Blunt (1986), “Unmasking the” Polite Smile” time as a continuous variable by recording the time that participants Situational and Personal Determinants of Managed Affect in spent on viewing the ad. Adult-Child Interaction,” Personality and Social Psychology The three-way interaction effect (smile type x rank x processing Bulletin, 12(March), 7-16. time) on consumers’ choice was significant (Wald’s 2χ (1) = 5.11, p = Cheng, Yimin, Anirban Mukhopadhyay, and Patti Williams (2020), .02). Specifically, when participants spent longer time (1 SD above “Smiling Signals Intrinsic Motivation,” Journal of Consumer the mean) viewing the ad, they chose to buy the tea more often if the Research, 46 (February), 915-935.

Cristol, Steven M., and Peter Sealey (1996), “Too much choice.” Small, Deborah A., and Nicole M. Verrochi (2009), “The Face of Marketing Management, 5 (1), 14. Need: Facial Emotion Expression on Charity Advertisements,” Dhar, Ravi, and Stephen M. Nowlis (1999), “The effect of time Journal of Marketing Research, 46 (December), 777-787. pressure on consumer choice deferral,” Journal of Consumer Wang, Z., Mao, H., Li, Y. J., & Liu, F. (2017). Smile big or not? Research, 25 (4), 369-84. Effects of smile intensity on perceptions of warmth and Hayes, Andrew F. (2013), Introduction to Mediation, Moderation, competence. Journal of Consumer Research, 43 (February), and Conditional Process Analysis, New York: Guildford. 787-805. Mueser, Kim T., Barry W. Grau, Steve Sussman, and Alexander J. Rosen (1984), “You’re Oly as Pretty as You Feel: Facial Expression as a Determinant of Physical Attractiveness,” Journal of Personality and Social Psychology, 46(2), 469-478. The Effect of Product Evaluations on Consumer Recycling Behavior Vivian (Jieru) Xie, Virginia Tech, USA Rajesh Bagchi, Virginia Tech, USA

EXTENDED ABSTRACT To investigate whether respect mediated the effect, we used While the industrial revolution has heralded an age of unprec- an Implicit Association Test (iatgen via Qualtrics, Carpenter et al. edented affluence and choice abundance, it has also unfortunately 2018) to examine the positive association between the concepts of increased wastefulness. Trashing products mean that not only did we respect and recycling (vs. trashing) in study 2. The goal of this study not utilize the full potential of the raw material, but that we will be was to assess if respondents associate “Recyclable materials” with imposing further stress on our landfills and incineration facilities. “Respect” and “Non-recyclable materials” with “Disrespect.” We Thus, it is of critical importance to understand how consumers can provided participants (N = 208) with 5 words belonging to each of be encouraged to recycle more. the four categories and ask them to categorize stimuli. In the “com- Consumers’ decisions to recycle (vs. trash) a product are sub- patible” block, “Recyclable material” and “Respect” were on the ject to a variety of influences, which range from consumer-centric one side, while “Non-recyclable material” and “Disrespect” were to societal-level factors. Consumer factors, such as knowledge (e.g., on the other side. In the “incompatible” block, the combinations Andrew et al. 2013) and concerns for the environment (Schultz and were switched. As shown in Table 1, the association occurred more Oskamp, 1996), product-level factors, such as product form distor- quickly in the compatible than in the incompatible block therefore, tion (Trudel and Argo 2013) and identity-relevance (Trudel et al. confirming our predictions. 2016), and finally, social-class and cultural variables (e.g., Laidley In study 3, we replicate these effects in a context with real re- 2013) have all been shown to affect consumers’ decisions to trash cycling behaviors, demonstrating external validity. We recruited 153 or recycle. undergraduates to complete this study in an individual lab room to We explore how a specific post-ownership experience, evalu- avoid distractions. We provided each participant with a piece of ori- ations (positive vs. negative), influences consumers’ disposal deci- gami paper and asked them to make an origami fox following a tuto- sions. We evaluate products all the time and even share our evalua- rial. Then they were required to take their finished origami fox to tions. While our evaluations play a fundamental role in affecting our the lab assistant. The lab assistant took pictures of their work and future engagement with products, these evaluations could also affect told them that an auto-grading system would score their work (but our disposal decisions—whether we trash or recycle a product? For in fact the grade was assigned randomly). Thus, half the participants example, while the evaluation of a drink’s taste is likely to affect received a positive grade (8.5/10), while the other half received a consumers’ propensity to repurchase the drink, would it also impact negative grade (4.5/10). We then asked them to dispose of their work their decision to recycle the container? in the bins (a recycling bin and a trash bin) in the individual lab From a material-value perspective, there should be no reason to room. After that, they indicated their respect level toward the origami expect any difference in disposal propensity. If consumers are aware work. Unbeknownst to participants, after they left, the lab assistant that a product (e.g., a drink’s bottle) is recyclable, then, normatively, recorded whether they recycled or trashed their fox and removed it their disposal decisions should be independent of the product’s eval- to retain the same presets for each participant. As shown in Table 1, uation (tastes good or bad). However, we argue that consumers will more of the participants who received a manipulated positive (vs. be more likely to recycle products that they evaluate positively (vs. negative) evaluation score recycled (vs. trashed) their origami work, negatively), and, thus, will be more likely to recycle a bottle when and respect toward their origami work mediated these effects. they evaluate the drink’s taste positively. We introduce a novel effect of product evaluations on consumer We propose that this effect emerges because people are more recycling behavior—consumers are more likely to recycle a prod- likely to respect a product with positive (vs. negative) evaluations, uct when it is positively evaluated. Overall, we believe this research and it is this respect that drives them to recycle (vs. trash). Respect makes important contributions from both theoretical and practical is earned and is often based on reasons (Dillon 2007). We propose perspectives. From a theoretical perspective, this research explores that a product associated with positive (vs. negative) evaluations will the impact of intangible characteristics related to the product itself— also be accorded greater respect by consumers. Furthermore, we be- product evaluations. We expand the literature on disposal decision lieve that products that are respected more will also be given a more making by demonstrating how a novel product-level characteristic respectful burial—recycling (vs. trashing) may be a more respectful — product evaluations — impact disposal decisions to trash or recy- way to dispose of a recyclable product. In summary, we propose that cle. From a practical perspective, we believe it is of great importance people are more likely to respect a product associated with positive to understand what drives consumers to recycle more. Our research (vs. negative) evaluations, which would increase likelihood of recy- adds to the growing discourse on environmental issues. cling (vs. trashing). In study 1 (N = 173), participants were randomly assigned to REFERENCES one of two conditions. In the positive (negative) evaluation condi- Carpenter, Tom, Ruth Pogacar, Chris Pullig, Michal Kouril, Stephen tion, participants were informed that a brand of bottled water was J. Aguilar, Jordan P. LaBouff, Naomi Isenberg, and Aleksandr very well (poorly) rated with an evaluation score of 4.8/5 (2.2/5) be- Chakroff (2008), “Survey-software Implicit Association cause most consumers liked the taste (did not like the taste). In both Tests: A Methodological and Empirical Analysis, “ Behavior cases the product was priced the same. Participants then indicated Research Methods, 51 (5), 2194-2208. how would they dispose of the bottle if they received a free sample Dillon, Robin S. (2007), “Respect: A Philosophical Perspective,” and finished it. As shown in Table 1, consumers were more likely to Gruppe. Interaktion. Organisation. Zeitschrift für Angewandte recycle the bottle, and believed the bottle was more worthy of recy- Organisationspsychologie (GIO), 38 (2), 201-12. cling when its taste was evaluated positively.

Greenwald, Anthony G., Brian A. Nosek, and Mahzarin R. Banaji Trudel, Remi, and Jennifer J. Argo (2013), “The Effect of Product (2003), “Understanding and Using the Implicit Association Size and Form Distortion on Consumer Recycling Behavior,” Test: I. An Improved Acoring Algorithm,” Journal of Journal of Consumer Research, 40 (4), 632-643. Personality and Social Psychology, 85 (2), 197-216. Trudel, Remi, Jennifer J. Argo, and Matthew D. Meng (2016), “The Laidley, Thomas M. (2013), “The Influence of Social Class and Recycled Self: Consumers’ Disposal Decisions of Identity- Cultural Variables on Environmental Behaviors: Municipal- Linked Products,” Journal of Consumer Research, 43 (2), Level Evidence from Massachusetts,” Environment and 246-264. Behavior, 45 (2), 170-97. Schultz, P. Wesley, and Stuart Oskamp (1991), “Effort as A Moderator of the Attitude-Behavior Relationship: General Environmental Concern and Recycling,” Social Psychology Quarterly, 59 (4), 375-83. Was the Past Heavier? The Impact of Product Temporal Orientation on Perceived Heaviness and Product Purchase Lan Xu, Wuhan University, China June Cotte, Western University, Canada Nan Cui, Wuhan University, China Yiran Jiang, Wuhan University, China

EXTENDED ABSTRACT Hypothesis 3: Consumers are more likely to buy a product with Consumers may use perceived heaviness as a clue to diagnose a past (future) temporal orientation when it is the durability and quality of a tangible product (Krishna, 2006; Lin, designed to be heavier (lighter). 2013; Togawa, Park, and Ishii, 2016). Extant research has found that product-related attributes such as size (Murray et al., 1891), color We tested our hypotheses through three lab experiments and lightness (Hagtvedt, 2019), or location of the product image on a a field experiment. Study 1 examined the impact of the product’s package (Deng & Kahn, 2009) affect individuals’ perception of prod- temporal orientation (past vs. future) on perceived product weight. uct weight. In this research, we theorize and demonstrate a new as- A total of 75 undergraduates participated in this experiment. The pect- the temporal orientation of a product (the perception of it as be- experiment followed a two-cell (product temporal orientation: past ing from the past or the future)-that influences consumers’ perception vs. future) between-subjects design. We manipulated the temporal of product weight, and ultimate product choice. orientation, informing participants that a newly-designed notebook Our theory about how product temporal orientation affects either “was launched in the market six months ago” (past) or “will perceptions of heaviness is a processing-based one. Compared to be launched in the market in six months” (future). Then participants the future, consumers can think of many more specific events that rated their perceived weight of the notebook using 9-point scales have happened in the past. Hence, one’s memories and the facts (1 = relatively light and 9 = relatively heavy). The results revealed related to one’s past are specific and detailed (Muehling, Sprott, & a significant effect of product temporal orientation on perceived Sprott,2004). Thinking about the past will inevitably trigger recol- weight (F(1,73)=37.62, p<0.001). Participants in the past condition lection and reflection on these specific facts and experiences (Van perceived the product to be heavier than those in the future condition Boven, Kane, & McGraw, 2010). However, thinking about the fu- (Mpast = 5.72, SDpast = 2.08, Mfuture = 3.03, SDfuture = 1.68). ture, no matter the extent of imagination or planning involved, is Study 2 examined the robustness of H1 and the mediating role more uncertain and less specific (Epley & Dunning, 2000; Wilson et of specificity of thinking in the impact of product temporal orienta- al., 2000). The different temporal orientation of products may trigger tion on perceived weight. A total of 114 undergraduates participated us to retrieve our experiences (past versus future) and we recollect in this experiment. The experiment followed a two-cell (product more specific facts from the past, which we argue feels heavier, and temporal orientation: past vs. future) between-subjects design. As may influence how we perceive the heaviness of a product we are predicted, participants in the past condition perceived the product considering. to be heavier. Mediation analysis showed that the indirect path of When consumers are considering a product that is temporally temporal orientation on perceived weight via specificity of thinking oriented to the past, they are more likely to retrieval their sensory was significant. experiences and memories about events that happened in the past Study 3 was to verify the impact of the match of temporal orien- (Deng, Han, and Wang, 2019). The immediate experience of recall- tation and product weight on purchase intention (H3). A total of 148 ing details of events evokes simultaneously muscular thinking, which undergraduates participated in this experiment. The experiment fol- renders inner experience felt to be heavy (Winnicotti, 1971). The lowed a 2 (product temporal orientation: past vs. future) × 2 (product more details consumers consider, the heavier they will feel during weight: heavy vs. light) between-subjects design. Participants were this thinking/processing. The feeling of heaviness will be projected asked to imagine a scenario of lamp purchase. Temporal orientation to the object which leads to thinking (Jostmann, Lakens, & Schubert, was manipulated by asking participants to look for either an antique 2009). In contrast, when exposed to a more futuristic product, con- lamp or a futuristic lamp. The product description was designed for sumers are less likely to perceive heaviness because less processing the manipulation of the product weight (2500g net weight vs. 500g of details happens when thinking about the future. net weight). As predicted, an ANOVA revealed a significant interaction of temporal orientation and product weight on purchase inten- Hypothesis 1: Compared to the product with a future temporal tion (F(1,144)=13.57, p<0.01). orientation, consumers will perceive a product Study 4, a field experiment, was designed to examine H3 in a with a past temporal orientation as heavier. real shopping situation. We examined consumers’ actual purchase by cooperating with a university souvenir shop selling bookmarks for Hypothesis 2: The specificity of thinking related to product four weeks. The experiment was a two-way mixed design, with a temporal orientation mediates the influence of two-level between-subjects factor (temporal orientation: past vs. fu- the product’s temporal orientation on perceived ture) and a two-level within-subjects factor (product weight: heavy weight. vs. light). The dependent variable was the real purchase of the bookmarks. Our results revealed a significant effect of the interaction of Based on processing fluency theory, consumers will be more temporal orientation and product weight (χ2 = 6.23, p<0.05) on prod- fluent in processing a weight-related stimulus that matches their cur- uct purchase. Specifically, in the past condition, a majority ofthe rent state of thinking of based on temporal orientation (Labroo et al., shoppers preferred the heavy bookmark (68.6%, [35/51]), whereas 2008). in the future condition, a majority of the shoppers chose the light bookmark (55.8% [29/52]), thus supporting H3.

In summary, four studies support the predicted relationship Labroo, A. A., Dhar, R., & Schwarz, N. (2008). Of Frog Wines and between product temporal orientation and perceived weight. Spe- Frowning Watches: Semantic Priming, Perceptual Fluency, cifically, compared to a product with a future temporal orientation, and Brand Evaluation. Journal of Consumer Research, 34(6), consumers perceive a product with a past temporal orientation to be 819–831. heavier, which was mediated by the specificity of thinking related to Lin, H. M. 2013. Does container weight influence judgments of temporal orientation. Furthermore, consumers had higher purchase volume. International Journal of Research in Marketing, intention, and more actual purchases, when the product’s temporal 30(3): 308–309. orientation matched its actual weight. Jostmann, N. B., Lakens, D., & Schubert, T. W. (2009). Weight as an embodiment of importance. Psychological science, 20(9), REFERENCES 1169–1174. Ariely, D., & Norton, M. I. (2009). Conceptual Consumption. Muehling, D. D., Sprott, D. E., & Sprott, D. E. (2004). The Annual Review of Psychology, 60(1), 475–499. power of reflection: An empirical examination of nostalgia Brown, L. J. (1985). On concreteness. Psychoanalytic Review, advertising effects.Journal of Advertising, 33(3), 25–35. 72(3), 379–402. Murray D J, Ellis R R, Bandomir C A, et al. 1999. Charpentier Deng, X., Han, B., & Wang, L. (2019). Up-Down versus Left- (1891) on the size-weight illusion[J]. Attention Perception & Right: The Effect of Writing Direction Change in East Asia Psychophysics, 61(8): 1681–1685. on Consumers’ Perceptions and Advertising. Journal of Taku Togawa, Jaewoo Park, & Hiroaki Ishii. (2016). Does Visual Advertising, 48(5), 437–456. Heaviness Convey Rich Flavor? Effect of Product Image Deng, X., & Kahn, B. E. (2009). Is Your Product on the Right Side? Location on Consumers’ Taste Expectation. Advances in The “Location Effect” on Perceived Product Heaviness and Consumer Research, 44, 762. Package Evaluation. Journal of Marketing Research (JMR), Van Boven, L., Kane, J., & McGraw, A. P. (2009). Temporally 46(6), 725–738. asymmetric constraints on mental simulation: Retrospection Epley, N., & Dunning, D. (2000). Feeling “holier than thou”: Are is more constrained than prospection. In K. D. Markman, W. self-serving assessments produced by errors in self- or social M. P. Klein, & J. A. Suhr (Eds.), Handbook of imagination and prediction? Journal of Personality and Social Psychology, mental simulation. (pp. 131–147). Psychology Press. 79(6), 861–875. Winnicott, D. W. (1971) Playing and reality. New York: Basic Hagtvedt, H. (2019). Dark is durable, light is user‐friendly: The books. impact of color lightness on two product attribute judgments. Wilson, T. D., Wheatley T, Meyers J M, et al. 2000. Focalism: A Psychology & Marketing,1-12. source of durability bias in affective forecasting[J]. Journal of Krishna, A. 2006. Interaction of senses: The effect of vision versus Personality & Social Psychology, 78(5): 821–836. touch on the elongation bias. Journal of Consumer Research, 32(4): 557–566. How do Consumers React to Company Moral Transgressions? The Role of Power Distance Belief and Empathy for Victims Haiyue (Felix) Xu, The Pennsylvania State University, USA Lisa E. Bolton, Pennsylvania State University, USA Karen Page Winterich, Pennsylvania State University, USA

EXTENDED ABSTRACT remained when including control variables and conducting robust- Company moral transgressions are prevalent in today’s mar- ness checks with different event windows. ketplace (e.g., Nike’s poor labor practices, Volkswagen’s emission In Study 2, participants completed a PDB manipulation (Zhang, scandal). With the Internet and social media, such company moral Winterich, and Mittal 2010) and read about two fitness brands, one transgressions rapidly receive public scrutiny (Crockett 2017), which with a moral transgression (poor labor practice), and the other with can cause considerable damage to transgressing firms (e.g., Frooman a non-moral transgression (poor product quality) and indicated their 1997). response. Consistent with our prediction, low PDB consumers made

However, not all consumers have the same level of concern to- less favorable evaluations (Mlow PDB =4.94 vs. Mhigh PDB =5.47, F(1, ward company moral transgressions. Moral norms are often subjec- 245)=5.41, p=.021). Participants also had a chance to win a $25 gift tively determined by one’s moral system (Trevino and Nelson 2016) card to the brand of their choice: low PDB consumers were margin- with moral judgments varying substantially across individuals and ally less likely to choose the morally transgressing brand (Mlow PDB 2 cultures (Haidt and Joseph 2004; Schein and Gray 2018). The cur- =68.6% vs. Mhigh PDB =79.0%, χ (1)=3.12, p=.077). rent research examines how, when, and why power distance belief Study 3 tests the empathy mechanism. Participants completed (hereafter, PDB), which is the extent to which people accept and en- the same PDB manipulation as study 2 and then read about a differ- dorse inequality and hierarchy in society (Hofstede 1980), influences ent moral transgression (environmental harm), followed by measure- consumer response following company moral transgressions. Equal- ment of their response and empathy toward victims (Batson 1987). ity is a prominent moral principle, and a moral transgression can be As expected, low-PDB participants responded less favorably (Mlow conceptualized as a powerful transgressor (corporations) harming PDB =1.93 vs. Mhigh PDB=2.34, F(1, 207)=4.53, p=.035) and expressed vulnerable victims (Gray and Wegner 2009). Hence, we propose that greater empathy (Mlow PDB=5.67 vs. Mhigh PDB=5.15, F(1, 207)=4.47, PDB will affect how individuals react to company moral transgres- p=.036), and the effect of PDB on consumer response was mediated sions with empathy toward the transgression victims as the underly- by empathy. ing mechanism. Study 4 uses an established empathy manipulation (high vs. Empathy is an other-oriented emotion that arises from witness- control) to demonstrate moderation-of-process (Spencer, Zanna, and ing someone in distress that subsequently motivates an effort to help Fong 2005). Participants read about a company moral transgression (Batson 2011; DeSteno 2015). We theorize that i) low-PDB individu- (poor labor practice), with an empathy manipulation embedded, and als are motivated to minimize social inequality and tend to feel empa- then indicated their response and PDB (measured with Yoo et al. thy for the vulnerable in society (Winterich and Zhang 2014), such as 2011). Results revealed the expected interaction between measured transgression victims, whereas high-PDB individuals tend to defend PDB and manipulated empathy (b=-.16, t(198)=-1.98, p=.049): low- social hierarchy and status quo (Jost, Banaji, and Nosek 2004) and er PDB participants responded less favorably in the control condition hence may perceive transgression victims as less warranting of care (b=.33, t(198)=2.80, p=.006) but, as expected, the PDB effect was and empathy (Lee, Winterich, and Ross 2014), and ii) the experience attenuated in the high empathy condition (b=.01, t(198) =.07, p>.05). (or lack) of empathy further influences how negatively consumers Study 5 manipulates victim salience by using two transgressions evaluate transgressing companies. Thus, low- (versus high-) PDB that vary naturally in victim salience: mistreating employees (high) consumers, due to their greater empathy for transgression victims, vs. disrespecting traditions (low). Consumer response and PDB were will have less favorable evaluations of transgressing companies. measured as in study 4. We found the expected interaction (b=.21, Consistent with this theorizing, we propose that the effect of t(305)=2.67, p=.008): lower PDB participants responded less favor- PDB is attenuated when the company moral transgression does not ably when victim salience was high (b=.49, t(305)=5.20, p < .001) have a salient victim as even low-PDB consumers do not focus on but not when victim salience was low (b=.08, t(305)=.67, p>.05). victims, thereby suppressing empathy. We also investigate company In Study 6, participants read about the a moral transgression response strategies and propose that neither an apology nor remedy (environmental harm) and were randomly assigned to either no re- alone dampens the PDB effect but a combined apology and remedy sponse, an apology, a remedy, or a combined apology and remedy, can attenuate the effect by signaling the company’s empathy for vic- with consumer response and PDB measured as in studies 4-5. Results tims. revealed the expected interaction between PDB and firm response In study 1, we examine search data regarding the Volkswagen (F(3, 397)=7.73, p<.001): an effect of PDB following no response emission scandal together with country-level cultural values (Hof- (b=.73, t(397)=7.73, p<.001), an apology (b=.74, t(397)=7.89, stede 2018) to test the impact of PDB on consumer response to the p<.001), or a remedy (b=.58, t(397)=5.73, p<.001), but not after a scandal. Using Google Trends, we performed the topic search of combined apology and remedy (b=.16, t(397)=1.65, p>.05). “Volkswagen emission scandal” across all global regions. The PDB In six studies, we find robust support for our theorizing across scores along with other cultural values were obtained from Hof- various consumer samples, operationalizations of PDB, and trans- stede’s (2018) website and matched to the search data along with gression contexts. In doing so, this research makes important con- other controls (e.g., political freedom, GDP). Based on a one-year tributions to the literature on corporate social (ir)responsibility and event window following the scandal, PDB had a negative effect on consumer/marketplace morality in a cross-cultural context. This re- the search index (b=-.41, t(54)=-3.93, p < .001); that is, people from search also has implications for marketers to understand when and lower PDB countries were more likely to search for the scandal, con- why consumers have negative reactions to company moral transgres- sistent with greater concern for the moral transgression. This pattern sions, as well as guidance for company response strategies.

Advances in Consumer Research 792 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 793 REFERENCES Lee, Saerom, Karenn Page Winterich, and William T. Ross (2014), Batson, C. Daniel (1987), “Prosocial Motivation: Is It Ever Truly “I’m Moral, but I Won’t Help You: The Distinct Roles of Altruistic?” Advances in Experimental Social Psychology, 20, Empathy and Justice in Donations,” Journal of Consumer 65-122. Research, 41 (3), 678-96. ______(2011), Altruism in Humans, Oxford University Schein, Chelsea, and Kurt Gray (2018), “The Theory of Dyadic Press. Morality: Reinventing Moral Judgment by Redefining Harm,” Crockett, M. J. (2017), “Moral Outrage in the Digital Age,” Nature Personality and Social Psychology Review, 22 (1), 32-70. Human Behaviour, 1 (11), 769-71. Spencer, Steven J., Mark P. Zanna, and Geoffrey T. Fong (2005), DeSteno, D. (2015), “Compassion and Altruism: How our Minds “Establishing a Causal Chain: Why Experiments Are Often Determine Who is Worthy of Help,” Current opinion in More Effective than Mediational Analyses in Examining Behavioral Sciences, 3, 80-83. Psychological Processes,” Journal of Personality and Social Frooman, Jeff (1997), “Socially Irresponsible and Illegal Behavior Psychology, 89 (6), 845-51. and Shareholder Wealth: A Meta-Analysis of Event Studies,” Trevino, Linda K., and Katherine A. Nelson (2016), Managing Business & Society, 36 (3), 221-49. Business Ethics: Straight Talk About How to Do It Right, John Gray, Kurt, and Daniel M. Wegner (2009), “Moral typecasting: Wiley & Sons. divergent perceptions of moral agents and moral patients,” Yoo, Boonghee, Naveen Donthu, and Tomasz Lenartowicz (2011), Journal of personality and social psychology, 96 (3), 505-20. “Measuring Hofstede’s Five Dimensions of Cultural Values Haidt, Jonathan, and Craig Joseph (2004), “Intuitive Ethics: How at the Individual Level: Development and Validation of Innately Prepared Intuitions Generate Culturally Variable CVSCALE,” Journal of International Consumer Marketing, Virtues,” Daedalus, 133 (4), 55-66. 23 (3-4), 193-210. Hofstede, Geert (1980), “Culture and Organizations,” International Winterich, Karen Page, and Yinlong Zhang (2014), “Accepting Studies of Management and Organization, 10 (4), 15-41. Inequality Deters Responsibility: How Power Distance ______(2018), “Compare Countries,” Hofstede Insights, Decreases Charitable Behavior,” Journal of Consumer 10 (4), https://www.hofstede-insights.com/product/compare- Research, 41 (2), 274-293. countries/. Zhang, Yinlong, Karen Page Winterich, and Vikas Mittal (2010), Jost, John T., Mahzarin R. Banaji, and Brian A. Nosek (2004), “A “Power Distance Belief and Impulsive Buying,” Journal of Decade of System Justification Theory: Accumulated Evidence Marketing Research, 47 (5), 945-54. of Conscious and Unconscious Bolstering of the Status Quo,” Political Psychology, 25 (6), 881–919. Getting Nothing Feels Fairer Than Getting Something: How Allocation Outcomes Invoke Different Allocation Rules Minzhe Xu, University of Florida, USA Bowen Ruan, University of Iowa, USA

EXTENDED ABSTRACT and is perceived to be fairer than the binary rule in most cases (Cook Most resources are limited, which is presumably why most so- and Hegtvedt 1983; Berkowitz et al. 1987), we predicted that par- cial systems evolve mechanisms for resource allocation (Cook and ticipants would perceive the perfectly proportional allocation as even Hegtvedt 1983). A primary goal of an effective allocation mechanism fairer than the none-all allocation. Second, we tested a boundary con- concerns how fair people perceive the allocation of resource to be. dition: if people are prompted to use the binary rule regardless of the Indeed, people care about fairness so much that they would sacri- actual resource allocation (e.g., performance itself is evaluated in a fice their own interests (and punish others) for the sake of fairness binary way, such as pass/fail), those faced with a small-large alloca- (Dawes et al. 2007). tion tend not to calculate or compare the proportions. We predicted A common practice to achieve fairness is to proportionally al- that in such case, the difference between the none-all-allocation and locate resource based on the performances of the entities involved the small-large-allocation conditions would be smaller. Study 3 rep- (Adams 1963; Hatfield et al. 1978; Ma and Roese 2013). Consider resented a 3 (allocation: none-all vs. small-large vs. proportional) × 2 a sales contest in which two salespersons—A and B—compete for (performance: binary vs. control) between-subjects design. Support- a total bonus of $1,000. A has sold 9,000 units in the contest, and ing our hypotheses, in the control condition, participants perceived B has sold 10,000 units. Based on the proportion rule, the fairest the small-large allocation to be less fair than the none-all allocation allocation of the bonus is A receiving $474 (9/19 of $1,000) and B (2.47 vs. 4.34, F(1, 567)=56.51, p<.001), which was less fair than receiving $526 (10/19 of $1,000), and the less an allocation deviates the proportional allocation (4.34 vs. 6.33, F(1, 567)=66.95, p<.001). from the fairest allocation, the fairer people should perceive it to be. Critically, in the binary-performance condition, the difference be- Therefore, it can be inferred that people perceive A receiving a small tween the small-large-allocation and the none-all-allocation condi- compensation amount (e.g., A receives $100 and B receives $900, a tions became smaller (4.37 vs. 4.95, F(1, 567)=5.60, p=.02; partial small-large allocation) to be fairer than A receiving no compensation interaction: F(1, 567)=13.61, p<.001). (e.g., A receives $0 and B receives $1,000, a none-all allocation). Study 4 examined another boundary condition. In a non-com- However, we find that the opposite is true: in a study (N=130), petitive setting where individuals collaborate rather than compete participants perceived A receiving no compensation ($0) to be fairer with each other, the binary allocation rule may not be applicable, than A receiving a small compensation amount ($100; 4.02 vs. 3.08, attenuating our effect. In a 2 (allocation: none-all vs. small-large) × F(1, 128) = 8.05, p = .01). Study 2 replicated this finding with an 2 (context: competition vs. non-competition) between-subjects de- incentive-compatible design. Participants (N=147) were invited to sign, participants (N=342) imagined that they submitted a research compete with another participant in a math-addition task to win an proposal to either a competition where researchers competed against extra bonus. They also learned, before the competition started, that each other for research grant or a symposium where researchers in- the person who answered more questions correctly would receive 50 teracted with each other. They learned that proposals with 80 or more cents, whereas the other person would receive 0 cents (the none-all- votes (out of 100) would receive a grant of $500,000, while those allocation condition) or 1 cent (the small-large-allocation condition). with fewer than 80 votes would receive either no grant (none-all) Participants were not really in a competition—they were all told at or $10,000 (small-large). They then learned that their own proposal end of the task that they lost (by 1-3 questions, randomly determined) received 78 votes and received no grant or $10,000. We found that in and thus received a bonus of 0 cents or 1 cent. Supporting our hy- the competition condition, participants perceived the none-all alloca- pothesis, participants in the none-all-allocation (vs. small-large-al- tion to be fairer than the small-large allocation (4.36 vs. 3.78, F(1, location) condition rated the allocation rule as fairer (4.28 vs. 3.57, 338)=4.58, p=.03), while in the non-competition condition, the effect F(1, 145)=5.13, p=.03) and were more likely to do the task one more disappeared (3.76 vs. 4.18, F(1, 338)=2.32, p=.13; interaction: F(1, time with the same participant and the same allocation rule (71% vs. 338) = 6.66, p = .01). 53%; χ2(1)=4.99, p=.03). This research presents a novel effect regarding resource alloca- We argue the reason is that a small-lage allocation (e.g., $100- tion that a small compensation amount (vs. no compensation) may $900) invokes the proportion rule, and prompts people to calculate the backfire. It contributes to the literature on fairness perception by proportion of the allocation (e.g., $100/$900=1/9) and compare it to demonstrating that people may infer different allocation rules from the proportion of performance (e.g., 9,000 units/10,000 units=9/10). the actual allocations of resource, which can have important conse- Because of the drastic difference between these two proportions, quences on fairness perception and behavior. people likely infer that the small-large allocation violates the proportion rule to a large extent and thus perceive it as unfair. In contrast, REFERENCES when the low performer receives no compensation, people tend to Adams, J. Stacy (1963), “Towards an Understanding of Inequity,” infer the allocation rule is binary (e.g., the winner takes all) and are Journal of Abnormal and Social Psychology, 67 (5), 422–36. less likely to pay attention to the proportion of performance. Because Berkowitz, Leonard, Colin Fraser, F. Peter Treasure, and Susan the actual allocation is binary and thus consistent with the inferred Cochran (1987), “Pay, Equity, Job Gratifications, and allocation rule, people perceive the none-all allocation as fair. Comparisons in Pay Satisfaction,” Journal of Applied Study 3 (N=573) had two goals. First, to test the proposed Psychology, 72 (4), 544-51. mechanism, we added another condition in which the resource al- Cook, Karen S. and Karen A. Hegtvedt (1983), “Distributive location perfectly followed the proportion rule. Although in both this Justice, Equity, and Equality,” Annual Review of Sociology, 9 and the none-all-allocation condition, the allocation aligned with the (1), 217–41. inferred allocation rule, because the proportion rule ensures equity

Dawes, Christopher T., James H. Fowler, Tim Johnson, Richard Ma, Jingjing and Neal J. Roese (2013), “The Countability Effect: McElreath, and Oleg Smirnov (2007), “Egalitarian Motives in Comparative versus Experiential Reactions to Reward Humans,” Nature, 446 (7137), 794–96. Distributions,” Journal of Consumer Research, 39 (6), Hatfield, Elaine, Elaine Hatfield Walster, G. William Walster, and 1219–33. Ellen Berscheid (1978), Equity: Theory and Research, Allyn & Bacon. Social Media Magnetism and Marketplace Pilgrimage Haibo Xue, East China Normal University, China Xin Zhao, Lancaster University Management School, UK

EXTENDED ABSTRACT cial capital that can then be competed and displayed in social media In this paper, we investigate why certain quintessential tourist space. It gives visual forms to consumers’ formless and exhausting spots, shops, and retailing outlets become an instant magnetic site on labor of waiting. In this case, consumers do not deliberately seek to social media and attract a large number of pious followers. Within slow down and to escape from the accelerated social life. It is ori- the urban consumptionscape in Shanghai and many other metropolis ented toward social media conspicuity and is an extrinsic form of de- in China, huge groups of consumers may spend an average of over celeration that does not aim at introspection and reflection. The cult 5 hours, and as long as 7 hours waiting in line in order to purchase shops that attract pilgrims are characterized by scarcity, difficulty of a cup of bubble tea, a snack, a Chinese pancake, or even a piece of access, ability of being traced on social media, or their potential for cookie. The zeal for such retail outlets is often fueled by Chinese visual transportation and transformation, as well as endorsement of social media such as Tiktok, the Little Redbook, WeChat, Weibo, and social media celebrity. We refer to the quintessential qualities of such online forums. Such avid desire can only be matched by the devo- pilgrimage space as social media magnetism. tions of religious pilgrimage (e.g., Turner and Turner 1978). In this paper, we extend previous literature on consumer de- In particular, we examine consumption rituals at these places celeration by highlighting the simulation of consumer deceleration and discuss implications for understanding pilgrimage in the market- through pilgrimage in the marketplace. Our notion of social media place as a simulated form of consumer deceleration (e.g., Husemann magnetism explains why certain retailing space becomes an instant and Eckhardt 2019; Rosa 2013). We seek to understand: whether or spot of such devotion. Together, we contribute to consumer culture not the perceived slowed-down temporal experience can be achieved theories on the sacralization of commercial spaces by highlighting in the marketplace and without the expected detachment from one’s the transient value of marketplace pilgrimage on social media in the social worlds as have been previously suggested. era of assess-based consumption (e.g., Bardhi and Eckhardt 2017). Pilgrimage is a journey to the sacred place containing the actual sacred relics from the past, which has the power to contaminate REFERENCES pilgrims with magical cure or positive energy (Turner and Turner Bardhi Fleura, and Eckhardt Giana M. (2017), “ Liquid 1978). Consumers seek spirituality and self-transformation through Consumption,” Journal of Consumer Research, 44 (3), religious pilgrimages (Hamilton and Higgans 2011; Higgans and 582–97. Hamilton 2016). More importantly, these physically demanding Hamilton, K.L. and Higgans, L. (2011), “Sacred Places: An and sacred journeys allow consumers to escape from their acceler- Exploratory Investigation of Consuming Pilgrimage,” in ated social life (Rosa 2013) and to engage in consumption practices D. W. Dahl, G.V. Johar, and S.M.J. van Osselaer (eds), NA that help them to slow down (Husemann and Eckhardt 2019). Such Advances in Consumer Research, 542-48. Vol 38. Duluth, MN: deceleration is a temporal consumption experience that is achieved Association for Consumer Research. through immersing in extended geographic journeys far away from Higgins, L. and Hamilton, K. (2016), “Mini-miracles: home and ideally from the market. In this paper, we argue that such Transformations of Self from Consumption of the Lourdes temporal experiences of slowing down may be achieved through the Pilgrimage,” Journal of Business Research, 69 (1): 25-32. simulation of deceleration in marketplace pilgrimage. Husemann, Katharina C and Giana M Eckhardt, Consumer In this study, we adopted both ethnographic and ethnographic Deceleration, Journal of Consumer Research, 45 (6), 1142– approaches (Kozinets, 2015), to understand marketplace pilgrimage. 1163 We used emergent design, multiple sites, purposive sampling, and Kozinets, Robert V. (2015). Netnography: Redefined: Sage. field interviews in the process of data collection. Our data collec- Spiggle, Susan. (1994). Analysis and interpretation of qualitative tion started in Spring 2017 and continued into Summer 2019. The data in consumer research. Journal of Consumer Research, research team visited various cult food shops in Shanghai and par- 21(3), 491-503. ticipated in the waiting. This helped them to reflect upon the observa- Rosa, Hartmut (2013), Social Acceleration: A New Theory of tions and to develop a better understanding of consumer experiences. Modernity, New York: Columbia University Press. We observed different waiting lines outside different cult food stores. Turner, Victor W. and Turner, E. (1978), Image and Pilgrimage in Pictures and interviews were combined with detailed observation Christina Culture: Anthropological Perspectives, New York, notes. Our total dataset consists of 35 depth interviews, 120 pages of NY: Columbia University Press. fieldnotes, 257 images from our fieldwork, and over 300 social media accounts that describe consumption experiences of these cult food stores. Our data analysis proceeded in an iterative process (Spiggle 1994). We focus on consumers’ experiences and seek to understand why consumers spent prolonged periods of time waiting in line. The experiences of prolonged waiting in line at the shops are commonly understood and interpreted by our informants as a “pilgrimage”. Many come with friends and collectively participate in the pilgrimage. We found that similar to religious pilgrimage, marketplace pilgrimage is structured by varied rituals, including anticipation, participation, evaluation, and dissemination. Marketplace pilgrimage involves efforts and bodily labor that transform consumer desire and consumption experience into an ephemeral form of so-

Advances in Consumer Research 796 Volume 48, ©2020 Stigma and Taint in Advertising Cagri Yalkin, Middle Eastern Technical University, North Cyprus Campus, Turkey Finola Kerrigan, University of Birmingham, USA

EXTENDED ABSTRACT real sense”. On the other hand, excessive working hours express Advertising sustains the contemporary system and people’s de- themselves as precarity: Goz, 41, account manager, illuminates the sire to consume despite recurring financial crises and the ongoing cli- precarious aspect of her job “..i’m trying to manage the feelings of mate crisis (e.g. Pradham 2009, UNFCC 2019). Advertisers remain everybody… trying to write emails that will not damage the relation- key cultural intermediaries and taste makers (Cronin 2004, Cayla ship with the client” and Fah, creative, seconds: “we do the ugly job… and Eckhardt 2008). The conduct of Cambridge Analytica, widely so many people’s feelings and irrelevant and irrational thoughts to covered in the press and through the documentary The Great Hack manage 24/7, via email, via whatsapp, via everything else that I feel (2019), have highlighted the ability of advertising in manipulating like I am just wasting the time”. Do also explains “you can’t just not all areas of public life. Corporate scandals have been blamed on take part in a tender, although it may not lead anywhere, you work on the advertising industry and its professionals (e.g. Business Insider it for months, concurrently with other tenders too”. 2016). Consumption’s direct link with climate crisis (see Carlson- The preliminary data analysis reveals that the advertisers from Kanyama and Gonzalez 2014), stereotyping and racism (see Thomas various ages and job specialisms view their job as stigmatized. In and Jones, 2019) thus require to further critical engagement with the particular, they have a dual view of their job as stigmatized and pre- advertising industry. carious; where precariousness feeds into their perceptions of taint. Previous literature has often focused on creativitiy in adver- This study extends the previous literature on creative identities (e.g. tising (e.g. Nyilasy, Kreshel, and Reid 2012), creative identities Hackley and Kover 2007, Gotsi 2010) by focusing on a middle to (Hirschman 1989, McLeod et al. 2009), gender in advertising prac- high income morally tainted line of work – advertising. Previous tice (Nixon and Crewe 2004, Windels and Lee 2012), advertisers’ studies focused on physically and socially stigmatized jobs (e.g. perceptions of gender portrayals in advertisements and ethics (Za- Dick 2005, Simpson et al. 2014), here, we extend the inquiry into the yer et al. 2015) and managing creatives (Gotsi et al. 2010, Hackley moral arena. Limitations of this study are that it is a snapshot in time and Kover 2007, Hirschman 1989). In particular, Nixon and Crewe and it is bound by its context. For example, in light of the Covid-19 (2004) and Hackley and Kover (2007) illustrate how advertisers ne- crisis, have the advertisers view of their jobs worsened? Future work gotiate competing identities. Here, we study how advertisers view could seek to investigate whether this perception also holds across their own jobs in light of the recent global state of affairs. other settings in the Global south and in the Global North and what Stigma and taint are associated with tasks of dubious virtue coping strategies advertisers cultivate to manage taint. (e.g., exotic dancer, bill collector, pawnbroker) or with jobs that are “deceptive, intrusive, confrontational, or that otherwise defy norms REFERENCES of civility” (Ashforth & Kreiner, 1999, p.415) (e.g., paparazzo, tele- Ashforth, Blake E. and Glen E. Kreiner (2014), “Dirty work and marketer). As such, advertising can easily be classified as a stigma- dirtier work: Differences in countering physical, social, and tized job as it is about managing sometimes deceptive perceptions moral stigma. Management and Organization Review, 10(1), that may not necessarily be based on the actual properties and ben- 81-108. efits of products/services/brands. There is limited literature related to Cadwalladr, Carole (2017), “The great British Brexit robber: how how the occupation holders themselves who are in morally stigma- our democracy was hijacked,” https://www.theguardian.com/ tized jobs view their jobs (see Zayer et. Al 2015 and Drumwright and technology/2017/may/07/the-great-british-brexit-robbery- Kamal 2016 for exceptions). If stigma and taint reduce occupational hijacked-democracy on March 3rd 2020. prestige (see Ashforth & Kreiner 2014), what are the ways in which Carlsson-Kanyama, Annika and Alejandro D. González (2009), advertisers view their profession and themselves? “Potential contributions of food consumption patterns to This on-going qualitative study draws on data from Turkey, an climate change”, The American journal of clinical nutrition, emerging economy with a workforce of 28 million. The studies on 89(May), 1704S-1709S. the advertising industry in the Global South are still limited. Excep- Chávez, Christopher (2012), “Hispanic agencies and profits of tions are Kaptan (2013), Cayla and Eckhardt (2008), and Drum- distinction: An examination of the advertising industry as a wright and Kamal (2016). Here, in-depth interviews with 20 adver- field of cultural production,” Consumption Markets & Culture, tisers are completed. A maximum variation purposeful sampling is 15(September), 307-325. used, where age, department within the agency, tenure and type of Cronin, Anne M. (2004), “Regimes of mediation: advertising agency (network vs. local boutique) is varied. The data are analysed practitioners as cultural intermediaries?” Consumption by the authors using a hermeneutic approach (Thompson 1989) and Markets & Culture, 7(December), 349-369. interpretation followed the guidelines of Miles and Huberman (1994) Dick, Penny (2005). Dirty work designations: How police officers and Spiggle (1994). account for their use of coercive force. Human relations, Our emergent findings point to a nuanced understanding of ad- 58(11), 1363-1390. vertisers’ own work as stigmatized and precarious, and the emerging Drumwright, Minette E. and Sara Kamal (2016), “Habitus, doxa, circumstances under which they may or may not feel the pressure and ethics: Insights from advertising in emerging markets in this brings. In particular, the advertisers view their jobs as stigma- the Middle East and North Africa,” Consumption Markets & tized and precarious. They view their job to be stigmatized to the Culture, 19(June), 172-205. degree of being immporal because they create artificial needs. Dar, a Ewen, Stuart (1976). Captains of Consciousness: Advertising and 40 year old senior creative, says: “afterall it is about selling people the Roots of Consumer Consciousness, McGraw-Hill: NY. crap they do not need… people look at me like I am evil”. Do, 40, is of a similar opinion “..it is what it is. It is not helping anyone in any

Gotsi, Manto, Andriopoulos, Constantine, Lewis, Marianne W. and Nyilasy, Gergely, Kreshel, Peggy J., and Leonard N. Reid (2012), Amy E. Ingram (2010), “Managing creatives: Paradoxical “Agency practitioners, pseudo-professionalization tactics, approaches to identity regulation,” Human Relations, and advertising professionalism,” Journal of Current Issues & 63(June), 781-805. Research in Advertising, 33(August), 146-169. Hackley, Chris & Arthur J. Kover (2007), “The trouble with Simpson, Ruth, Hughes, Jason, Slutskaya, Natasha and Maria creatives: Negotiating creative identity in advertising Balta (2014). Sacrifice and distinction in dirty work: agencies,” International Journal of Advertising, 26(January), men’s construction of meaning in the butcher trade. Work, 63-78. employment and society, 28(5), 754-770. Haug, Wolfgang F. (1986), Critique of commodity aesthetics: Thomas, Kevin D. and Naya Jones (2019), “Critical reflexivity: appearance, sexuality and advertising in capitalist society. Teaching about race and racism in the advertising classroom,” Minneapolis: University of Minnesota Press. Advertising & Society Quarterly, 20(2), doi:10.1353/ Hirschman, Elizabeth C. (1989). Role-based models of advertising asr.2019.0013. creation and production. Journal of Advertising, 18(June), Tuncay Zayer, Linda and Catherine A. Coleman (2015), 42-53. “Advertising professionals’ perceptions of the impact of McLeod, Charlotte, O’Donohoe, Stephanie and Barbara Townley gender portrayals on men and women: A question of ethics?. (2009), “The elephant in the room? Class and creative careers Journal of Advertising, 44(July), 1-12. in British advertising agencies,” Human Relations, 62(July), UNFCC (2019). Fashion for Global Climate Action, accessed at 1011-1039. https://unfccc.int/climate-action/sectoral-engagement/fashion- Nixon, Sean and Ben Crewe (2004), “Pleasure at work? Gender, for-global-climate-action consumption and work‐based identities in the creative Windels, Kasey and Wei-Na Lee (2012), “The construction of industries,” Consumption Markets & Culture, 7(June), 129- gender and creativity in advertising creative departments,” 147. Gender in Management: An International Journal, 27(November), 502-519. Time in the Past is Real and Honest: The Role of Nostalgia in Enhancing Authenticity Shiyu Yang, University of Illinois at Urbana-Champaign, USA Jia Chen, University of Illinois at Chicago, USA Yu-Wei Lin, University of Illinois at Urbana-Champaign, USA

EXTENDED ABSTRACT Across six studies, we examined how nostalgic consumptions Authenticity, typically defined as being “real”, “genuine”, or enhanced experienced authenticity. We first demonstrated that prod- “true” (Dutton 2003; Sharpley 1994; Taylor 1991; Vannini and Fran- ucts that made consumers feel authentic (vs. products that consumers zese 2008), is a desirable state for individuals (Gino, Kouchaki, and liked) were more nostalgic (Study 1 N=63). Building on this finding, Galinsky 2015) and a cherished attribute of products (Newman and we further tested the causal effect of nostalgia on experienced au- Dhar 2014). Many recent social criticisms argue that our society is thenticity using both controlled (Study 2 N=107) and freely recalled experiencing an authenticity deficit because of the overflow of mean- consumptions (Study 3 N=190). Moreover, we identified the un- ingless market offerings (Boyle 2004). To overcome this deficit, derlying mechanism by showing that nostalgic consumptions could people increasingly search for authenticity in brands, products, and remind consumers of their intrinsic nature, which in turn led to in- experiences (Arnould and Price 2000; Beverland 2005; Morhart et al. creased experienced authenticity (Study 3). We also found that self- 2015). As a result, authenticity has “…overtaken quality as the pre- concept clarity moderated the effect, such that the positive relation- vailing purchasing criterion” (Gilmore and Pine 2007, 5) and become ship between nostalgia and experienced authenticity was stronger “one of the cornerstones of contemporary marketing . . .” (Brown, among consumers with ambiguous self-views (Study 4 N=51). We Kozinets, and Sherry 2003, 21). In response to consumers’ craving then applied the findings to a purchase decision context and showed for authenticity, marketers are enthusiastic in cultivating products that highlighting the importance of authenticity would lead consum- and services that are experienced as authentic (Goldman and Papson ers to favor nostalgic products over novel ones (Study 5 N=159). 1996; Grayson and Shulman 2000). Finally, applying natural language processing to analyzing over 5000 It is therefore important to ask the following questions: From real-world products on Kickstarter, we showed that invoking nostal- consumers’ perspective, what kind of products can connect them with gia when pitching for a product could potentially make the product a real, honest and genuine self-view (i.e., self-authenticity)? From appear more authentic (Study 6 N=52997). marketers’ perspective, how to make their products experienced by Taken together, our research contributed to the literature on au- consumers as real, honest, and genuine (i.e., product authenticity)? thenticity and nostalgia by unveiling the relationship between them, These two questions are interrelated, as prior research has shown that the underlying mechanism, and the boundary condition. Moreover, consuming authentic products can increase self-authenticity (Hahl, our findings have implications for marketers who want to appeal to Zuckerman, and Kim 2017); also, a product that connects consumers consumers with authentic market offerings and consumers who are with their true self is perceived to be symbolically authentic (Morhart in search for authenticity. Also, to the extent that authenticity is a et al. 2015). We therefore ask the following question encompass- positively valenced psychological state, the current research also ing both self and product authenticity: how to increase experienced contributes to the literature on consumer well-being. authenticity in consumptions? We posit that inducing nostalgia in consumptions can lead consumers to experience a greater level of REFERENCES authenticity. Arnould, Eric J. and Linda L. Price (2000), “Questing for Self A plethora of research has documented the benefits of nostalgia, and Community,” The why of consumption: Contemporary such as social connectedness (Wildschut et al. 2010), meaningfulness perspectives on consumer motives, goals and desires, 1(1), (Routledge et al. 2012), empathy (Zhou et al. 2011), and patience 140. (Huang, Huang, and Wyer Jr 2016). A majority of consumer research Baldwin, Matthew, Monica Biernat, and Mark J. Landau (2015), has shown that nostalgia toward a brand could increase brand prefer- “Remembering the Real Me: Nostalgia Offers a Window to the ence, respect, trust ( Merchant and Rose 2013; Muehling and Sprott Intrinsic Self,” Journal of personality and social psychology, 2004), and the likelihood of purchase and recommendation (Nam et 108(1), 128–47. al. 2016). However, as far as we know, there is limited research on Beverland, Michael (2005), “Brand Management and the Challenge how nostalgia may influence consumers’ experience of authenticity. of Authenticity,” Journal of Product & Brand Management, Drawing upon research linking past autobiographical memories 14(7), 460–61. with intrinsic self-focus (Baldwin, Biernat, and Landau 2015) and Boyle, David (2003), Authenticity Brands, Fakes, Spin and the Lust pure self-view (Gino and Desai 2012), we argue that nostalgic con- for Real Life, philpapers.org. sumptions will offer a window to one’s original self, a self that is Brown, Stephen, Robert V. Kozinets, and John F. Sherry (2003), real, honest and true, a self that preserves one’s intrinsic nature as a “Teaching Old Brands New Tricks: Retro Branding and the person that is not yet tainted by any external constraints. The recol- Revival of Brand Meaning,” Journal of marketing, 67(3), lection of past experiences can set the tone for understanding and 19–33. anchoring individual identity. Following these arguments, we predict Gilmore, James H. and B. Joseph Pine (2007), Authenticity: What that consumptions that have been woven into consumers’ personal Consumers Really Want, Harvard Business Press. histories are likely to be perceived as embodiment of their true self Gino, Francesca., & Desai (2012), “Memory lane and morality: (i.e., self-authenticity); a logical corollary is that, products that can How childhood memories promote prosocial behavior”. elicit nostalgic feelings are likely to be perceived as real, honest, and Journal of personality and social psychology, 102(4), 743. genuine (i.e., product authenticity). We therefore posit that inducing Gino, Francesca, Maryam Kouchaki, and Adam D. Galinsky nostalgia in consumptions (e.g., nostalgic descriptions or advertise- (2015), “The Moral Virtue of Authenticity: How Inauthenticity ments) can potentially enhance experienced authenticity. Produces Feelings of Immorality and Impurity,” Psychological science, 26(7), 983–96.

Goldman, Robert and Stephen Papson (1996), Sign Wars: The Newman, George E. and Ravi Dhar (2014), “Authenticity Is Cluttered Landscape of Advertising, Guilford Press. Contagious: Brand Essence and the Original Source of Grayson, Kent and David Shulman (2000), “Indexicality and the Production,” JMR, Journal of marketing research, 51(3), Verification Function of Irreplaceable Possessions: A Semiotic 371–86. Analysis,” The Journal of consumer research, 27(1), 17–30. Routledge, Clay, Tim Wildschut, Constantine Sedikides, Jacob Juhl, Hahl, Oliver, Ezra W. Zuckerman, and Minjae Kim (2017), “Why and Jamie Arndt (2012), “The Power of the Past: Nostalgia as Elites Love Authentic Lowbrow Culture: Overcoming High- a Meaning-Making Resource,” Memory , 20(5), 452–60. Status Denigration with Outsider Art,” American sociological Sharpley, R. (1994), “Tourism, Tourists,” & society. Huntington, review, 82(4), 828–56. Cambridgeshire: ELM. Huang, Xun (irene), Zhongqiang (tak) Huang, and Robert S. Wyer Taylor, Charles (1992), The Ethics of Authenticity, Harvard Jr. (2016), “Slowing down in the Good Old Days: The Effect University Press. of Nostalgia on Consumer Patience,” The Journal of consumer Vannini, Phillip and Alexis Franzese (2008), “The Authenticity of research, 43(3), 372–87. Self: Conceptualization, Personal Experience, and Practice,” Merchant, Altaf and Gregory M. Rose (2013), “Effects of Sociology Compass, 2(5), 1621–37. Advertising-Evoked Vicarious Nostalgia on Brand Heritage,” Wildschut, Tim, Constantine Sedikides, Clay Routledge, Jamie Journal of business research, 66(12), 2619–25. Arndt, and Filippo Cordaro (2010), “Nostalgia as a Repository Morhart, Felicitas, Lucia Malär, Amélie Guèvremont, Florent of Social Connectedness: The Role of Attachment-Related Girardin, and Bianca Grohmann (2015), “Brand Authenticity: Avoidance,” Journal of personality and social psychology, An Integrative Framework and Measurement Scale,” Journal 98(4), 573–86. of consumer psychology: the official journal of the Society for Zhou, Xinyue, Tim Wildschut, Constantine Sedikides, and Kan Consumer Psychology, 25(2), 200–218. Shi (2012), “Nostalgia: The Gift That Keeps on Giving,” The Muehling, Darrel D., David E. Sprott, and David E. Sprott (2004), Journal of consumer research, 39(1), 39–50. “THE POWER OF REFLECTION: An Empirical Examination of Nostalgia Advertising Effects,”Journal of advertising, 33(3), 25–35. Nam, Jiyeon, Yun Lee, Nara Youn, and Kyoung-Min Kwon (2016), “Nostalgia’s Fulfillment of Agentic and Communal Needs: How Different Types of Self-Concepts Shape Consumer Attitudes toward Nostalgia,” Journal of Consumer Behaviour, 303–13, http://dx.doi.org/10.1002/cb.1568. The Moderating Effect of Regulatory Focus on Consumer Preferences for Price Discounts and Bonus Packs Jun Yao, Macquarie University, Australia Di Wang, Queensland University of Technology, Australia Brett A.S. Martin, Queensland University of Technology, Australia

EXTENDED ABSTRACT in the promotion-orientation condition. These results indicate that In today’s increasingly competitive marketplace, firms deploy a regulatory fit mediates only the effect of prevention orientation on range of sales promotion tactics to encourage purchases by provid- preference for promotion format, but not the effect of promotion ori- ing consumers with extra benefits or incentives. The sales promotion entation on preference. The direct effect in the promotion-orientation literature suggests that monetary tactics, such as price discounts, and condition can be explained by consumers’ inherent preference for a non-monetary tactics, such as bonus packs, distinctively influence bonus pack over a price discount, due to the prospect theory-based consumers’ purchase behaviors. Bonus packs which offer more prod- framing conceptualization. uct volumes are generally preferred by consumers over price dis- Study 2 was designed to replicate the findings shown in Study counts which offer lower prices (Diamond and Sanyal 1990; Mishra 1 in a field setting. In addition, instead of using percentages, we used and Mishra 2011), because the former tend to be framed as pure gains an equivalent absolute value in the two promotion formats. The study while the latter tend to be framed as reductions in losses according adopted a 2 (promotion format: discount vs. bonus pack) × 2 (regula- to prospect theory’s value function (Kahneman and Tversky 1979). tory orientation: promotion vs. prevention) between-participants de- We propose that regulatory orientation (Higgins 2000) sign. We collaborated with a medium-sized supermarket to conduct plays a moderating role in consumers’ preferences between bonus this study over a one-week period. We set up a table in the entrance/ packs and price discounts. Specifically, the default preference for a exit area of the supermarket and put a sign on the table reading, “By bonus pack does not hold for individuals with a prevention orien- completing our marketing survey, you can receive a $5 voucher.” tation. A price discount provides an opportunity to reduce the eco- Participants were asked to read three sentences either related to their nomic cost and thus aligns with a prevention orientation to minimize hopes and aspirations (i.e. “I want to do grocery shopping,” “It is losses. In turn, individuals with a prevention orientation are expected enjoyable to do grocery shopping,” “Doing grocery shopping is to experience a regulatory fit between the prevention orientation and interesting.”) or to their duties and obligations (i.e. “I ought to do the price discount, resulting in a preference for a price discount over grocery shopping,” “It is an obligation to do grocery shopping,” “Do- a bonus pack. ing grocery shopping involves a feeling of duty.”). Participants were In contrast, a bonus pack provides incremental product benefits then told that today the supermarket was running an in-store promo- and thus aligns with a promotion orientation to maximize gains, in- tion for a store gift voucher. The promotion was framed as either a dividuals with a promotion orientation are expected to experience a discount (i.e. “you buy a $12 gift voucher and get $2 off, you pay regulatory fit between the promotion orientation and the bonus pack, $10 cash”) or a bonus pack (i.e. “you pay $10 cash to get $2 bonus, resulting in a preference for a bonus pack over a price discount. Such you get a $12 gift voucher”). Participants indicated their choice of a preference is in line with the default preference in favor of a bonus whether to purchase the gift voucher. back based on differential framing (i.e., a bonus pack is framed as Results showed an interaction effect. In the prevention-orienta- a pure gain while a price discount is framed as a reduction in loss). tion condition, 58.5% of participants in the discount condition and However, we argue the regulatory fit cue may not be as accessible as 40.0% of participants in the bonus pack condition chose to purchase the differential framing cue to influence the preference. the gift voucher. In contrast, in the promotion-orientation condition, We present two studies, including a field study, to test the pro- 40.6% of participants in the discount condition and 58.5% of partici- posed effects. Study 1 investigates the joint effect of regulatory ori- pants in the bonus pack condition chose to purchase the gift voucher. entation and promotion format on purchase intention and the mediat- Again, participants in the prevention-orientation condition felt the ing role of regulatory fit. We recruited 387 participants on Mturk. discount promotion was right more than the bonus pack. Those in The study adopted a 2 (promotion format: discount vs. bonus pack) × the promotion-orientation condition felt no difference in terms of 2 (regulatory orientation: promotion vs. prevention) between-partici- how right participants felt in the discount and bonus pack conditions. pants design, with two product replicates to enhance generalizability. Similar to Study 1, results indicate that regulatory fit mediates only Results demonstrated that neither promotion format nor regu- the effect of prevention orientation on preference for promotion for- latory orientation had main effects, but there was a significant in- mat, but not the effect of promotion orientation on preference. teraction effect. Participants in the prevention-orientation condition One limitation of our research is that regulatory orientation was were more likely to purchase the product when the promotion was primed across two studies. To enhance the robustness of our findings, in the discount format than in the bonus pack format. Participants further research could assess consumers’ chronic regulatory orienta- in the promotion-orientation condition were more likely to purchase tion to see whether the moderating effect of regulatory orientation on the product when the promotion was in the bonus pack than in the consumer preferences for promotion tactics still hold. discount format. The analysis of regulatory fit revealed a significant interaction effect such that participants in the prevention-orientation REFERENCES condition felt the discount was more “right” than the bonus pack. Diamond, W. D. & Sanyal, A. (1990). The effect of framing on the There was no difference in regulatory fit between the discount and choice of supermarket coupons,” in M. E. Goldberg, G. Gorn, bonus pack in the promotion-orientation condition. Results of the & R. W. Pollay (Eds.), Advances in Consumer Research (Vol. mediation analysis showed a significant moderated mediation effect. 17, pp. 488-493). Provo, UT: ACR. Specifically, the indirect effect was significant while the direct ef- Higgins, E. T. (2000). Making a good decision: Value from fit. fect was not significant in the prevention-orientation condition. The American Psychologist, 55(11), 1217-1230 . indirect effect was not significant, but the direct effect was significant

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis Mishra, A., & Mishra, H. (2011). The influence of price discount of decision under risk. Econometrica, 47 (2), 263-291. versus bonus pack on the preference for virtue and vice foods. Journal of Marketing Research, 48(1), 196-206. The Devil You Know: The Problem of Equilibrium Goals Siyuan Yin, The Wharton School, University of Pennsylvania, USA Cait Lamberton, The Wharton School, University of Pennsylvania, USA

EXTENDED ABSTRACT to control conditions, people in the high social- and cognitive-cost Why do we stay in unsatisfying relationships with brands (Er- conditions were less likely to switch away from equilibrium goals dem and Keane 1996, Erdem, Keane, and Sun 2008), hold onto los- (bhigh social cost = -7.41, p < .001; bhigh cognitive cost = -3.04, p = .038) and peo- ing stocks even when they have little likelihood of improving (Odean ple in the low social- and cognitive-cost conditions were more likely

1998, Shefrin and Statman,1985), take extra flights to stay with air- to do so (blow social cost = 4.13, p = .005; blow cognitive cost = 5.19, p < .001). lines that delay us regularly, or remain in frustrating jobs or inter- Thus, manipulating both types of switching costs can independently personal relationships? We conceptualize such “devil you know” shift individuals’ likelihood to persist in equilibrium goals. situations as equilibrium goals, in which we prioritize status quo To replicate this finding in a more consequential setting, Study 3 maintenance even when changing our state would be more likely to (n = 228) established a “work partnership” with cost-related interven- offer positive utility than remaining. tions in the lab. The experiment followed a two (cognitive cost: high Three studies have been completed to substantiate the unique vs. low) by two (social cost: high vs. low) between-subject design. nature of equilibrium goals. We first identify consistent characteris- Participants worked together in “teams” to obtain points, such that tics of equilibrium goals across domains. Study 1 (n = 460) followed they could maximize their final bonus that would be determined by a five (scenario: job, relationship, diet, budget, cancer treatment) by their own points and their team points. All participants were assigned two (goal: equilibrium vs. disequilibrium) between-subject design. the role of “receiver” and received points allocated by their paired For example, participants in the equilibrium job condition read: deciders who was ostensibly on the same team. We held constant “Suppose that your current job is boring and involves tedious work point allocations across conditions such that an equilibrium goal situ- and you have been working in the position for 5 years. You could ation was created for all participants: the decider was systematically change jobs, but you choose to continue to go to work and do well underallocating to participants, such that little utility could be gained enough to stay in this job.” By contrast, participants in the disequilib- by staying in the partnership. After three rounds of point allocations, rium job condition read the same information with reversed actions. participants read, in the high (low) cognitive (social) cost condition, Participants reported that in both goal conditions they were pur- “The average rating of cognitive (social) costs of switching from pre- suing a goal to an equivalent extent and had a similar level of past in- vious surveys was 7 (3) out of 10.” Participants then had the opportu- vestment in pursuit of that goal. However, disequilibrium goals (e.g. nity to switch partners who were on an opposing (the same) team in changing jobs) involved a higher focus on change, stronger motiva- the high (low) social cost conditions. Despite the fact that participants had evidence that their decid- tion, and higher anticipated utility than equilibrium goals (bfocus on change er was unlikely to offer a high chance to maximize their outcomes, = -.023, p < .001; bmotivation = -.026, p = .002; butility = -.033, p < .001). By contrast, remaining in equilibrium as opposed to disequilibrium only 58.7% decided to switch partners, leaving 41.3% pursuing the equilibrium goal. Participants in the low social-cost conditions were goals required greater self-control (bself-control = .039, p < .001). Thus, it appears that equilibrium goals create an undesirable situation for significantly more likely to switch partners, thus releasing an equi- consumers: while their self-control costs are high, their motivation librium goal, than those in the high social-cost conditions (blow social cost and utility are low. = 1.85, p < .001). Participants in the low cognitive-cost conditions To prompt consumers to release their equilibrium goals, we were marginally more likely to switch partners than those in the high examined literature on maintenance goals (Yang, Stamatogianna- cognitive-cost conditions (blow cognitive cost = .91, p = .08). As before, no kis, and Chattopadhyay 2015) and consumer lock-in (Johnson and interaction was observed (binteraction = -.70, p = .30). Russo 1984, Johnson, Bellman, and Lohse 2003, Zauberman 2003). We show that manipulating perceived goal-switching costs ef- This literature suggested that consumers in equilibrium goals may fectively shifted people away from equilibrium goals in both mar- perceive higher switching costs than necessary. If this is the case, ketplace and collaboration contexts. Our findings contribute to tra- lowering perceptions of social and cognitive switching costs should ditional goal and motivation literature with a new framework that weaken consumers’ hold on equilibrium goals. helps understand the mechanisms of equilibrium goal-pursuit behav- Study 2 (n = 1782) tested this prediction using a three (cogni- ior. Future research will extend these findings to other consequential tive cost: control, high, low) by three (social cost: control, high, low) equilibrium goals, in the hope that we can learn more about ways to by three (scenario: airline, retailer, partner relationship) between- enhance consumers’ goal pursuit and overall well-being. subject design. For example, in the airline scenario, the equilibrium goal related to staying in an airline reward program that regularly REFERENCE provided less-than-optimal service. We manipulated the perceived Erdem, Tülin. and Michael P. Keane (1996), “Decision-Making social and cognitive switching costs as high or low via cost-related under Uncertainty: Capturing Dynamic Brand Choice messages. No cost information was provided in the control condi- Processes in Turbulent Consumer Goods Markets,” Marketing tions. Participants reported how likely they would be to end their Science, 15(1), 1–20. relationship and start over with another airline or do what it takes Erdem, Tülin, Michael P. Keane, and Baohong Sun (2008), “A to maintain their relationship from 0 (definitely keep status) to 100 Dynamic Model of Brand Choice When Price and Advertising (definitely end the relationship). Signal Product Quality,” Marketing Science, 27(6), 1111–25. A mixed-effects model predicting this likelihood with cognitive- Johnson, Eric J. and J. Edward Russo (1984). “Product Familiarity and social-cost conditions as predictors and scenarios as random ef- and Learning New Information,” Journal of Consumer fects revealed significant main effects of cognitive costs (F(2, 1771) Research, 11(1), 542-550. = 14.17, p < 0.001) and social costs (F(2, 1771) = 32.65, p < .001), but no significant interaction (F(4, 1771) = .47, p = .75). Compared

Johnson, Eric J., Steven Bellman, and Gerald L. Lohse (2003), Yang, Haiyang, Antonios Stamatogiannakis, and Amitava “Cognitive lock-in and the power law of practice,” Journal of Chattopadhyay (2015), “Pursuing Attainment versus Marketing, 67 (April), 62-75. Maintenance Goals: The Interplay of Self-Construal and Odean, Terrance (1998), “Are Investors Reluctant to Realize Their Goal Type on Consumer Motivation,” Journal of Consumer Losses?” Journal of Finance, LIII (5), 1775–1798. Research, 42 (1), 93–108. Shefrin, Hersh and Heir Statman (1985), “The disposition to Zauberman, Gal (2003), “The Intertemporal Dynamics of sell winners too early and ride losers too long: Theory and Consumer Lock-In,” Journal of Consumer Research, 30 evidence,” Journal of Finance, 40 (July), 777–790. (December), 405–419. Advances in Consumer Research (Volume 48) / 805 The Effect of Social Identity Conflict on Consumers’ Planning Time Horizons Yiqi Yu, Guanghua School of Management, Peking University, China Ying Zhang, Guanghua School and Management, Peking University, China

EXTENDED ABSTRACT Using the Midlife in the United States survey, study 4 found Consumers often hold multiple social identities (Crisp and that a less clear sense of self was associated with shorter planning Hewstone 2007). For example, immigrants are simultaneously iden- horizons among 2654 respondents. Further, study 5 directly mea- tified with their country of origin and the new country; students may sured our proposed mechanism, after the using the same manipula- hold identities associated with the multiple schools that they have tion of conflict as Study 2. Results suggested that experiencing social attended. When these identities are simultaneously activated and the identity conflict increased participants’ preferences for a short-term contents of these identities are perceived incompatible, consumers video streaming membership (MConflict = 3.05 (1.50) vs. MNo conflict = experience social identity conflict (Hirsh and Kang 2016). 3.79 (1.59), p = .011). Furthermore, bootstrapping analyses revealed Despite the ubiquity and importance of consumers experiencing that an unclear sense of self mediated this effect (95% CI = [-0.6973, such conflict, research that examines whether and how this experi- -0.0031]). ence may affect consumer choices remain relatively scarce. In this In study 6, we employed a 2 (social identity conflict: conflict vs. research, we explore how social identity conflict might alter one’s no conflict) ×2 (bolstering a clear sense of self: yes vs. no) between- perception of self, and in turn affect his/her planning time horizon subject design and randomly assigned 231 participants to one of the in consumption decisions. Specifically, we predict that experiencing four conditions. We manipulated identity conflict as study 5. Partici- social identity conflict leads consumers to focus more on here and pants were then asked to describe a personal possession (Morrison now, thus showing greater preference for short-term products and and Johnson 2011). Specifically, to bolster a clear sense of self, par- service plans. This preference shift occurs because experiencing so- ticipants were asked to describe a possession that reflects one of his/ cial identity conflict undermines a clear sense of self (Campbell et al. her important social identities. In the control condition, participants 1996), which lays the foundation of a temporally stable self (Dunlop were asked to describe a possession which they found useful. We 2017; Jiang, Chen, and Sedikides 2019). A less enduring sense of self then recorded their choice of financial products. We found that when in the temporal dimension, in turn, leads to less willingness to plan a clear sense of self was not bolstered, participants who experienced for the distant future. social identity conflict showed greater preference for the short-term Study 1 surveyed 225 Mturkers to directly test the relationship financial products than did those who experienced no such conflict between social identity conflict and planning horizon. Participants (MConflict = 2.98 (2.05) vs. M No conflict = 3.92 (2.06), p = .019). However, indicated the extent to which and how often they felt that their mul- when participants experienced a clear sense of self, the difference tiple social identities were in conflict, after we introduced what was disappeared (MConflict = 3.83 (2.05) vs. MNo conflict = 3.75 (1.39), p = social identity conflict. We then used three items to measure partici- .642). pants’ planning time horizons (e.g. “I focus more on the proximate To summarize, by showing consumers who experience social future when I am making the decisions”) and found that social iden- identity conflict are more likely to adopt short planning time hori- tity conflict was positively correlated with shorter planning time ho- zons, our work makes important theoretical contributions and yields rizons (b = 0.128, p = .001), controlling for demographic variables. practical implications for consumer decision-making and well-being. Study 2 employed a lab study where 94 participants first read a diary written by a college student about her experience after mov- REFERENCES ing to a different city from her hometown and imagined themselves Campbell, Jennifer D., Paul D. Trapnell, Steven J. Heine, Ilana being the student. Participants were randomly assigned to a conflict M. Katz, Loraine F. Lavallee, and Darrin R. Lehman (1996), condition where the diary described that the college student often “Self–Concept Clarity: Measurement, Personality Correlates, experienced identity conflict because of the different norms of the and Cultural Boundaries,” Journal of Personality and Social two places, or a no conflict condition, where the diary described that Psychology, 70 (1), 141–56. the student comfortably navigated between the two identities and no Crisp, Richard J., and Miles Hewstone (2007), “Multiple Social conflict experience was mentioned. Next, in an ostensibly unrelated Categorization,” Advances in experimental social psychology, task, participants were told to write down their future plans. We in- 39, 163-254. tentionally kept the instructions vague and did not specify any time Dunlop, William L. (2017), Self-Concept Clarity, Springer, Cham. period for the “future”. Their answers were coded based on the time Hirsh, Jacob B. and Sonia K. Kang (2016), “Mechanisms of frame and we found that, participants in the conflict condition wrote Identity Conflict: Uncertainty, Anxiety, and the Behavioral down more plans about proximate future than did those in the no Inhibition System,” Personality and Social Psychology conflict condition: 71.8% vs. 47.6%,p < 0.001. Review, 20(3), 223-44. Study 3 moved to school identity. A total of 104 graduate stu- Jiang, Tonglin, Zhansheng Chen, and Constantine Sedikides dents were randomly assigned to either a social identity conflict (2019), “Self–Concept Clarity Lays the Foundation for Self– condition where they imagined that they were watching a champi- Continuity: The Restorative Function of Autobiographical onship match between their undergraduate alma mater and their cur- Memory,” Journal of personality and social psychology. rent school, or a friendly exhibition match between the two teams. Morrison, Kimberly Rios and Camille S. Johnson (2011), “When Then, participants were asked to indicate their preferences between What You Have Is Who You Are: Self-Uncertainty Leads two financial products: a one-year CD offering 1.12% APY and a Individualists to See Themselves in Their Possessions,” three-year CD offering 1.28%. Results showed that participants in Personality and Social Psychology Bulletin, 37 (5), 639–51. the conflict condition were less likely to choose the three-year product (MConflict= 2.14 vs. MNo conflict= 3.09, p = .018; higher value denoted preference for the long-term product). How Mere Manager Response Influences Product Evaluations Vincentia Yuen, University of Miami, USA Michael Tsiros, University of Miami, USA

EXTENDED ABSTRACT the two MR presence conditions (Mmere = 4.70, F(1, 255) =2.431, As firms gradually notice the impact of electronic word of p=.016 and Mcontent = 4.75, F(1, 255)=2.776, p=.006) were signifi- mouth (eWOM) on business performance, online review manage- cantly higher than that in the control condition (Mcontrol = 4.34). As ment has become more prominent in marketing practice. Rather than expected, we found no significant difference in product evaluations subtly intervene in eWOM by promotional reviewing (Mayzlin, Do- between the two MR presence conditions (F(1, 255)=.351, p = .726). ver, and Chevalier 2014), firms start to actively engage in eWOM Study 2 (N= 273) tested the hypothesized mechanism (H2) with conversations by responding to customer reviews to restore customer a 2(MR: presence vs. control) between-subjects design. We mea- satisfaction (Gu and Ye 2014) and influence future reviews (Proser- sured product evaluations (α = .969), trust (α= .960), and perceived pio and Zervas 2017) via a manager response (MR) feature. recovery (e.g., “I expect the [firm] to try to make up for any nega- While MRs are, in theory, directed at reviewers, Murphy (2019) tive issues mentioned in the reviews,” α = .843) respectively. As pre- finds that 97% of review readers also attend to MRs when they evalu- dicted, the presence of MR increased product evaluations (4.08 vs. ate products online. For this reason, even if managers are responding 3.79; t(271) = 1.94, p =.053). A serial mediation analysis (PROCESS to prior failures, the impact of such highly observable MRs should model 6; Hayes 2018) confirmed our hypothesized process (b = .24; no longer be confined to the focal recipients who wrote the initial 95% CI from .0611 to .4273) that the presence of MR made consum- reviews but should also extend to all customers who observe such ers perceive the firm will address the issues mentioned in the review responses. In the current research, we build on service recovery lit- (b = .48, t = 2.65, p <.01), which enhanced customer trust toward the erature and hypothesize that the mere presence of MR —with no firm (b = .64, t = 18.73, p < .001) and thus boosted product evalua- knowledge of the actual content—can positively shape consumer tions (b = .78, t = 19.87, p< .001). responses. Lastly, study 3 (N= 231) further examined the hypothesized It is well-established that a successful recovery from service underlying process by directly manipulating perceived recovery failure could restore customer satisfaction and enhance repurchase through review valence with a 3 (MR presence: negative reviews intentions (Grewal, Roggeveen, and Tsiros 2008). When firms re- vs. both positive and negative reviews vs. control) between-subjects cover in a way that builds customer trust, customers are more likely design. If, as we suggest, MR increases product evaluations by mak- to make confident predictions about the firms’ future recovery being the firm seem more likely to address the negative issues raised haviors and thus commit themselves to ongoing relationships (Kim, in reviews and lead to higher trust, then having managers also reply Kim, and Kim 2009). As trust could also be acquired by observing or to positive reviews should weaken this positive effect from the MR. learning from previous interactions (e.g., conflicts) that the firm has We measured product evaluations (α = .898), trust (α= .916), and had with others in similar situations (Holmes 1991; Tax et al. 1998), perceived recovery (α = .836) in a randomized order. As predicted, we argue that previous recovery episodes should enhance prospec- planned contrasts revealed that product evaluations in the condition tive customer trust and enhance product evaluations on the basis of with MRs to negative reviews (M = 4.94) were significantly higher a positive expectation of solving any potential future failures. For- than that in the condition with MRs to both positive and negative mally, we hypothesize: reviews (M = 4.45; F(1, 228) = 2.573, p=.011) and the control condition (M = 4.43; F(1, 228) = 2.618, p=.009). We found no significant Hypothesis 1: The mere presence (vs. absence) of manager difference in product evaluations between the latter two conditions response to negative reviews enhances product (F(1, 228)= .104, p = .917). The serial mediation analysis (PROCESS evaluations. model 6) again confirmed that perceived recovery and customer trust mediated the effect of MR on product evaluations (b = .27; 95% CI Hypothesis 2: The effect of the presence of manager response from .0902 to .4711). Overall, H1 and H2 were supported. on product evaluations in H1 is mediated by the In sum, this research contributes to the eWOM literature by increase in perceived recovery and subsequent showing how firms could proactively intervene in the impact of on- customer trust toward the firm. line reviews on consumer responses. To date, no research has exam- Three studies examine the hypotheses and the proposed psy- ined the psychological processes underlying the impact of MRs on chological mechanism. In all studies, participants first read identi- prospective customer evaluations. Hence, by extending key findings cal product information along with actual hotel or restaurant reviews in the service recovery literature, this research yields important in- (where we only manipulated the presence of MR) and responded to sight into the process explanation and shows how the mere signal of measures on seven-point Likert scales. It is worth noting that we ex- response to a negative prior review could influence product evalua- amine the impact of the mere presence of MR, rather than the effects tions. From a substantive perspective, this research offers clear im- driven by the MR content itself. plications to practitioners, particularly those under pressure to man- Study 1 (N= 258) tested the central hypothesis that the mere age their online reputation. presence of MR enhances product evaluations with a 3(MR: mere REFERENCES presence vs. with content vs. control) between-subjects design. As Grewal, Dhruv, Anne L. Roggeveen, and Michael Tsiros (2008), one might question how our findings might compare to the more “The Effect of Compensation on Repurchase Intentions in popular practice of having MR response present publicly, we also Service Recovery,” Journal of Retailing, 84 (4), 424-34. examine a condition in which (a widely used) response content (acknowledgment and apology) is shown in this study. After reading the stimulus, participants indicated their product evaluations (α = .940). Results supported our prediction (H1). Product evaluations in

Gu, Bin, and Qiang Ye (2014), “First Step in Social Media: Murphy, Rosie (2019), “Local Consumer Review Survey: Online Measuring the Influence of Online Management Responses Reviews Statistics & Trends,” https://www.brightlocal.com/ on Customer Satisfaction,” Production and Operations research/local‐consumer‐review‐survey/. Management, 23 (4), 570-82. Proserpio, Davide, and Georgios Zervas (2017), “Online Reputation Holmes, John G. (1991), “Trust and the Appraisal Process in Close Management: Estimating the Impact of Management Relationships,” Advances in Personal Relationships, 2, 57- Responses on Consumer Reviews,” Marketing Science, 36 (5), 104. 645-65. Kim, Taegoo Terry, Woo Gon Kim, and Hong-Bumm Kim (2009), Tax, Stephen S., Stephen W. Brown, and Murali Chandrashekaran “The Effects of Perceived Justice on Recovery Satisfaction, (1998), “Customer Evaluations of Service Complaint Trust, Word-Of-Mouth, and Revisit Intention in Upscale Experiences: Implications for Relationship Marketing,” Hotels,” Tourism Management, 30 (1), 51-62. Journal of Marketing, 62 (2), 60-76. Mayzlin, Dina, Yaniv Dover, and Judith Chevalier (2014), “Promotional Reviews: An Empirical Investigation of Online Review Manipulation,” American Economic Review, 104 (8), 2421-55. Does Space Speak Quantities? The Impact of Interspace on Product Size Estimates Yuli Zhang, Stockton University, USA Hyokjin Kwak, Drexel University, USA Marina Puzakova, Lehigh University, USA Charles R. Taylor, Villanova University, USA

EXTENDED ABSTRACT Study 2 aims to replicate the findings of Study 1. Participants Academic research has recognized the general significance of were asked to imagine going to a store and considering buying mixed space in a retail setting and marketing communications by devot- nuts and then were presented with a shelf image displaying the ficti- ing attention to the study of the downstream consequences of the tious brand of mixed nuts. We manipulated interspatial (non-inter- amount of space allocated to products in a display (e.g., shelf display, spatial) product display similar to that of Study 1, with only slight POP/advertising display) (Dreze, Hoch, and Purk 1994; Parker and modifications. Study 2A (n = 160) kept the number of products and Lehmann 2011). However, this past streams of work have primarily the amount of white space between conditions constant. Study 2B focused on either variation in the amount of space dedicated to a (n = 87) held the size of the display between conditions constant. As product display itself (e.g., more or less amount of display space) or expected, participants estimated that one product package contained the alteration in the space-to-product ratio (e.g., high or low intersti- a higher number of servings of mixed nuts in the interspatial (vs. tial space) in a display. Our research significantly differs from these non-interspatial) display condition (Study 2A: Minterspatial = 11.38 vs. prior investigations in that we examine the impact of the existence Mnon-interspatial = 9.26; t(158) = 1.99, p < .05; Study 2B: Minterspatial = 6.54 of space between products (hereafter referred to as “interspace”) as vs. Mnon-interspatial = 5.28; t(85) = 2.09, p < .05). Mediation analysis opposed to the overall space surrounding products. In our research, revealed a significant indirect path from interspatial product display we hold the amount of display space constant and demonstrate that to product size estimates through attention (Study 2A: 95% CI = the existence or nonexistence of interspace in a product display has [.0525, 1.0726]; Study 2B: 95% CI = [.0319, .8221]). unique consequences for judgments of product size. We posit that Study 3 provides further process evidence by manipulating indian interspatial (vs. a non-interspatial) product display is likely to in- viduals’ attention level. Participants (n = 155) first viewed a one-tier crease consumers’ perceived product size. shelf image displaying HK shampoo. Interspatial (non-interspatial) Prior research on visual processing demonstrates that space be- product display was manipulated in a manner similar to that of Study tween entities facilitates an entity’s recognition or increases its de- 1. At the same time, participants’ level of attention was manipulated lineation from the background (Qiu, Sugihara, and von der Heydt via a distraction task. In the distraction (no distraction) condition, a 2007; Rayner 2011). In this regard, prior research points out that popular song was played (no music was played) while participants objects are more readily recognizable when they are separated by were exposed to the HK shampoo display. After viewing the image, background space than when they are connected to each other or participants then indicated their perceived product size. As expected, overlapping (Qiu et al. 2007). In situations where distinction from when participants were not distracted, they estimated the product to the background is enhanced, this causes increased salience of the contain more quantity in the interspatial (vs. non-interspatial) display target stimulus (Nothdurft 2000), which in turn affects the allocation condition (how many washes per bottle: Minterspatial = 58.55 vs. Mnon- of attention to that stimulus (Bacon and Egeth 1994). Overall, prior interspatial = 37.71; F(1, 150) = 9.15, p < .01; how long to use up (days): research suggests that when space isolates objects in a visual field Minterspatial = 68.81 vs. Mnon-interspatial = 47.97; F(1, 150) = 6.74, p < .02). from the background they are more salient and, thus, attract more at- However, when participants were distracted, there was no significant tention. We further theorize that such increased attention to products difference in perceived product size between the two conditions (Fs in an interspatial (vs. a non-interspatial) product display subsequent- < 1). ly results in greater size perceptions. According to prior research, an From a theoretical point of view, our research contributes to the object appears to be larger and/or of greater quantity when people’s literature by identifying a novel spatial factor, the mere existence of attention is automatically directed to this object (Folkes and Matta interspace in a display, that has unique consequences for judgments 2004). Hagtvedt and Brasel (2017) also show that objects with highly of product size. Our work also provides significant managerial guide- saturated colors attract more attention and, thus, are perceived to be lines for the use of space between products in a product display. larger. Taken together, these findings point out that increased attention devoted to an object prioritizes processing mental resources to REFERENCES make size estimation better. Bacon, William F. and Howard E. Egeth (1994), “Overriding Study 1 examines the core effect of interspace on product size stimulus-driven attentional capture,” Perception & estimates and the underlying mechanism. Participants (n = 103) were Psychophysics, 55 (5), 485-496. instructed to view a print ad which included an image of three cans Dreze, Xavier, Stephen J. Hoch, and Mary E. Purk (1994), “Shelf of a fictitious brand of wafers. In the interspatial (non-interspatial) management and space elasticity,” Journal of Retailing, 70 (4), display condition, the cans were presented with a perceptually recog- 301-326. nizable space (vs. no space) between them. Participants then reported Folkes, Valerie and Shashi Matta (2004), “The effect of package their perceived product size. As predicted, participants in the inter- shape on consumers’ judgments of product volume: attention spatial (vs. non-interspatial) display condition estimated a greater as a mental contaminant,” Journal of Consumer Research, 31 (2), 390-401. quantity of wafer rolls per product package (Minterspatial = 23.06 vs. Hagtvedt, Henrik and S. Adam Brasel (2017), “Color saturation Mnon-interspatial = 17.21; t(101) = 2.21, p < .05). Mediation analysis revealed a significant indirect path from interspatial product display to increases perceived product size,” Journal of Consumer product size estimates through attention (95% CI = [.2317, 3.9667]). Research, 396-413. The effect of interspatial product display on product size estimates through product popularity and quality were not significant.

Nothdurft, Hans-Christoph (2000), “Salience from feature contrast: Qiu, Fangtu T., Tadashi. Sugihara, and Rüdiger von der Heydt additivity across dimensions,” Vision Research, 40 (10-12), (2007), “Figure-ground mechanisms provide structure for 1183-1201. selective attention,” Nat Neurosci, 10 (11), 1492-1499. Parker, Jeffrey R. and Donald R. Lehmann (2011), “When shelf- Rayner, Alan David (2011), “Space cannot be cut—why self- based scarcity impacts consumer preferences,” Journal of identity naturally includes neighbourhood,” Integrative Retailing, 87 (2), 142-155. Psychological and Behavioral Science, 45 (2), 161-184. The Pursuit of Pragmatic Truth at the Cost of Literal Truth Charles Y. Zhang, University of California, Riverside, USA Norbert Schwarz, University of Southern California, USA

EXTENDED ABSTRACT interview would be used as part of a call for international charity In communication, listeners expect speakers’ utterances to be donations; the other half were told that the interview would be used truthful and speakers expect their utterances to be accepted as true. to inform the coordinators of international rescue teams. All partici- Philosophers and linguists consider it a core principle of cooperative pants were asked to 1) choose which interview should be issued, and conversation that speakers should only say what they believe to be 2) identify which statement was more accurate. For donors, it is ir- true (Grice, 1975). Grice (1989) noted that his other maxims of con- relevant if the damage pertains to the seaport or airport as long as the versation (relation, quantity, and manner) “come into operation only severity of damage is comparable. In contrast, whether the seaport on the assumption that this maxim of quality is satisfied” (p. 27). or the airport is destroyed is crucial for rescue teams, who need to In the present research, we challenge this view and introduce a approach the island. Accordingly, while most participants in both distinction between literal and pragmatic truth. To tell a literal truth, conditions considered “a warehouse and the airport were destructed” a speaker should avoid saying anything she does not believe to be less accurate, they chose to issue this statement to potential donors part of the fact. To tell a pragmatic truth, a speaker should provide but not to rescue teams. information that is as close to the fact as possible on the variable that Study 4 provides process evidence that a speaker’s choice of ut- is most relevant to the audience. Although the goals of achieving terance is driven by the extent to which she believes that the utterance literal and pragmatic truth are often aligned, they can conflict in con- is pragmatically true. Participants learned that a medicine could lead texts where the fact cannot be entirely and precisely communicated, to side effects such as dry mouth and heart attack. The official mis- e.g., e.g., due to a lack of time, technical difficulties, or the recipi- spoke the side effects either as “dry mouth and liver failure,” “heart ent’s lack of background knowledge. We propose that, in these cases, attack and itchy skin,” or “dry mouth and itchy skin.” All participants speakers will, and will be expected to, form utterances that maximize were given a choice of broadcasting the original version or to trim off the pragmatic truth, even at the cost of literal truth. In the present the second side effect of the corresponding statement. They also rated research, we pitch statements that are literally true but pragmatically the severity of the four side-effects. When the speaker mentioned false against statements that are literally false but pragmatically true. dry mouth and liver failure, most participants chose to broadcast this We demonstrate the existence of situations in which pragmatically (literally false) statement over the trimmed version. Their decision to true statements are considered more appropriate than literally true do so was mediated by their perceived severity of the side effect. In statements and discuss the implications of the findings for marketing contrast, significantly fewer participants chose to broadcast one of contexts. the other literally false statements (heart attack and itchy skin or dry In study 1, participants were told that a medicine was recently mouth and itchy skin). found to have side effects, including dry mouth and heart attack. An In sum, the present research highlights the distinction between official misspoke in a radio interview that “the medicine may cause pragmatic and literal truth. When literal truth conflicts with prag- side effects such as dry mouth and liver failure.” The editor’s only matic truth under conditions of communication constraints, speakers choices were to either broadcast this interview or a trimmed version choose the pragmatically true statement over the literally true state- of it: “the medicine may cause side effects such as dry mouth.” The ment. original statement is literally false but pragmatically truer because liver failure and heart attack are comparable in terms of their se- REFERENCES verity. As predicted, the majority of participants considered that the Grice, H. P. (1975). Logic and conversation. In Speech acts (pp. trimmed version better reflected the facts, but chose the full version 41-58). Brill. to broadcast. Grice, H. P. (1989). Studies in the Way of Words. Harvard In study 2, participants imagined being a burglar whose task University Press. it is to warn a partner of any approaching cops. They had agreed to Horn, L. (1984). Toward a new taxonomy for pragmatic inference: position their arms horizontally if no cop is in sight and vertically if Q-based and R-based implicature. Meaning, form, and use in a cop is approaching. Participants were told that, although no cop context: Linguistic applications, 11, 42. was approaching, they spotted a hidden surveillance camera when Johnson, E. J., & Goldstein, D. (2003). Medicine. Do defaults save their partner was about to commit the crime. In this situation, the lives?. Science (New York, NY), 302(5649), 1338-1339. gesture that indicates an approaching cop could serve as a warning Levinson, S. C., Stephen, C., & Levinson, S. C. (2000). even though no cop was seen; it would hence serve the pragmatic Presumptive meanings: The theory of generalized purpose while being literally false. As expected, most participants conversational implicature. MIT press. chose to hold the arms vertically, thus conveying a warning, while Thomson, J. J. (1985). The trolley problem. The Yale Law Journal, also agreeing that the explicit prerequisite for this signal (cop ap- 94(6), 1395-1415. proaching) was not met. An utterance can be relevant to different listeners in different ways. Statements that are pragmatically true to some listeners may not be pragmatically true to others, which should affect speakers’ choice of utterances. In study 3, participants learned that a typhoon hit a Pacific island and destroyed a warehouse and the only seaport. Two local officials were interviewed. One stated that a warehouse was destroyed and the other stated that a warehouse and the only airport were destroyed. Half of the participants were told that the

Advances in Consumer Research 810 Volume 48, ©2020 Maximizing Strengths or Fixing Weaknesses? Evaluatees Tend to Fix Weaknesses More Than Evaluators Appreciate Xiaoying Zheng, Nankai University, China Yanping Tu, University of Florida, USA Jingting Yuan, China Construction Bank, China Jing Xu, Peking University, China

EXTENDED ABSTRACT sensitivity to potentially threatening information and negative out- Everything has its strengths and weaknesses. For instance, a comes: apprehensive individuals detect negative information faster restaurant may receive an above average rating on food quality, but (Bradley, Mogg and Lee 1997), interpret ambiguous stimuli in a an average rating on service; a job candidate may have better analyti- more negative way (Calvo and Castillo 1997; Eysencket al. 1991), cal skills than writing skills. When given an opportunity to improve, and perceive higher risks of negative events (e.g., failing an exam, do people (e.g., the restaurant manager, the job candidate) prioritize catching a disease) (Butler and Mathews 1987; Mitte 2007). Another maximizing strengths (e.g., food quality, analytical skills) or fixing stream of research that informs our proposition is about how moti- weaknesses (e.g., service, writing skills)? Moreover, often times, vational orientation (prevention vs. promotion focus; Higgins 1997) these decisions are made to cater to others’ preferences. For instance, changes thinking. Specifically, apprehension is associated with a the restaurant manager would like to win consumers’ heart; the job prevention-focus (Klenk, Strauman, and Higgins 2011), a “vigilant, candidate hopes to get recruiters’ vote. Do the restaurant manager negative-outcome-focused motivational state,” which makes people and the job candidate make improvement decisions optimally? attend to negative information and work towards avoiding negative We examine decision contexts like these, which involve an outcomes (rather than achieving positive outcomes). evaluatee/seller who provides products for others to evaluate or se- Thinking guides action. If follows that, when evaluatees worry lect, and an evaluator/buyer who evaluates or selects among products about the negative prospects of being rejected or disliked, they act on provided by evaluatees. Evaluatees often make decisions to improve the aspect that is more likely to cause it. Because weaknesses (rather than strengths) are the more likely causes of the negative prospects, an original product (PO) that has strengths and weaknesses. Should they prioritize maximizing strengths (MS) or fixing weaknesses evaluatees will act on weaknesses to ensure that they provide no (FW)? Sometimes the answer is obvious: if the weakness falls below grounds for rejection. the acceptable threshold, then FW will be imperative; if the return for However, evaluators, the other party in the evaluatee-evalua- improving a certain attribute is higher, then effort should be devoted tor dyad, do not experience such apprehension and are less likely to that attribute. However, in situations where the original product to adopt a negative focus when choosing a product. Compared to (1) is reasonably good such that the weakness is not alarmingly unac- evaluatees, their attention is less likely to be drawn to the weakness ceptable and (2) the return of improving either the strength and the attribute, and, consequently, they would not appreciate evaluatees’ weakness is similar, what should evaluatees do? To illustrate (figure effort of FW as much. 1), consider, in a 5-star evaluation system, a product at the original In sum, our central hypothesis is that, when making improvement decisions, evaluatees tend to fix weaknesses more than evalu- state PO (4.3, 4.0), scoring 4.3 on attribute A (i.e., the strength) and 4.0 on attribute B (i.e., the weakness). Suppose the evaluatee can ators appreciate, because they are in a more apprehensive state and improve the rating of one (and only one) attribute by 0.3. Should they engage in more negative thinking. improve attribute A or B? To answer this question, our empirical approach is to present Study 1: Job Candidate vs . Recruiter the final products of the two strategies to evaluators, ask evaluators Study 1 (N = 173) tests our basic hypothesis in the job mar- to choose, and assess the match. Specifically, maximizing strengths ket context, using a one-factor two-level (evaluatee vs. evaluator) will lead to a product with more polarized attributes, some extremely between-participants design. In the evaluatee condition, participants imagined themselves strong and some relatively weak (PMS); fixing weaknesses will lead to hunting for a job and preparing for two equally important job en- a product with more balanced attributes, all moderately strong (PFW). trance tests (A and B). They learned that the total score of each test We measure whether evaluators prefer PMS or PFW, as a proxy of their appreciation of evaluatees’ effort of MS or FW. was 150, and that, in a mock test, they scored 120 for test A and 100 for test B. Participants further supposed that they could improve the Evaluatees Prefer Fixing Weaknesses More Than score on either test by 20 points, and indicated whether they would Evaluators Appreciate like to improve test A (the strength) or test B (the weakness). We propose that evaluatees may choose FW more than evalu- In the evaluator condition, participants imagined being a re- ators appreciate, for the following reasons. To begin with, it is well cruiter and choosing between two candidates. They learned that the documented that people feel apprehensive (also described as anx- total score of each test was 150, and that, candidate 1 scored 140 on ious, uneasy, or worried in literature) when being judged by others test A and 100 on test B (i.e., the end state of the evaluatee had they (i.e., evaluation apprehension; Rosenber 1969). Feeling apprehen- chosen to improve test A), whereas candidate 2 scored 120 on both sive, evaluatees may consider more about the negative prospects of tests (i.e., the end state of the evaluatee had they chosen to improve their decision (e.g., being rejected or disliked by evaluators), rather test B). Evaluators indicated which candidate they would like to re- than the positive prospects (e.g., being selected or favored by evalu- cruit. ators). Our proposition is informed by two streams of research. Evaluatees who chose to work on test A [test B] were coded as First, mood alters cognition (Mayer et al. 1992; Johnson and Tver- preferring MS [FW]; evaluators who selected candidate 1 [candidate sky 1983): people attend to mood-congruent information, generate 2] were coded as preferring MS [FW]. A Chi-Square test showed that mood-congruent thoughts, and expect mood-congruent event to be evaluatees preferred FW (63%; 53 out of 84) more than evaluators more likely to happen. More specifically, apprehension heightens

(45%; 40 out of 89; χ2(1) = 5.73, p = .017, odds ratio = 2.09), lending the lower rated aspect of these apartments” and “I made the decision support to our basic hypothesis. based on the lower rated aspect of these apartments.” in the evaluator condition). (1 = completely disagree, 7 = completely agree). Study 2: Vacation Spot Manger vs . Consumer Replicating the findings in Studies 1 and 2, the results showed Study 2 (N = 402) aimed to explore the main effect in a market- that evaluatees (hosts) preferred FW (49%, 98 out of 201) more than er-consumer setting. We also examined whether (1) the importance evaluators (guests) (30%, 60 out of 200; χ2(1) = 14.77, p < .001; odd of the weakness attribute and (2) whether PFW received perfectly bal- ratio = 2.22). Evaluatees (M = 4.11, SD = 1.40) experienced a more anced scores on both attributes had an impact. It used a 2 (evaluatee negative mentality than evaluators (M = 3.31, SD = 1.31) (t(399) = vs. evaluator) x 2 (PFW has perfectly balanced attribute scores: yes vs. 5.94, p < .001), and adopted a more negative focus (M = 4.36, SD = no; conceptual replication factor) x 2 (importance of the weakness 1.77) than buyers (M = 3.39, SD = 1.55) (t(399) = 5.84, p < .001). attribute: high [food]vs. low [entertainment]; conceptual replication The planned serial mediation (PROCESS Model 6; Hayes 2013) factor) between-participants design. confirmed that negative mentality and negative focus mediated the In the evaluatee conditions, participants imagined themselves impact of role (evaluatee vs. evaluator) on preference for FW (95% being a manager of a vacation spot with food and entertainment (ab- CI = [.1185, .3355]). breviated to entmt.) as the two major attributes. They can improve the rating of one dimension by 0.2. In the evaluator conditions, par- Study 4: Car Manufacturers vs . Car Buyers ticipants imagined themselves making a choice between two vaca- Although for simplicity and clarity, we conducted our analysis tion spots which differed in ratings on food and entertainment. In- and most empirical work in a two-products and two-attributes con- formation on attribute values shown to evaluatees and evaluators text where evaluatees choose to improve one attribute, our theory respectively is in appendix A. could make predictions in multiple-products and multiple-attributes We coded participants’ preferences for MS/FW. The choice contexts where evaluatees can choose to improve multiple attributes. share for FW broken by cells is shown in appendix B. A 2 (evaluatee In such contexts, MS means increasing the variance across attributes vs. evaluator) x 2 (PFW has perfectly balanced attribute scores: yes (i.e., making attribute values more polarized), whereas FW means vs. no) x 2 (importance of the weakness attribute: high [food]vs. low reducing the variance across attributes (i.e., making attribute values [entertainment]) logistic regression with choice (0 = MS, 1 = FW) more balanced). as the dependent variable yielded neither a three-way interaction nor Study 4 offered such an empirical test, modeling real decisions any two-way interactions (all ps > .350). Thus, we collapsed data on by car manufactures (evaluatees) and buyers (evaluators). We linked the latter two factors and ran a chi-square test, finding that evaluatees (1) a unique proprietary data set of multi-attribute rating information preferred FW (72%, 149 out of 208) marginally more than evaluators (fuel consumption, handling, horsepower, exterior styling, interior (63%, 122 out of 194; χ2(1) = 3.50, p = .061, odds ratio = 1.49). design, comfort, and space), scraped from a leading customer re- Collectively, studies 1 and 2 established the main effect, as well view website, of the top 20% major cars and (2) their corresponding as demonstrated its generalizability. monthly sales volumes in the Chinese market, between January 2014 and October 2019 (70 months). Study 3: Testing the Underlying Mechanism Car manufacturers make decisions to improve car models on a Study 3 (N = 401) aimed to test the proposed mechanism: eval- yearly basis, and thus we coded attribute rating at year-level. Their uatees consider more about the negative prospects of their decisions, choice of FW [MS] would manifest as a decrease [an increase] of leading them to act on the more likely cause (i.e., weakness). In addi- variance across attributes (Var) from time t-1 to time t. We note the tion, we generalized the effect to a new domain: Airbnb, with a host possibility that the whole industry might be shifting to produce cars and a guest as the seller-buyer dyad. This study used a one-factor with more balanced (or more polarized) attributes due to external two-level (evaluator vs. evaluatee) between-participants design. factors, and thus standardized Var based on the average variance of In the evaluator condition, participants assumed the role of an all the car models in the given year (SVar). We reason that, if car Airbnb host who was considering whether to improve the amenities manufacturers adopt a FW strategy, then a car with an above-average or the interior design (i.e., two focal attributes of their apartment). variance last year should have a larger decrease in variance this year. Participants learned that their apartment was currently rated 4.0 and Put formally, our dependent variable was the difference of standard- 4.3 on interior design and amenities, respectively, on Airbnb’s 5-star ized variance between year t and year t-1 for car i (), and our main rating system. They further learned that they had the resources to predictor was standardized variance of car i in year t-1 (). We in- improve the customer rating of one, and only one, attribute by 0.3. cluded rating valence (), sales volume (), and price () of car i in year Participants then chose one attribute to improve. In the evaluatee t as control variables. We also incorporated fix effects for car () and condition, participants assumed the role of an Airbnb guest who was for time (). Equation is shown below. choosing between two apartments. Apartment A was rated 4.0 on interior design and 4.6 on amenities (i.e., PMS). Apartment B was rated 4.3 on both interior design and amenities (i.e., P ). 𝑆𝑆𝑆𝑆𝑆𝑆𝑟𝑟�� 𝑆𝑆𝑉𝑉𝑎𝑎𝑟𝑟�� 𝛽𝛽� � 𝛽𝛽�𝑆𝑆𝑆𝑆𝑆𝑆𝑟𝑟�� 𝛽𝛽�𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉�� 𝛽𝛽�𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉�� 𝛽𝛽�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃�� 𝛼𝛼� � 𝑓𝑓� �� FW If manufacturers choose FW [MS], then should be negative Afterward, participants indicated (1) whether they experienced [positive]. Our theory predicts that should be negative. Regression a negative mentality (“I was worried that guests may not like my results supported our theory (see appendix C). apartment” and “I felt uncertain about what kind of apartment would Car buyers make purchase decisions based on customer review be preferred by guests.” in the evaluatee condition; “I was worried information. Because we obtained monthly sales data, we coded that I may not like the apartment I chose” and “I felt uncertain about attribute rating data at monthly level. Our dependent variable was what kind of apartment I prefer.” In the evaluator condition) and (2) sales volume of car i at month t (), and our main predictor was the whether they adopted a negative focus (“I was worried that guests accumulative variance at month t-1 of car i (). Note that we used ac- may focus more on the lower rated aspect of my apartment” and “I cumulative reviews, because when a buyer reads reviews at time t, was worried that guests may not choose my apartment because of its she gets reviews from time 1 to time t. Our control variables included lower rated aspect.” in the evaluatee condition; “I focused more on accumulated valence from time 1 to time t (), accumulated volume Advances in Consumer Research (Volume 48) / 813

from time 1 to time t (), and price at time t of car i (). In addition, (Weaver, Garcia and Schwarz 2012; Woolley and Fishbach 2018). to control for the impact of the change in the entire industry, we Last but not the least, our research offers clear recommendations

factored in accumulated industry-level variance (Var_industryit) by for managers that they should give prioritize maximizing strengths averaging variances in all the accumulated reviews for all cars from rather than fixing weaknesses. time 1 to time t. We also incorporated fix effects for car () and for time (). Equation is shown below. REFERENCES Bower, G. H. (1981). “Mood and Memory,” American      Psychologist, 36, 129-148. �� 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆  Johnson, Eric J., and Amos Tversky (1983), “Affect, �� 𝑉𝑉𝑉𝑉𝑉𝑉_𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 � �𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 � �𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 � �𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 Generalization, and the Perception of Risk,” Journal of � �𝑉𝑉𝑉𝑉𝑉𝑉_𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖�� If consumers appreciate FW [MS], then should be negative Personality and Social Psychology, 45(1), 20-38. [positive]. Our theory predicts that should be positive. Regression Klenk, Megan M., Timothy J. Strauman, and E. Tory Higgins results supported our theory (see appendix D). (2011), “Regulatory Focus and Anxiety: A Self-regulatory Model of GAD-depression Comorbidity,” Personality and General Discussion Individual Differences, 50(7), 935-943. Across three lab studies and two secondary data sets collected Rust, Roland T., Anthony J. Zahorik, and Timothy L. Keiningham in both the US and China, we document a robust effect that evalu- (1995), “Return on Quality (ROQ): Making Service Quality atees choose to fix weaknesses more than evaluators appreciate. We Financially Accountable,” Journal of Marketing, 59(2), 58-70. further identify the underlying psychological process: evaluatees ex- Weaver, Kimberlee, Stephen M. Garcia, and Norbert Schwarz perience apprehension and focus more on negative prospects. (2012), “The Presenter’s Paradox,” Journal of Consumer Our research adds to literature on improvement decision, which Research, 39(3), 445-460. previously only focused on establishing methods to calculate fi- Woolley, Kaitlin, and Ayelet Fishbach (2018), “It’s About Time: nancial returns post hoc (Rust, Zahorik and Keiningham 1995). It Earlier Rewards Increase Intrinsic Motivation,” Journal of enriches literature on how roles shape preferences and decisions Personality and Social Psychology, 114(6), 877-885. Donating Cash or Donating Objects? How Donation Type Determines Donation Likelihood Xiaozhou (Zoe) Zhou, IE Business School, Spain Dilney Gonçalves, IE Business School, Spain

EXTENDED ABSTRACT or in in-kind. A third ad was created which instructed people in its Charities have been concerned about in-kind donations (vs. description to imagine how their cash donation could be used to get cash donations) shipped to disaster-affected areas displacing more face masks. Chi-square tests revealed that the likelihood of clicking urgent shipments and incurring high logistic costs (Brooks 2018; on the in-kind advertisement was higher than the cash advertisement Piper 2019). However, a pre-test (30 participants) revealed that more (χ2 = 16.89, p < .01). People were also more likely to click the cash than half of the participants believed that making in-kind donations advertisement that encouraged imagination (χ2 = 6.13, p < .02) com- (i.e., food/supplies) was a better way to help disaster-affected area. pared to that which simply requested cash. We did not observe any In this research, we sought to provide evidence for and explain the difference between in-kind and cash advertisement with imagination underlying mechanism of donors’ preference for in-kind donations. (χ2 = 1.07, p = .30). Research of imagery has recognized that stimulus plays an Study 3 examined the moderated mediation with mental imagi- important role in eliciting imagination (MacInnis and Price 1987; nal ability as a moderator. Participants were presented with a mostly Spears and Yazdanparast 2014). Moreover, researchers demonstrate identical charitable advertisement. The only difference was the type that stimulus with higher concreteness is better at inducing imagina- of donation the charity sought (in-kind or cash). We measured par- tion (Burns, Biswas, and Babin 1993; Yoo and Kim 2014). Drawing ticipants donation likelihood, contribution confidence (Macdonnell on previous findings, we propose that in-kind donations, which are and White 2015), and imaginal ability (α = .73, scale developed both tangible and concrete, are more likely to induce donors’ imagi- by Childers, Houston, and Heckler (1985)). A regression with donation about how their donation can be used. Since people have high- nation type, imaginal ability, and their interaction on contribution er confidence about the likelihood of an event happening if they have confidence (the mediator) revealed a significant interaction effect (β previously imagined the same situation (Anderson 1983; Carroll = 1.31, p = .01). The moderated mediation through contribution con- 1978), we hypothesize that asking for in-kind donations will increase fidence was also significant (indirect effect = .99, SE = .42; 95% CI donors’ contribution confidence about their donation’s helpfulness, [ .2172, 1.8927]). Participants with high imaginal ability indicated thus increasing donation intention. higher donation likelihood for in-kind donation as these increased Consumers vary in their imaginal processing ability (Ernest their contribution confidence. 1977). Studies that examined the effect of imagery-eliciting strate- Study 4 examined moderated mediation with construal-level as gies on advertising found that advertisements that instructed consum- a moderator. Mental construals were manipulated by social distance. ers to imagine were more effective to participants with high imaginal The charity was either collecting donations (cash or in-kind) for peo- ability (Burns et al. 1993). We thus predict that donors with high ple in the participants’ hometown (low construal) or Indonesia (high imaginal ability will show a stronger preference for in-kind dona- construal). A regression with donation type, construal level, and their tions by better using the stimulus (the in-kind donations) to imagine interaction on contribution confidence revealed a significant -inter an end result. Low imaginal ability donors, in contrast, have limited action effect (β = .96, p = .01). The overall moderated mediation ability to use stimulus to imagine a similar situation. Consequently, was also supported (indirect effect = .71, SE = .27; 95% CI [ .1929, we hypothesize that donors’ imaginal ability will moderate the effect 1.2589]). Participants with high mental construal indicated higher of donation type on donors’ perceived contribution confidence. contribution confidence and donation intention for in-kind donation. According to construal-level theory, low-level construal leads We did not observe differences in contribution confidence between people to construe events in a more concrete manner (Trope and donation types when participants use low construal. Liberman 2010). Since the ease of imagination increases as con- Past research mostly focused on the difference between cash do- creteness increases (Sherman et al. 1985), we expect that the ease of nation and volunteering (Liu and Aaker 2008). Our research fills the imagination increases as mental construal decreases. Thus, we hy- gap by examining the difference between cash and in-kind donation. pothesize that the effect of donation type on contribution confidence Moreover, our work adds to consumer imagination literature, partic- will be attenuated by low construal because low construal encour- ularly in the area of donation. We also provide practical implications. ages imagination regardless of donation type. Our studies suggest that nonprofits can create advertisements that In studies 1a and 1b participants were told to evaluate a situa- elicit imagination about how cash donations can be used to encour- tion where a charity, which was collecting donations for people who age cash donations. The results that construal levels can influence just suffered an earthquake, approached them at a supermarket. We people’s preferred donation type also provide guidance to nonprofits employed a between-subjects design in study 1a and an independent- who target a wide range of donors. There are situations where non- samples t-test revealed that donation likelihood was significantly profits would prefer physical donations, and sometimes nonprofits might want to collect in-kind donations from nearby donors and cash higher for in-kind donations (Mphysical = 3.94, SD = 1.11; Mcash = 3.43, SD = 1.35; t(159) = -2.60, p = .01). Study 1b used a within-subjects donation from long-distance donors. In this case, different advertise- design. A paired-samples t-test showed that participants still pre- ments could be designed. ferred in-kind donations (Mphysical= 3.78, SD = 1.27; Mcash = 2.98, SD = 1.30; t(79) = - 4.05, p < .001). REFERENCES Study 2 examined how people react to nonprofits’ ads that re- Anderson, Craig A. (1983), “Imagination and Expectation: quested either cash or in-kind donation by using Google Adwords The Effect of Imagining Behavioral Scripts on Personal to post charity advertisements. Two advertisements were created Influences,” Journal of Personality and Social Psychology, to raise donations for Chinese residents living in the epidemic area 45(2), 293–305. during COVID-19 and the advertisements differed only in the first Brooks, Julia (2018), “Want to Help after Hurricanes? Give Cash, headline where we manipulated the donation request as either in cash Not Diapers.,” The Conversation.

Burns, Alvin C, Abhijit Biswas, and Laurie A Babin (1993), “The MacInnis, Deborah J and Linda L Price (1987), “The Role of Operation of Visual Imagery as a Mediator of Advertising Imagery in Information Processing: Review and Extensions,” Effects,”Journal of Advertising, 22(2), 71–85. Journal of Consumer Research, 13(4), 473–491. Carroll, John S (1978), “The Effect of Imagining an Event on Piper, Kelsey (2019), “Why Disaster Relief Is so Hard,” Vox, Expectations for the Event: An Interpretation in Terms of https://www.vox.com/future-perfect/2018/10/15/17927688/ the Availability Heuristic,” Journal of Experimental Social disasters-cyclone-idal-relief-mozambique. Psychology, 14(1), 88–96. Sherman, Steven J, Robert B Cialdini, Donna F Schwartzman, Childers, Terry L., Michael J Houston, and Susan E Heckler and Kim D Reynolds (1985), “Imagining Can Heighten or (1985), “Measurement of Individual Differences in Visual Lower the Perceived Likelihood of Contracting a Disease: The Versus Verbal Information Processing,” Journal of Consumer Mediating Effect of Ease of Imagery,” Personality and Social Research, 12(2), 125–34. Psychology Bulletin, 11(1), 118–27. Ernest, Carole H (1977), “Imagery Ability and Cognition: A Critical Spears, Nancy and Atefeh Yazdanparast (2014), “Revealing Review,” Journal of Mental Imagery, 1(2), 181–215. Obstacles to the Consumer Imagination,” Journal of Consumer Liu, Wendy and Jennifer Aaker (2008), “The Happiness of Giving: Psychology, 24(3), 363–72. The Time-Ask Effect,”Journal of Consumer Psychology, Trope, Yaacov and Nira Liberman (2010), “Construal-Level Theory 35(3), 543–557. of Psychological Distance,” Psychoanalytic Review, 117(2), Macdonnell, Rhiannon and Katherine White (2015), “How 440–63. Construals of Money Versus Time Impact Consumer Yoo, Jungmin and Minjeong Kim (2014), “The Effects of Online Charitable Giving,” Journal of Consumer Research, 42(4), Product Presentation on Consumer Responses : A Mental 551–63. Imagery Perspective,” Journal of Business Research, 67(11), 2464–72. Can a Single Digit Make Consumers Eat Less? The Influence of Calorie Presentation Format on Consumers’ Unhealthy Food Consumption Li Zhou, Hunan Normal University, China Leilei Gao, Chinese University of Hong Kong, China Guowei Zhu, Hunan University, China

EXTENDED ABSTRACT negative information. Thus, for a given vice food item, we would ex- To alleviate the problem of obesity and promote consumers’ pect a round-number presentation of caloric information (as opposed well-being, many countries require food companies to provide nu- to precise presentation) to intensify consumers’ negative evaluations. tritional information on food packages (FDA 2016; European (EU) Finally, we expect people’s evaluative judgment of food items No 1169/2011). Yet, with regard to information on calorie content in to have a direct impact on their consumption intention such that particular, different countries have different regulations (Appendix negative (positive) evaluation of a given food item will decrease A). Take the original flavored Pringles potato chips as an example, (increase) consumption (e.g., Kivetz and Simonson 2002; Werten- companies selling in the EU market are required to present the calo- broch 1998). Therefore, it is expected that a round-number presenta- rie content of Pringles as a precise integer (154 calories per serving), tion of caloric information (as opposed to precise presentation) will whereas companies in the US market have to round the number of intensify consumers’ negative evaluations of unhealthy food items, calories to the nearest multiple of 10 (150 calories per serving). An and subsequently decrease consumption intention. We examined this interesting question that arises is, does this tiny difference in the pre- proposition in four studies. sentation of calorie information affect consumers’ food consumption intentions and actual behavior? Study 1a A round number is defined as any integer that is a multiple of Method 10 (where numbers that end in more zeroes are considered to be Participants (N=114, 45.6% female, Mage=36.65, SD=10.19) re- “rounder” than numbers that end in fewer zeroes). Consumers per- cruited on Amazon M-Turk were randomly assigned to one of two ceive numbers that end with non-zero digits to be more precise than conditions (round vs. precise calorie information) in a between-sub- round numbers (Thomas, Simon, and Kaidiyali 2010; Pena-Marin ject design. and Bhargave 2016). Research suggests that people process round All participants were shown the front and back package views numbers more easily (or fluently) than they process precise numbers of a brand of peanut butter flavored sandwich crackers, with either (Kettle and Häubl 2010; King and Janiszewski 2011; Wadhwa and round (200 calories) or precise (202 calories) calorie information dis- Zhang 2014, 2019). The heightened processing fluency associated played on the back-pack nutrition facts label (Appendix B). Partici- with round numbers is commonly attributed to the high frequency pants were then asked to estimate the number of crackers they would of usage of round numbers in daily life (Dehaene and Mehler 1992; like to eat (between 0 and 30). As hunger has been reliably shown to Fassbender et al. 2014. affect people’s food consumption level (Garg, Wansink, and Inman Prior research does not provide an obvious indication of wheth- 2007; Tangari et al. 2019), in all studies we also recorded partici- er round (vs. precise) presentation of calorie content is likely to pants’ hunger level (“How hungry are you at the moment?”; 1= not positively or negatively affect consumers’ perceptions of a particular at all, 7= extremely hungry), and included hunger as a covariate in food item. On the one hand, some studies suggest that processing flu- all the related analyses. ency can evoke positive affective reactions, which individuals might misattribute to the target stimulus, leading them to evaluate it more Results favorably (Winkielman and Cacioppo 2001; Winkielman et al. 2003; Consistent with our expectation, participants indicated that they Schwarz 2004). This account would suggest that presenting calorie would eat fewer crackers when calorie information was presented information in round (vs. precise) numbers might elicit more positive as a round number (M=8.67, SD=6.37) than as a precise number affect, leading to more positive evaluations of the food item. Yet, oth- (M=11.12, SD=8.8; F(1, 111)=4.37, p=.039, η2=.04).). er researchers downplay the role of positive affect triggered by processing fluency, and instead emphasize the general capacity of pro- Study 1b cessing fluency to magnify one’s judgment of a stimulus, whatever Study 1B sought to replicate the findings of Study 1A using a that judgment might be (Jacoby and Dallas 1981; Jacoby, Kelley, and different product and a different presentation method of calorie in- Dywan 1989; Witherspoon and Allan 1985). This stream of research formation. To assure that the effect was not only observed when the (Albrecht and Carbon 2014; Carbon and Albrecht 2016) suggests that precise number was bigger than the round number, Study 1B adopted a higher level of processing fluency can activate a clearer interpreta- two precise presentation conditions where the precise numbers were tion of the most salient source of the stimulus that is task-relevant at either slightly higher or slightly lower than the round number. Finally, the time of judgment, thus intensifying one’s original evaluations of in Study 1B, we measured the perceived healthiness and perceived that stimulus. In other words, under heightened processing fluency, calorie level of the unhealthy food item and examined whether it was an individual will evaluate positive attributes of the stimulus even these judgments that were magnified by the round presentation of more positively, and negative attributes more negatively (King and calorie numbers and subsequently caused the effect. Janiszewski 2011; Wadhwa and Zhang 2014). Method In the case of unhealthy foods, consumers usually have mixed Participants recruited via Amazon M-Turk (N=201, 46.8% evaluations. Such foods offer immediate hedonic enjoyment, but also female, Mage=36.79, SD=11.13) were randomly assigned to one of high numbers of calories that conflict with long-term goals of staying three conditions (round, precise-low, and precise-high) in a between- healthy or losing weight (e.g., Read, Loewenstein, and Kalyanara- subject design. man 1999; Wertenbroch 1998). When processing calorie information They were told to look at the package design of a newly launched of unhealthy foods, we suggest that consumers are likely to treat it as ice cream brand and indicate how much ice cream they would like to Advances in Consumer Research 816 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 817 eat at the moment (between 0 to 250 grams). The per-serving calo- Results and Discussion rie content was 200 calories, 196 calories, and 204 calories for the Replicating the findings of Studies 1A and 1B, we found an ef- round, precise-low, and precise-high conditions respectively (Ap- fect of calorie information roundness (F(1, 121)=5.03, p= .027, η2= pendix C). Next, they evaluated the perceived healthiness of the food .04). Participants ate significantly fewer chips when the calorie in- item (“In your opinion, how healthy is the ice cream?”; 1=Not at all formation was presented in a round number (M=5.33, SD=3.53) than healthy, 7=Very healthy)” and the perceived calorie level of the ice a precise number (M=7.2, SD=4.73). cream (“How do you perceive the calorie level of the ice cream?”; In addition, there was no difference in self-reported guilt as- 1=Extremely low, 7=Extremely high). sociated with eating chips (p = .332). Nonetheless, analysis on risk perception revealed an effect of calorie presentation condition (F(1, Results and Discussion 122)= 10.06, p= .002, η2= .08). Specifically, compared to present- As expected, there was a significant difference in consump- ing the potato chip’s calorie content in a precise number (M=4.30, tion intention across the three conditions (F(2, 197)=3.48, p=.03, SD=1.46), participants perceived greater health-related risks when η2= .033). Participants in the round-number condition intended to the calorie content was presented in a round number (M=5.14, consume less ice cream (M=101.46, SD=75.38) than those in the SD=1.49). precise-high condition (M=138.06, SD=80.88; F(1, 197)= 5.07, p= We also conducted a mediation analysis following the PRO- .025, η2= .03) and those in the precise-low condition (M=128.36, CESS macro (calorie presentation format à risk perception à con- SD= 73.68; F(1, 197)=5.33, p= .022, η2= .03). Participants in the two sumption volume; model 4, Hayes, 2018). The effect of calorie pre- precise-number conditions, however, did not differ in terms of their sentation on risk perception was significant (b= .65, SE= .17, t= 3.76, intended consumption level (F < .01, p = .966). p = .001, 95% CI= [.31, 1.00]), so was the effect of risk perception In addition, calorie presentation format did not affect the health- on food consumption (b= -.31, SE= .10, t=-3.18, p =.001, 95% CI= iness perception of the ice cream (p= .163), nor the perceived calorie [-.50, -.12]). Participants’ subjective judgment of risk factor fully level of the ice cream (p= .431). mediated the effect of calorie presentation on food consumption (b= In summary, Study 1A and Study 1B provide convergent evi- -.20, SE= .08, p = .017, 95% CI= [-.41, -.08]). dence that a round-number presentation of caloric information (as In summary, Study 2 replicated the findings of Study 1 with opposed to precise presentation) decreased consumption intention. consumers’ actual consumption behavior. In addition, Study 2 found However, the lack of between-condition difference in perceived un- that it was the risk judgment associated with consuming unhealthy healthiness and perceived calorie level of the ice cream suggests that food that was magnified when calorie information was presented as the effect was not caused by the amplification of these judgments. a round (vs. precise) number, and subsequently decreased consump- It is possible that they may not be the primary consideration when tion volume. consumers process calorie information on nutrition labels. STUDY 3 Study 2 Thus far, our findings are consistent with the conceptualization Study 2 was designed to replicate the findings of Study 1 with that processing fluency increases attention paid to calorie informa- real consumption behavior. In addition, we continued to examine the tion, which allows consumers to form a more negative interpretation underlying mechanism of the effect by measuring participants’ risk of the information, and subsequently decreases consumption inten- perception and guilt associated with eating unhealthy foods. tion. If this is the case, then the effect should be prominent among Method those who do not habitually pay attention to food calorie informa- Students from a large university (N=124, 51.6% female, tion, but attenuated when consumers’ default attention to food calorie information is high. We tested this hypothesis in Study 3. Mage=21.34, SD=2.01) participated in the lab study and were randomly assigned to two conditions (round vs. precise). As a cover sto- Method ry, participants were asked to taste a brand of newly launched potato Participants (N=211, 49.3% Female, Mage= 30.01, SD= 7.23) chips and provided their evaluations. Each participant was provided were randomly assigned to two conditions (calorie presentation: a cup of water to rinse their mouth, a plate containing 20 original fla- round vs. precise). We also measured participants’’ default attention vored potato chips, and a food label that presented the calorie infor- paid to food calorie information and treated as an individual differ- mation either as a round (2000 KJ/100g) or precise (2021KJ/100g) ence moderator. number. Participants were given five minutes to taste the chips and Participants were presented the front package view of a ficti- fill out a short questionnaire measuring their food evaluations (e.g., tious brand of potato chips with calorie information presented either taste, saltiness, crispiness). After the food-tasting task, all partici- as a round (800 KJ) or precise number (813KJ). Next, they indicated pants returned the questionnaire and any leftover food to the experi- the amount of potato chips they felt like to eat (0 ~ 70 grams). Then, menter and answered some additional questions. participants’ default attention levels were measured with three ques- Specifically, participants’ risk perceptions were assessed (“Will tions, “How often do you check calorie information when choos- eating chips on a regular basis increase the risk of weight gain?” and ing food items?” (1= never, 7= always); “How sensitive are you to “Will eating chips on a regular basis increase the risk of long-term calorie information when choosing food items?” (1= not at all sen- health problems?” 1= not at all, 7= definitely). The two measures sitive, 7= extremely sensitive); “To what extent do you base your were highly correlated and were combined for later analyses (r = snack purchase decision on the calorie information?” (1= not at all, .77). Participants were also asked, “How guilty do you feel to eat the 7= totally). These three measures were combined for later analyses chips in today’s study?” (1= not at all guilty, 7= extremely guilty). considering the high correlation among them (r = .93). After the participant left the lab, the experimenter counted each participant’s leftover potato chip pieces and recorded their consumption Results and Discussion amount to the nearest half chip. A bootstrapping method with PROCESS macro (Model 1, Hayes 2013) was conducted with food consumption as DV, calorie presentation format (0=precise, 1=round) and attention to calorie 818 / Can a Single Digit Make Consumers Eat Less? The Influence of Calorie Presentation Format on Consumers’ Unhealthy Food Consumption information as the IV. Results showed a significant main effect of Garg, Nitika, Brian Wansink and J. Jeffrey Inman (2007), “The caloric number roundness on food consumption (b=-6.82, SE=3.01, Influence of Incidental Affect on Consumers’ Food Intake,” t=-2.26, p= .025, 95% CI= [-12.75, - .88]) such that participants in- Journal of Marketing, 71 (1), 194-206. dicated that they would eat fewer potato chips when the calorie in- Hayes, Andrew F. (2013), Introduction to Mediation, Moderation, formation was presented as a round than precise number. In addition, and Conditional Process Analysis: A Regression-Based there was a significant interaction between calorie number round- Approach, New York, NY, US: Guilford Press. ness and default attention level (b=4.75, SE=1.80, t=2.63, p= .009, ______(2018), Introduction to Mediation, Moderation, and 95% CI= [1.19, 8.31]). For participants whose default attention level Conditional Process Analysis: A Regression-Based Approach, to food calorie information was low (b=-14.95, SE=4.42, t=-3.38, 2nd ed. New York, NY, US: Guilford Press. p= .001, 95% CI= [-23.66, -6.23]) or median (b=-6.82, SE=3.01, Jacoby, Larry L. and Mark Dallas (1981), “On the Relationship t=-2.26, p= .025, 95% CI= [-12.75, - .88]), presenting the calorie between Autobiographical Memory and Perceptual Learning,” number in round (vs. precise) number decreased their consumption Journal of Experimental Psychology: General, 110 (3), 306- intention. However, for participants whose default attention level to 40. calorie information was already high, presenting calorie informa- Jacoby, Larry L., Colleen M. Kelley and Jane Dywan (1989), tion as a round or precise number did not cause any difference in “Memory Attributions,” in Varieties of Memory and consumption intention (b=-1.32, SE=4.21, t=.31, p= .754, 95% CI= Consciousness: Essays in Honour of Endel Tulving, ed. H. L. [-6.97, 9.61]). Roediger III and F. I. M. Craik, Hillsdale, NJ, US: Lawrence In summary, Study 3 provides strong support to our theory and Erlbaum Associates, Inc., 391-422. suggest that the effect is primarily driven by consumers who are ha- Kettle, Keri and Gerald Häubl (2010), “Numeric Fluency and bitually inattentive to food calorie information. Preference,” in Na - Advances in Consumer Research, Vol. 37, ed. M. C. Campbell, J. Inman and D. Rik Pieters, MN: General Discussion Association for Consumer Research, 150-52. In recent years, considerable attention has been paid to the ef- Kim, Eojina, Liang Tang, Chase Meusel and Manjul Gupta (2018), ficient presentation of nutrition information on food packages. For “Optimization of Menu-Labeling Formats to Drive Healthy example, research has suggested presenting nutrition labels on front Dining: An Eye Tracking Study,” International Journal of (vs. back) packages (Van Kleef et al. 2008), and informing consum- Hospitality Management, 70, 37-48. ers of the duration of exercises needed to burn out the calories (Dow- King, Dan and Chris Janiszewski (2011), “The Sources and ray et al. 2013; Kim et al. 2018). The current research shows that Consequences of the Fluent Processing of Numbers,” Journal simply presenting calorie information in round (vs. precise) numbers of Marketing Research, 48 (2), 327-41. can nudge consumers to reduce their unhealthy food consumption, Kivetz, Ran and Itamar Simonson (2002), “Self-Control for especially for those who do not habitually pay attention to calorie the Righteous: Toward a Theory of Precommitment to information on food packages. Indulgence,” Journal of Consumer Research, 29 (2), 199-217. Our studies also show that the calorie roundedness effect is trig- Pena-Marin, Jorge and Rajesh Bhargave (2016), “Lasting gered by greater processing fluency associated with round (vs. pre- Performance: Round Numbers Activate Associations of cise) calorie numbers, which in turn magnifies perceived health risks Stability and Increase Perceived Length of Product Benefits,” of unhealthy foods, and subsequently reduces consumers’ consump- Journal of Consumer Psychology, 26 (3), 410-16. tion intention and actual consumption volume. Read, Daniel, George Loewenstein and Shobana Kalyanaraman (1999), “Mixing Virtue and Vice: Combining the Immediacy REFERENCES Effect and the Diversification Heuristic,”Journal of Albrecht, Sabine and Claus-Christian Carbon (2014), “The Fluency Behavioral Decision Making, 12 (4), 257-73. Amplification Model: Fluent Stimuli Show More Intense but Schwarz, Norbert (2004), “Metacognitive Experiences in Consumer Not Evidently More Positive Evaluations,” Acta Psychologica, Judgment and Decision Making,” Journal of Consumer 148, 195-203. Psychology, 14 (4), 332-48. Carbon, Claus-Christian and Sabine Albrecht (2016), “The Fluency Tangari, Andrea Heintz, My Bui, Kelly L. Haws and Peggy J. Liu Amplification Model Supports the Gane Principle of Arousal (2019), “That’s Not So Bad, I’ll Eat More! Backfire Effects Enhancement,” The Behavioral and Brain Sciences, 39, e204. of Calories-Per-Serving Information on Snack Consumption,” Dehaene, Stanislas and Jacques Mehler (1992), “Cross-Linguistic Journal of Marketing, 83 (1), 133-50. Regularities in the Frequency of Number Words,” Cognition, Thomas, Manoj, Daniel H. Simon and Vrinda Kadiyali (2010), 43 (1), 1-29. “The Price Precision Effect: Evidence from Laboratory and Dowray, Sunaina, Jonas J. Swartz, Danielle Braxton and Anthony Market Data,” Marketing Science, 29 (1), 175-90. J. Viera (2013), “Potential Effect of Physical Activity Based Van Kleef, Ellen, Hans Van Trijp, Frederic Paeps and Laura Menu Labels on the Calorie Content of Selected Fast Food Fernández-Celemín (2008), “Consumer Preferences for Front- Meals,” Appetite, 62, 173-81. of-Pack Calories Labelling,” Public Health Nutrition, 11 (2), Fassbender, Catherine, Sebastien Houde, Shayla Silver-Balbus, 203-13. Kacey Ballard, Bokyung Kim, Kyle J. Rutledge, J. Faye Wadhwa, Monica and Kuangjie Zhang (2014), “This Number Just Dixon, Ana-Maria Iosif, Julie B. Schweitzer and Samuel Feels Right: The Impact of Roundedness of Price Numbers on M. Mcclure (2014), “The Decimal Effect: Behavioral and Product Evaluations,” Journal of Consumer Research, 41 (5), Neural Bases for a Novel Influence on Intertemporal Choice 1172-185. in Healthy Individuals and in Adhd,” Journal of Cognitive ______(2019), “When Numbers Make You Feel: Impact of Round Neuroscience, 26 (11), 2455-468. Versus Precise Numbers on Preventive Health Behaviors,” Organizational Behavior and Human Decision Processes, 150, 101-11. Advances in Consumer Research (Volume 48) / 819

Wertenbroch, Klaus (1998), “Consumption Self-Control by Winkielman, Piotr, Norbert Schwarz, Tedra A. Fazendeiro and Rolf Rationing Purchase Quantities of Virtue and Vice,” Marketing Reber. (2003), “The Hedonic Marking of Processing Fluency: Science, 17 (4), 317-37. Implications for Evaluative Judgment,” in The Psychology of Winkielman, P. and J. T. Cacioppo (2001), “Mind at Ease Puts Evaluation: Affective Processes in Cognition and Emotion, a Smile on the Face: Psychophysiological Evidence That ed. J. Musch and K. C. Klauer, Mahwah, NJ, US: Lawrence Processing Facilitation Elicits Positive Affect,”Journal of Erlbaum Associates Publishers, 189–217. Personality and Social Psychology, 81 (6), 989-1000. Witherspoon, Dawn and Lorraine G. Allan (1985), “The Effect of a Prior Presentation on Temporal Judgments in a Perceptual Identification Task,” Memory & Cognition, 13 (2), 101-11. Digital Nudging to Reduce Screen Time Consumption Laura Zimmermann, IE University, Spain Michael Sobolev, Cornell University, USA

EXTENDED ABSTRACT dom of choice directly or nudge consumers to reduce screen time Mobile phones are our constant companions and an integral part less favorably than apps, which only provide information. We inves- of our daily lives. Despite the huge advantages of mobile devices, re- tigated consumer perceptions of three app design ideas (within-sub- search has started to uncover negative effects on behavior, including jects: tracking vs. blocking vs. greyscale) in terms of overall prefer- performance (Felisoni & Godoi, 2018; Lanaj, Johnson, & Barnes, ence, knowledge efficacy and reduction efficacy (scales from 0-100). 2014), emotional health (Lee, Chang, Lin, & Cheng, 2014), sleep Participants (N=139; Prolific Academic) evaluated three app design patterns (Thomée, Härenstam, & Hagberg, 2011) as well as cognitive ideas with the goal of helping consumers spend less unproductive capacity (Ward, Duke, Gneezy, & Bos, 2017). A significant propor- time on their phones. Within-subjects ANOVA showed that partici- tion of people want to reduce their screen time, yet they are not able pants had a clear preference for the purely informational tracking to achieve the goal. As a solution, both Apple and Google released app (F(2, 276)=15.74, p<.001), and thought it was most effective in time management applications with the goal of helping consumers increasing self-awareness (F(2, 276)=50.98, p<.001). However, in understand and change their mobile phone habits. terms of reduction efficacy, the blocking app was perceived as most Despite recent controversies regarding the correlational nature effective F( (2, 276)=18.55, p<.001). of studies investigating screen time consumption, there is a lack of Study 3 employs systematic experimentation of interventions systematic experimentation of interventions to reduce screen. Al- to reduce screen time as well as potential consequences for GPA and though a plethora of possible solutions are suggested in the popu- wellbeing. We conducted an RCT with three conditions among a lar press to reduce screen time, almost none have been examined student population (N=112) over three weeks and altered the choice in a randomized fashion. One approach to help consumers lower architecture of mobile consumption by manipulating existing mobile their screen time is altering the choice architecture by manipulat- features to help consumers lower their screen time. In the control ing existing features on mobile devices, for example, by incorporat- condition subjects simply tracked their screen time with the help of ing frictions in the usage process which disrupt habitual, automatic an app. Additionally, we examined two commonly proposed digital behavior and stimulate a more mindful interaction with the phone nudges varying along a spectrum of coerciveness: (1) greyscale: al- (Cox, Gould, Cecchinato, Iacovides, & Renfree, 2016). Such digital tering the color of the screen, and (2) self-commitment: setting goals nudges do not eliminate consumers’ freedom of choice to use their to limit time spent on addictive apps. Objectively measured screen phone (as opposed to blocking applications) but nevertheless lead time was obtained through screenshots. individuals in a particular direction (Okeke, Sobolev, Dell, & Estrin, As hypothesized, simply tracking phone usage had no effect on 2018; Thaler & Sunstein, 2009). screen time (M=283 min). However, the greyscale (M=233 min) and, This research addresses the question of whether screen time to a lesser extent, self-commitment (M=268 min) digital nudges led tracking apps effectively educate consumers about their digital be- to significant reduction in daily screen time (F(2,109)=3.21, p=.044). havior on mobile, and whether this leads to a reduction of screen time Subjects in the greyscale condition reported lower enjoyment of so- (study 1: longitudinal field study). Second, we investigate consum- cial media activities during the intervention (F(2,131)=3.35, p=.038), ers’ attitudes towards app designs varying in terms of their coercive- suggesting a process mechanisms through which greyscale is effec- ness (study 2: experimental survey). Third, we study digital nudges tive. that alter the choice architecture as a method to reduce screen time While our intervention was effective in reducing screen time, it consumption (study 3: longitudinal randomized trial). Finally, we did not have any immediate causal effects on sleep quality, happiness examine to what extent reducing objectively measured smartphone and stress level (all ps=ns). We also found no difference for GPA usage is linked to wellbeing and academic performance. (F(2,109)=0.15, p=.858). Nevertheless, average daily screen time Study 1 investigated the effect of mobile phone screen time was correlated with GPA (r=-0.263, p<.0.001). tracking apps on users’ self-awareness, their actual time spent on Our findings offer insights for consumer researchers and prac- mobile, and long-term consequences for grade point average (GPA). titioners interested in curbing mobile phone overload. By showing Participating students (N=246) completed two screen time estima- that screen time is linked to GPA we add to the growing literature on tion surveys before and after (T1 & T2) a mobile phone tracking consequences of mobile phone overuse. This research further con- project which was part of a university workshop. Students down- tributes to the ‘quantified self’ literature (Lupton, 2016) by showing loaded a tracking app and submitted screenshots of their usage indi- that the quantification of mobile phone usage with the help of track- cators. At T1, before the tracking project, estimated screen time was ing app has benefits for self-awareness, but no impact on behavior not correlated with actual screen time (r=.018, p=.783). At T2, after change. Finally, we add to the literature on digital nudges, by show- tracking, participants’ estimates of mobile screen time were highly ing that greyscale mode can be a viable technique to reduce mobile correlated with their actual screen time (r=0.691, p<.001). Longer use. Taken together, our research advances knowledge about effec- tracking durations led to better calibration (β=1.63, SE=.46, t=3.55, tive solutions to curb digital overload and potential consequences of p<.001). However, this did not lead to a reduction of overall screen screen time reduction. time. In fact, the findings suggest the opposite (β=1.56, CI [-.22, 3.35], t=1.71, p=.087), an increase in screen time. Higher mobile REFERENCES phone screen time was linked to lower GPA (β=-0.0004, SE=0.0002, Cox, A. L., Gould, S. J. J., Cecchinato, M. E., Iacovides, I., & p=.031). Renfree, I. (2016). Design frictions for mindful interactions: Study 2 investigated consumers’ attitudes towards app design The case for microboundaries. In Proceedings of the 2016 idea varying in terms of coerciveness (Thaler & Sunstein, 2009). We CHI Conference Extended Abstracts on Human Factors in hypothesized that consumers evaluate app designs, which limit free- Computing Systems.

Felisoni, D. D., & Godoi, A. S. (2018). Cell phone usage and Thaler, R. H., & Sunstein, C. R. (2009). Nudge: Improving academic performance: An experiment. Computers & decisions about health, wealth, and happiness: Penguin. Education, 117, 175-187. Thomée, S., Härenstam, A., & Hagberg, M. (2011). Mobile phone Lanaj, K., Johnson, R. E., & Barnes, C. M. (2014). Beginning the use and stress, sleep disturbances, and symptoms of depression workday yet already depleted? Consequences of late-night among young adults-a prospective cohort study. BMC public smartphone use and sleep. Organizational Behavior and health, 11(1), 66. Human Decision Processes, 124(1), 11-23. Ward, A. F., Duke, K., Gneezy, A., & Bos, M. W. (2017). Brain Lee, Y.-K., Chang, C.-T., Lin, Y., & Cheng, Z.-H. (2014). The dark drain: The mere presence of one’s own smartphone reduces side of smartphone usage: Psychological traits, compulsive available cognitive capacity. Journal of the Association for behavior and technostress. Computers in human behavior, 31, Consumer Research, 2(2), 140-154. 373-383. Lupton, D. (2016). The quantified self: John Wiley & Sons. Okeke, F., Sobolev, M., Dell, N., & Estrin, D. (2018, September) Good vibrations: can a digital nudge reduce digital overload?. In Proceedings of the 20th international conference on human- computer interaction with mobile devices and services. Engagement with Virtue and Vice Information Through the Day Ozum Zor, Rutgers University, USA Kihyun Hannah Kim, Rutgers University, USA Ashwani Monga, Rutgers University, USA

EXTENDED ABSTRACT vice-virtue asymmetry for tweets from CNN when we randomly as- A growing number of consumers access information using so- sign participants to morning and evening time periods. We observe cial media platforms, such as Twitter. These platforms have become the same asymmetry in study 5 where we measure actual engagement the dominant locations to engage with new information. Consumers of Twitter users with several magazines through likes (Study 5A) and spend more than two hours per day on social media (Bayindir and followers (Study 5B). Finally, in study 6, we measure actual engage- Kavanagh 2018). Users mostly engaged with frivolous information ment through likes by using vice versus virtue information from the in the early days of social media, but it is now a source of news for Twitter account of a newspaper. In addition, we observed that this 68% of U.S. adults (Matsa and Shearer 2018). Thus, it is critical to asymmetry becomes stronger when day temperatures are higher, pro- understand how different types of information are driving engage- viding evidence for depletion as the underlying reason. ment. We specifically focus on vice (i.e., tempting, but with few later Our results add to the findings from social media research. To benefits) versus virtue information (i.e., not tempting, but offering our knowledge, we are the first to demonstrate asymmetric engage- later benefits). ment with vice versus virtue content throughout the day on social We draw from research on self-control. Similar to muscles media. By examining vice versus virtue content online, we also con- weakening after physical exertion, self-control weakens after subse- tribute to research on self-control. We believe that our study is the quent efforts (Muraven and Baumeister 2000). Individuals are faced first to show depletion effects using large-scale social media datasets. with multiple self-control challenges through the day, as they try to The current research has implications for consumers and practi- avoid tempting activities and foods. Such behavior is depleting (Vohs tioners. Our findings can help consumers, and public policy makers et al. 2008), weakening self-control as the day progresses. For in- understand how the time of day changes sensitivity to different kinds stance, individuals are more likely to disregard professionalism rules of information. On the flip side, our results are helpful for companies or engage in unethical behavior later in the day (Dai et al. 2015; trying to decide when to release which type of information to best Kouchaki and Smith 2014). What is not yet understood is how self- capture consumers’ attention. control depletion may manifest for online information engagement. One exception is Hofmann et al.’s (2012) examination of a broad set REFERENCES of activities, including TV watching and web surfing. They observed Bayindir, Nisa and Duncan Kavanagh (2018), “The Latest Social an increased desire to engage in several media activities in the eve- Media Trends to Know in 2019,” https://www.globalwebindex. ning hours. Will such a pattern emerge similarly for engagement with com/reports/social. online information, specifically, tweets on Twitter? We argue that the Dai, Hengchen, Katherine L. Milkman, David A. Hofmann, and outcome will depend on what the tweet is about. That is, the type of Bradley R. Staats (2015), “The Impact of Time at Work and information will matter. Time Off From Work on Rule Compliance: The Case of Hand Specifically, we discover an asymmetry in the engagement with Hygiene in Health Care,” Journal of Applied Psychology, 100 vice versus virtue information. A relative vice offers immediate plea- (3), 846–62. sure at the moment of consumption but may have negative conse- Hofmann, Wilhelm, Kathleen D. Vohs, and Roy F. Baumeister quences later on; in contrast, consumption of a relative virtue may not (2012), “What People Desire, Feel Conflicted About, and Try be pleasurable at the moment but provides benefits in the long-term to Resist in Everyday Life,” Psychological Science, 23 (6), (Khan and Dhar 2007; Khan, Dhar, and Wertenbroch 2005; Siddiqui, 582–8. May, and Monga 2017). While a bulk of the vice-virtue research has Khan, Uzma and Ravi Dhar (2007), “Where There Is a Way, Is been conducted on foods, this distinction can apply equally to infor- There a Will? The Effect of Future Choices on Self-Control,” mation content. Thus, a salacious gossip in the Cosmopolitan could Journal of Experimental Psychology: General, 136 (2), be considered vice information by many who are tickled by the in- 277–88. stantaneous pleasure of reading the article but know that there is no Khan, Uzma, Ravi Dhar, and Klaus Wertenbroch (2005), “A long-term benefit from it. However, an article on financial markets Behavioral Decision Theory Perspective on Hedonic and in Forbes could be considered virtue information by many who find Utilitarian Choice,” in Inside Consumption: Frontiers of it hard to understand but appreciate the long-term benefit of becom- Research on Consumer Motives, Goals, and Desires, ed. ing more knowledgeable on the topic. We argue that the self-control S. Ratneshwar and David Glen Mick, London: Routledge, depletion arising at the end of the day is likely to reduce the resources 144–65. needed to process the Forbes article, but makes one more likely to Kouchaki, Maryam and Isaac H. Smith (2014), “The Morning succumb to the immediate gratification from reading theCosmopoli - Morality Effect: The Influence of Time of Day on Unethical tan article. Thus, we expect that as the day wears on, consumers will Behavior,” Psychological Science, 25 (1), 95–102. engage more with vice content and less with virtue content. Matsa, Katerina Eva and Elisa Shearer (2018), “News Use We observe such a pattern across six studies—four experiments Across Social Media Platforms 2018,” http://www. and two studies using large-scale Twitter datasets. In study 1, we find journalism.org/2018/09/10/news-use-across-social-media- initial evidence of engagement shifting away from virtue and toward platforms-2018/. vice as morning turns to evening. Studies 2 and 3 replicate this result Muraven, Mark and Roy F. Baumeister (2000), “Self Regulation with student participants in a lab setting, where we simply measure and Depletion of Limited Resources: Does Self Control the time at which participants complete the survey considering vice Resemble a Muscle?” Psychological Bulletin, 126 (2), 247–59. versus virtue magazines and websites. Study 4 also shows the same

Siddiqui, Rafay A., Frank May, and Ashwani Monga (2017), “Time Vohs, Kathleen D., Roy F. Baumeister, Brandon J. Schmeichel, Window as a Self-Control Denominator: Shorter Windows Jean M. Twenge, Noelle M. Nelson, and Dianne M. Tice Shift Preference toward Virtues and Longer Windows toward (2008), “Making Choices Impairs Subsequent Self-Control: Vices,” Journal of Consumer Research, 43 (6), 932–49. A Limited-Resource Account of Decision Making, Self- Regulation, and Active Initiative,” Journal of Personality and Social Psychology, 94 (5), 883–98. 824 / Engagement with Virtue and Vice Information Through the Day Special Sessions

Digitized Big Data and Machine Learning Approaches To Consumer Behavior Chair: Ada Aka, University of Pennsylvania, USA

Paper #1: Computational Consumer Segmentation and Brand Nave, Rentfrow, and Bhatia combine data from Facebook and Management: Leveraging Social Media, Digitized Language IMDb to investigate associations between movie characteristics and Data, and Machine-Learning to Identify Consumer Personality consumer personalities. They find that key features of movies’ plots Segments, Facets of Brand Image, and Predictors of Brand predict the personality profiles of consumers who are fans of those Liking movies, above-and-beyond demographic variables, movie genre, and Ada Aka, University of Pennsylvania, USA movie metadata. Their findings thus reveal robust links between per- Christopher Olivola, Carnegie Mellon University, USA sonality and movie preferences, with implications for media market- Sudeep Bhatia, University of Pennsylvania, USA ing recommendation systems. Gideon Nave, University of Pennsylvania, USA Goetz, Ebert, Mueller, Gladstone, and Matz examine how the personality of consumers’ social context (i.e. the personality of Paper #2: We Are What We Watch: Movies Contents Predicts those around them) shape their spending decisions above and be- The Personality Of Their Social Media Fans yond their own individual traits. Using spending records of 111,336 Gideon Nave, University of Pennsylvania, USA participants across 374 Local Authority Districts (LAD), she finds a Jason Rentfrow, Cambridge University, UK congruency between geographic regions’ aggregate scores on a given Sudeep Bhatia, University of Pennsylvania, USA personality trait and the collective spending on categories associated Paper #3: Spending Reflects Not Only Who We Are But Also with that trait. Furthermore, individual-level analyses demonstrate Who We Are Around: The Joint Effects of Individual and that regional personality of a participant’s home LAD predicts indi- Geographic Personality on Spending vidual spending above and beyond individual personality. Goetz Friedrich, Cambridge University, UK Wang, Lu, Li, Khamitov, and Bendle investigate the relation- Ebert Tobias, Mannheim University, Germany ship between vocal tones of persuaders and online persuasion using Mueller Sandrine, Columbia University, USA Kickstarter videos. Applying a novel audio mining technology, they Gladstone Joe, University College London, UK find that concentration, lack of stress, and stable emotions influence Sandra Matz, Columbia University, USA funding success. Next, they conduct a controlled experiment to show Paper #4: The Role of Vocal Tone in Online Persuasion: A how greater signs of stress or extreme emotion hurt while greater Crowdfunding Enquiry signs of concentration help persuasion success. Lastly, they identify Xin (Shane) Wang, Western University, Canada perceptions of competence as a mediator in this relationship between Shijie Lu, University of Houston USA vocal tones and persuasion success. Xi Li, City University of Hong Kong, Hong Kong, China Altogether, the papers presented in this session demonstrate Mansur Khamitov, Nanyang Technological University, how novel machine learning approaches applied to digitized big data Singapore can provide important new insights to consumer behavior. More spe- Neil Bendle, Western University, Canada cifically, each of the papers use a unique digitized data platform and apply varying novel quantitative techniques to answer consumer- SESSION OVERVIEW relevant research questions. This session will appeal to a broad audi- In today’s world, consumers leave unprecedented amounts of ence who are interested in learning about novel approaches to study- digitized footprints every day when they use social media platforms, ing consumer behavior, and will also be relevant to all researchers as review a movie on IMDb, make purchases using their credit cards, well as practitioners interested in consumer personality, cognition, or decide to help fund a new project on Kickstarter. The availabil- and decision-making. ity of digitized ‘Big Data’ and the recent advances in data science created numerous novel opportunities to ask many consumer-related Computational Consumer Segmentation and Brand questions. Our session concentrates on multiple different kinds of Management digitized ‘Big Data’ and applies cutting-edge machine learning techniques to gain insights about various aspects of consumer cognition EXTENDED ABSTRACT and behavior. The emergence of social media has fundamentally changed the Aka, Olivola, Bhatia, and Nave shed light on the relationship way consumers interact with brands and, as a result, how brands are between consumer personality and brand liking as revealed in large managed (e.g., Appel et al. 2020; Batra and Keller 2016). For ex- online datasets. Using machine learning models trained on large- ample, every day, millions of social media users (e.g., on Facebook, scale digitized language data and Facebook ‘likes’ data, they predict Instagram, and Twitter) express their attitudes towards branded con- consumer personality profiles for brands above and beyond demo- tent through ‘likes’ (De Vries et al. 2012; John et al. 2017). These graphic variables, or other objective and subjective brand character- indicators are often viewable to their friends, family, colleagues, and istics. They show how these ‘black box’ models can be interpreted acquaintances, and are available to brand managers in real-time. As in terms of the images associated with brands liked by particular a consequence, firms can efficiently monitor the favorability of their consumer personality segments. Finally, they also demonstrate self- brands and how this favorability evolves over time. In this paper, we congruity effects in consumer liking of brands. use several recently assembled datasets drawn from social media and news media. We analyze these data using techniques from machine learning, to study how consumer psychographic characteristics (their

Advances in Consumer Research 825 Volume 48, ©2020 826 / Digitized Big Data and Machine Learning Approaches To Consumer Behavior personality) relate to brand image and brand liking. In particular, we of them were consistent with the notion that consumers ‘like’ brands analyze a unique dataset of Facebook users who rated their (own) that reflect their personality. We test this conjecture more formally personality characteristics and examine the relationships between in the next study. different brands and the personality of social media users who “liked” In Study 3, we carried out a similar analysis, but with personali- those brands. This allows us to determine how personality traits are ty words instead of the LIWC constructs, in order to directly examine associated with brand liking. We go further by utilizing a word em- whether brand personality dimensions relate to consumer personality beddings model built using large scale language data (Mikolov et al. traits. Doing so allows us to evaluate the extent to which consumers 2013; see reviews in Bhatia et al. 2019) and an established semantic tend to like brands that “fit” (i.e., are congruent with) their own per- dictionary (the LIWC) (Pennebaker and Graybeal 2001; Pennebaker sonality traits. Our analyses using dimensions of human personality et al. 2003) to predict consumer personality and to identify the main found evidence for congruity between BCAs and BCPPs for three of concepts and themes that each brand is associated with, in the mind the Big Five dimensions: Openness, Agreeableness and Emotional of consumers. This, in turn, allows us to analyze the relationships Stability (thus, for example, brands strongly associated with Open- between consumer characteristics and brand liking, for hundreds of ness adjectives were also liked by individuals high in Openness). brands, both within and across different product categories. We pres- In sum, this paper showcases, through three studies, the power ent three studies. The goal of Study 1 was to evaluate our machine and promise of combining big data and machine learning methods to learning approach for predicting consumer personality. Specifically, examine three important topics in consumer research: consumer psy- we tested how well we could predict the average personality profiles chographic segmentation, brand image, and consumer-brand self- of Facebook users who ‘liked’ different brands. We call this 5-di- congruity. The capacity to accurately and inexpensively assess the mensional personality profile “Brand Consumer Personality Profile” BCPP of brands, using publicly available word embedding represen- (BCPP). We used high-dimensional embedding representations for tations, not only has important practical managerial applications for those brands (i.e., brand embeddings), derived from large-scale natu- marketing strategy, but also enhances our theoretical understanding. ral language data. The brand embeddings served as predictor variables in five Ridge regression models (one for each personality di- We Are What We Watch: Movies Contents Predicts mension) that attempted to predict (out-of-sample) the BCPPs of the The Personality Of Their Social Media Fans brands. Due to the flexibility inherent in machine learning modeling, we evaluated all our models in terms of their out-of-sample predic- EXTENDED ABSTRACT tive power—i.e., their ability to predict Facebook user personalities With the growing prevalence of on-demand video streaming for novel brands, not previously shown to the model. The results services, the volume of personalized media consumption is rapidly of this analysis plots observed vs. predicted BCPPs for each brand increasing. Despite the ubiquity of personalized video consumption, and for the five personality dimensions. Our approach achieves high and although understanding the psychological characteristics of me- accuracy rates, with an average out-of-sample correlation of 0.62 dia audiences is fundamental to many aspects of marketing, relative- across the five personality dimensions. To evaluate our approach ly little is known about the associations between movie preferences against other established methods, we tested competitor models that and individual differences in psychological traits. The current work used various quantified brand characteristics (e.g., brand age, brand addresses this gap in knowledge by systematically studying the per- personality trait ratings) as predictor variables (Lovett et al. 2014). sonality profiles of movie fans. Additional analyses of the robustness of our approach involved mod- Our analysis uses data collected via myPersonality, a Facebook els based on user demographics and brand categories. Our results app through which roughly 3.5 million users took personality ques- showed that these methods predict BCPPs above-and-beyond brand tionnaires and consented to share their Facebook data with research- category information, consumer demographics, and brand-level ers. Using the myPersonality database, we created personality pro- characteristics (including both In Study 2, we sought to examine and files for 854 movies, by averaging the Big Five personality scores of interpret the information contained in brand embeddings that gives all users who “liked” these movies on Facebook (we included all of rise to these successful predictions. Embedding vectors quantify the the movies in the myPersonality dataset that were liked by at least extent to which words and concepts are associated with each other in 250 people). Similarly, we estimated the fans’ demographic profiles language, and thus by extension, in the minds of consumers. In this by calculating their average age, percentage of females, and the per- way, embedding vectors provide an approach to measuring brand as- centage of individuals who reported being in a relationship. sociations and images automatically from natural language data. By We obtained rich representations of movies from IMDb (www. studying the words and concepts that make up these images, we can imdb.com), an online database that contains information about over try to characterize the variability in BCPPs across different brands, 5 million titles of movies and TV episodes . Each IMDb entry rep- and understand how this variability relates to brand images. That is, resents a title, and includes genre information (a total of 21 genres), we can “dig into” our brand embeddings data in order to determine and user-generated plot keywords that describe “any notable object, which attributes, traits, concepts, and constructs are most associated concept, style or action that takes place during a title”. IMDb also with brands that are ‘liked’ by consumers who are high (vs. low) on includes additional information about the movies, such as box office a particular personality dimension. income and critics ratings, to which we refer as “metadata”. We identified a number of distinct associations between psy- Our primary objective was to test whether movies’ character- chological constructs and BCPPs. For example, the BCPP for Open- istics (genre, metadata and content, as captured by plot keywords), ness correlated with brand construct associations (BCA) and brand contain information about the personality of their Facebook fans. To word associations for experiences, creativity and art (in other words, this end, we conducted “leave one out” cross validated predictions brands whose consumers were higher in Openness were closer in the (Stone, 1978; Zhang, 1993) to predict the Big Five profiles associ- word embedding space to words and constructs related to experi- ated with each movie, using different combinations of the following ences, creativity and art). Likewise, Agreeableness correlated with variables: (1) the movie keywords (2) the demographic profiles of pro-sociality and Extraversion with apparel and parties. While not the movies’ fans; (3) the movies’ IMDb metadata variables; (4) the all of the associations we identified have simple explanations, many movies’ genre. Personality predictions for each movie were carried Advances in Consumer Research (Volume 48) / 827 out by fitting regularized linear regressions with the least absolute personality variance that could potentially be explained by the other shrinkage and selection operator (LASSO (Tibshirani, 1996) to the factors (Sackett & Yang, 2000). We hope that the growing availabil- data of all other movies, and recording the out-of-sample predictions ity of video-streaming services (e.g., Netflix, Hulu, Amazon Prime) under different model specifications. The results reported here were will facilitate the generation of datasets that would allow exploring obtained using a LASSO with regularization parameter value of ƛ = the associations between personality and preferences for movies us- 0.001. Similar results were obtained using different ƛ values. Predic- ing actual watching behavior and with less selection bias. tive accuracy was calculated as the part of the variance in personality that could be explained by the independent variables. Spending Reflects Not Only Who We Are But Also Who Our results indicate that all of the Big Five traits could be pre- We Are Around: The Joint Effects of Individual and dicted from the keywords (all p-values <0.001), and that the vari- Geographic Personality on Spending ance in fans personalities that was predicted by plot keywords was significantly greater than what could be explained by metadata and EXTENDED ABSTRACT genre in all of the Big Five traits. Furthermore, adding plot keywords Research shows that people’s personality traits play and impor- significantly improved the predictive accuracy incrementally above tant role in their spending decisions. Previous research has shown all of the other variables, for all of the Big Five traits. that people tend to spend money in a way that is aligned with their Our findings provide marketers means for reaching audiences personality (Aaker, 1999, Govers & Schoormans, 2005; Matz, Glad- of specific personality profiles via cinema advertising (Phillips & stone, & Stillwell, 2016) and personal values (Hill & Howell, 2014). Noble, 2007), branded entertainment (Hudson & Hudson, 2006) or An extraverted person, for example, is more likely to spend money product placement (Gupta & Lord, 1998). While personality has on social activities such as going to bars and restaurants, while an been shown to be indicative of consumer’s attitudes and behavior introverted person is more likely to spend money on books (Glad- (Arbuthnot, 1977; Matz, Kosinski, Nave, & Stillwell, 2017; Moon, stone et al., 2019; Matz et al., 2016). This is because spending not 2002; Wheeler, Petty, & Bizer, 2005), its use in marketing has been only reflects functional need-fulfillment but also reflects who we are limited due to the questionnaire-based nature of personality assess- as individuals (self-congruity theory, Sirgy, 1985): We buy products ment. Although recent research has shown that personality can be not only for what they can do, but also for what they mean to us approximated from individual’s digital footprints (e.g., Facebook (Levy, 1959). Consistent with this idea, individuals high in extraver- Likes, Instagram photos and Twitter language), marketers seldom sion, agreeableness and openness prefer to spend money on experi- have access to dataset that contain both individual-level digital foot- ences (Howell, Pchelin, & Iyer, 2012; Mehmetoglu, 2012), whereas prints and personality measures. Our approach overcomes this barri- individuals low in agreeableness, conscientiousness and emotional er by allowing marketers to assess the personality profiles of movies’ stability prefer to spend money on possessions (Zhang, Howell, Cap- audiences based on the movies’ attributes alone (as measured using rariello, & Guevarra, 2014). plot keywords), and without having to collect data about individuals. At the same time spending is also an inherently interpersonal, Our work also contributes to the literature of content recommen- social behavior and--at least in part--performative. What people buy dation and prediction of viewing behaviors. Typical individual-level is not only guided by their own preferences, but also by cultural recommendation systems (or prediction models) requires observing norms (de Mooij & Hofstede, 2002; Henry, 1967; Kacen & Lee, each consumer’s past behavior (i.e., using panel data). However, 2002), social expectations (Martineau, 1958; Mathews & Slocum, such data may not be available in the particularly important applica- 1969; Rook & Fisher, 1995; van Kempen, 2004), status concerns tion of making recommendations to new consumers. However, in (Brown, Bulte & Zhang, 2011; Martineau, 1958; Veblen, 1899), some cases it is possible to approximate the personalities of new con- and people’s social networks (Garling, Kirchler, Lewis, & van Raaj, sumers (e.g., from their tweeter language), and provide them with 2009; Sheth, 1967; Wang & Xiao, 2009). Spending can be used to personality-based recommendations. satisfy social needs, such as reinforcing and expressing self-iden- Our study has several limitations, which provide a fertile tity to others, and allowing one to differentiate oneself and to as- ground for future investigations. First, while our findings suggest sert one’s individuality (Ball & Tasaki, 1992; Belk, 1988; Fournier, that part of the variance in movie preferences is linked to stable 1998; Richins & Rudmin, 1994). Consequently, what people buy is personality traits, movie preferences are likely also influenced by not only determined by their own preferences, but also their socio- environmental context. For example, the movie industry is highly cultural context. seasonal (Einav, 2007), and people were shown to prefer romantic In this work, we integrate these separate research streams by movies when it is colder (Hong & Sun, 2012). Continuous inves- investigating the extent to which the personality of consumers’ social tigation of how context variables affects dynamic aspects of movie context (i.e. the personality of those around them) influence their preferences may benefit from studying interactions between context spending decisions above-and-beyond their own individual traits. variables and personality. Specifically, we test the impact of individual and regional personal- A second limitation concerns our use of Facebook Likes is ity as drivers of spending behavior in two steps. First, we analyze proxies for movie preferences. While Likes are active, ecological- spending records from 111,336 participants (over 31 million unique ly valid expressions of preferences, they do not necessarily reflect transactions) to test whether spending behaviors (i.e., what people what people de facto watch, and they might be affected by social and spend their money on) are regionally clustered and whether these cultural factors that do not immediately reflect the content of mov- regional clusters of spending are reliably predicted by regionally ag- ies, such as signalling and impression management. Our reliance on gregated personality traits (H1). This step is necessary as examining Facebook likes might also bias our study because specific types of the relationship between regional personality and spending behavior people may be more disposed to sharing their Facebook data with the is only meaningful if people in different areas actually spend money MyPersonality app, or to liking movie-related contents on Facebook. differently. Second, we analyze individual personality and spending However, such selection bias would be expected, in general, to atten- of 1,716 participants, to test whether regionally aggregated person- uate the associations we observed, because it would restrict the range ality predicts individual spending behavior above-and-beyond indi- of personalities that we observe and thus constrain the amount of vidual personality (H2). 828 / Digitized Big Data and Machine Learning Approaches To Consumer Behavior

Our findings show that personality-related spending is mean- tions to studying this important consumer-relevant phenomenon but ingfully clustered on the geographic level. The effects are strongest also somewhat changes the process. Online those making persuasion for the traits of extraversion and openness, for which we find strong attempts (who we will call persuaders) cannot be assessed through geographical clusters of highly prototypical spending in and around such things as handshakes and eye contact. Yet as technology pro- Metropolitan London, while low (i.e., introverted and less open) gresses this gives access to more cues that allow for intuitive assess- spending clusters North of London, in the Southwest, Wales and the ment (Nisbett and Wilson 1977) by those subject to a persuasion at- North of England. For agreeableness, we find significant hotspots tempt (who we will call receivers). Receivers tend to use assessments of prototypical agreeable spending in the very Southwest of Eng- beyond the message content (Hall 1980) to judge whether persuaders land, Southern Wales as well as in pockets of Northern England and are worthy of their support. Since online videos allow receivers of Scotland, while low, disagreeable spending clusters in Metropolitan a persuasion attempt to hear the vocal tones that the persuader uses, London and North of London. For conscientiousness, high proto- these vocal tones are likely to give cues that are used by the receivers typical spending clusters in the very South and North of England and to inform the persuasion attempt success. the English Midlands, while low, unconscientious spending strongly Prior work has examined how certain vocal character- cluster in Southern Wales and parts of Scotland. For neuroticism, istics, such as speech rate, pitch, volume, and tone, can influence high prototypical spending clusters in pockets of Scotland, South- individuals’ perceptions of speaker traits and character (Apple et ern Wales and the English Midlands, while low, emotionally stable al. 1979; Chattopadhyay et al. 2003; Cheng et al. 2016; Klofstad et spending primarily clusters around London. Importantly, those al. 2012, 2015; Oleszkiewicz et al. 2017; Street et al. 1983; Signo- spending clusters map onto clusters found for the regional personal- rello 2019; Tigue et al. 2012; Wiener and Chartrand 2014). A vast ity of participant. We capture this relationship more objectively by majority of this important work has been conducted in traditional, calculating the zero- order correlations between the two variables face-to-face contexts. This precludes understanding all the nuances on the aggregated LAD level (N = 374). We find a significant posi- and intricacies of the now prevalent online and digital persuasion tive relationship for all Big Five personality traits. These correlations attempts. Second, virtually all relevant prior research did not venture were medium-sized for openness (r = .34, p < .001), extraversion (r beyond assessing attitudes and behavioral intentions to systemati- = .47, p < .001), agreeableness (r = .33, p < .001) and neuroticism cally tap into real-world funding outcomes that are more relevant to (r = .34, p < .001) and small for conscientiousness (r = .17, p < .01). ultimate persuasion success. Third, prior work has predominantly To test whether regionally aggregated personality scores predict employed labor-intensive and/or semi-automated approaches to cod- prototypical individual level spending above-and-beyond individual ing and interpreting vocal characteristics which limited scalability personality, we used a subset of 1,716 participants for which individ- and researchers’ ability to consider fine-grained detail. These ben- ual spending and personality information was available. Specifically, efits are now afforded by advanced machine learning techniques and we employed multilevel analyses (individuals nested in 374 LADs) computational methods. to regress prototypical individual level spending on individual per- We first used novel audio mining technology to investigate sonality, gender, age and total spending amount (level 1 predictors) Kickstarter videos, looking for indications related to the receiver’s and aggregated personality (level 2 predictor). We z-standardized all assessment of the persuader’s concentration, stress, and lack of ex- non-binary predictors (i.e., all except gender) and specified random treme emotion. These indications were assessed for connections to intercepts and random slopes for individual personality. successful persuasion, i.e. funding of the project. We collaborated As hypothesized, regional personality positively predicted pro- with Nemesysco Ltd., an Israel based high-tech firm whose QA5 totypical spending above all individual level predictors for all Big system (Nemesysco 2015) was commercially applied in call cen- Five traits. For conscientiousness (β = .06, p < .01) and neuroticism ters and sold as “the most sophisticated, flexible, cutting edge voice (β = .06, p < .05) effect sizes reached roughly one third of the ef- analysis technology available today” (Nemesysco 2016). The soft- fect size of individual personality. For openness (β = .08, p < .001) ware not only allows analysis of many more audio files than could and agreeableness (β = .10, p < .001), the regional level effect ap- be achieved using human coders, it also uses standardized signal proached half the size of the individual level effect. Finally, for ex- processing algorithms to extract and combine attributes from voices traversion, the regional level effect (β = .17, p < .001) even reached identifying, amongst other things, stress, cognitive processes, and the same size as the individual level effect. In other words, prototypi- emotional reactions. Taken together, the study showed a significant cally extraverted spending to the same extent reflects the personality relationship between concentration, lack of stress, and lack of exces- of the regional context as the personality of the individuals them- sive emotionality on funding success. selves. These results provide support for our second hypothesis that Next, we conducted a controlled experiment to validate regionally aggregated personality scores predict individual spend- our findings from the secondary data and to extend them by investi- ing on associated purchase categories above-and-beyond individual gating if perceptions of competence mediate the differential relation- level personality for all of the Big Five personality traits. ship between vocal tones and persuasion attempt success. We used Taken together, our findings empirically support the proposition a pretested “Coolest cooler brand” actual Kickstarter pitch recorded that spending behavior is a reflection of personality traits as both a by a research assistant that was made to sound either concentrat- personal and environmental characteristic. ed, stressed, extremely emotional, or neutral (control). Overall, we found that a funding request was less likely to succeed when the The Role of Vocal Tone in Online Persuasion: voice of a persuader requesting funds in a pitch shows greater signs A Crowdfunding Enquiry of stress or extreme emotion compared to a control. In contrast, a funding request was more likely to succeed when the voice of the re- EXTENDED ABSTRACT questor shows greater signs of concentration (vs. control). The find- A major stream within consumer research involves under- ings of study 2 also support our theorizing regarding a focal mecha- standing persuasion attempts and reactions to these (Campbell and nism. Specifically, we documented that a concentrated vocal tone Kirmani 2000; Friestad and Wright 1994; Petty and Cacioppo 1986). (vs. control) generates greater perceptions of persuader competence, The profusion of online interactions provides additional observa- thus, leading to increased willingness to fund and improved brand Advances in Consumer Research (Volume 48) / 829 evaluations. Conversely, a stressed or extremely emotional vocal Chattopadhyay, Amitava, Darren Dahl, Robin J. B. Ritchie, and tone (vs. control) reduced perceived persuader competence, thereby Kimary N. Shahin (2003), “Hearing Voices: The Impact of lowering willingness to fund and brand evaluations. Lastly, study 2 Announcer Speech Characteristics on Consumer Response to also explored the possibility that given trust and processing fluency Broadcast Advertising,” Journal of Consumer Psychology, 13 are central considerations in consumer decision-making in the mar- (3), 198–204. ketplace (Chaudhuri and Holbrook 2001; Graf et al. 2018; Khamitov Chaudhuri, Arjun and Morris B. Holbrook (2001), “The Chain of et al. 2019; Morgan and Hunt 1994; Reber et al. 1998; Winkielman Effects from Brand Trust and and Cacioppo 2001), the resulting levels of trust or distrust in a per- Cheng, Joey T., Jessica L. Tracy, Simon Ho, and Joseph Henrich suader might have guided a particular peer-to-peer funding decision. (2016), “Listen, Follow Me: Since vocal tone dimensions did not seem to foster different levels of Crowley, Ayn E. and Wayne D. Hoyer (1994), “An Integrative processing fluency or persuader trust, the latter are unlikely to drive Framework for Understanding Two-Sided Persuasion,” our findings. Journal of Consumer Research, 20 (4), 561–74. Our findings offer insight to the persuasion literature, rein- De Vries, Lisette, Sonja Gensler, and Peter S.H. Leeflang forcing that speaker characteristics matter (Crowley and Hoyer 1994; (2012), “Popularity of Brand Posts on Brand Fan Pages: An Eisend 2006; Rucker et al. 2014; Tormala and Petty 2004; Swartz Investigation of the Effects of Social Media Marketing,” 1984; Wilson and Sherrell 1993). Our enquiry documents that speak- Journal of Interactive Marketing, 26 (2), 83–91. er characteristics in online persuasion attempts can not only be read- Dynamic Vocal Signals of Dominance Predict Emergent Social ily inferred from the voice heard but, more importantly, matter so Rank in Humans,” Journal of Experimental Psychology: much in first impressions as to exert influence on real world financial General, 145 (5), 536–47. outcomes. In doing this our work speaks to persuasion in the online Einav, L. (2007). Seasonality in the U.S. motion picture industry. world which is an increasingly fruitful topic given the profusion of Rand Journal of Economics, 38(1), 127–145. online interactions and how they provide additional opportunities to Eisend, Martin (2006), “Two-Sided Advertising: A Meta-Analysis,” studying persuasion (Grewal and Stephen 2019; Kupor and Tormala International Journal of Research in Marketing, 23 (2), 2018). In contrast to the work focused on offline persuasion environ- 187–98. ments, another key contribution of our research is to provide an early Friestad, Marian and Peter Wright (1994), “The Persuasion illustration that the effectiveness of online persuasion efforts can Knowledge Model: How People Cope with Persuasion be computationally mined and captured at a much larger scale with Attempts,” Journal of Consumer Research, 21 (1), 1–31. more fine-grained detail using an automatic novel method. Here, we Graf, Laura K. M., Stefan Mayer, and Jan R. Landwehr (2018), provide one of the pioneering accounts of how computers can pre- “Measuring Processing Fluency: One versus Five Items,” cisely predict the general effectiveness of online persuasion attempts Journal of Consumer Psychology, 28 (3), 393–411. based on automatically extracted vocal tone characteristics and their Grewal, Lauren and Andrew T. Stephen (2019), “In Mobile We interplay with the extent of video stimulation and brightness. In do- Trust: The Effects of Mobile Versus Nonmobile Reviews ing this we contribute by highlighting audio-mining’s potential in on Consumer Purchase Intentions,” Journal of Marketing academic consumer research (Hobson et al. 2012; Mayew and Ven- Research, 56 (5), 791–808. katachalam 2012). We, using technology now available to marketers, Gupta, P. B., & Lord, K. R. (1998). Product placement in movies: have fascinating new methods of quantifying key elements of the The effect of prominence and mode on audience recall. human voice. Using these could benefit numerous academic fields Journal of Current Issues & Research in Advertising, 20(1), with consumer research taking the lead. 47–59. Hall, Judith A. (1980), “Voice Tone and Persuasion,” Journal of REFERENCES Personality and Social Psychology, 38 (6), 924–34. Appel, Gil, Lauren Grewal, Rhonda Hadi, and Andrew T. Stephen Hobson, Jessen L., William J. Mayew, and Mohan Venkatachalam (2020), “The Future of Social Media in Marketing,” Journal of (2012), “Analyzing Speech to Detect Financial Misreporting,” the Academy of Marketing Science, 48 (1), 79–95. Journal of Accounting Research, 50 (2), 349–92. Apple, William, Lynn A. Streeter, and Robert M. Krauss (1979), Hong, J., & Sun, Y. (2012). Warm it up with love: The effect of “Effects of Pitch and Speech Rate on Personal Attributions,” physical coldness on liking of romance movies. Journal of Journal of Personality and Social Psychology, 37 (5), 715–27. Consumer Research, 39(2), 293–306. Arbuthnot, J. (1977). The roles of attitudinal and personality Hudson, S., & Hudson, D. (2006). Branded entertainment: A new variables in the prediction of environmental behavior and advertising technique or product placement in disguise? knowledge. Environment and Behavior, 9(2), 217–232. Journal of Marketing Management, 22(5-6), 489–504. Batra, Rajeev, and Kevin L. Keller (2016), “Integrating Marketing John, Leslie K., Oliver Emrich, Sunil Gupta, and Michael I. Norton Communications: New Findings, New Lessons, and New (2017), “Does ‘Liking’ Lead to Loving? The Impact of Joining Ideas,” Journal of Marketing, 80 (6), 122–45. a Brand’s Social Network on Marketing Outcomes,” Journal Bhatia, Sudeep, Russell Richie, and Wanling Zou (2019), of Marketing Research, 54 (1), 144–55. “Distributed Semantic Representations for Modeling Human Khamitov, Mansur, Xin (Shane) Wang, and Matthew Thomson Judgment,” Current Opinion in Behavioral Sciences, 2931–36. (2019), “How Well Do Consumer-Brand Relationships Drive Brand Affect to Brand Performance: The Role of Brand Loyalty,” Customer Brand Loyalty? Generalizations from a Meta- Journal of Marketing, 65 (2), 81–93. Analysis of Brand Relationship Elasticities,” Journal of Campbell, Margaret C. and Amna Kirmani (2000), “Consumers’ Consumer Research, 46 (3), 435–59. Use of Persuasion Knowledge: The Effects of Accessibility Klofstad, Casey A., Rindy C. Anderson, and Stephen Nowicki and Cognitive Capacity on Perceptions of an Influence Agent,” (2015), “Perceptions of Competence, Strength, and Age Journal of Consumer Research, 27 (1), 69–83. Influence Voters to Select Leaders with Lower-Pitched Voices,” PloS ONE, 10 (8), e0133779. 830 / Digitized Big Data and Machine Learning Approaches To Consumer Behavior

Klofstad, Casey A., Rindy C. Anderson, and Susan Peters (2012), Phillips, J., & Noble, S. M. (2007). Simply captivating: “Sounds Like a Winner: Voice Pitch Influences Perception of Understanding consumers’ attitudes toward the cinema as an Leadership Capacity in Both Men and Women,” Proceedings advertising medium. Journal of Advertising, 36(1), 81–94. of the Royal Society B: Biological Sciences, 279, 2698–704. Reber, Rolf, Piotr Winkielman, and Norbert Schwarz (1998), Kupor, Daniella and Zakary Tormala (2018), “When Moderation “Effects of Perceptual Fluency on Affective Judgments,” Fosters Persuasion: The Persuasive Power of Deviatory Psychological Science, 9 (1), 45–48. Reviews,” Journal of Consumer Research, 45 (3), 490–510. Rucker, Derek D., Zakary L. Tormala, Richard E. Petty, and Pablo Matz, S. C., Kosinski, M., Nave, G., & Stillwell, D. J. (2017). Briñol (2014), “Consumer Conviction and Commitment: An Psychological targeting as an effective approach to digital Appraisal-Based Framework for Attitude Certainty,” Journal mass persuasion. Proceedings of the National Academy of of Consumer Psychology, 24 (1), 119–36. Sciences of the United States of America, 114(48), 12714– Signorello, Rosario (2019), “Voice in Charismatic Leadership,” in 12719. The Oxford Handbook of Voice Studies, ed. Nina Eidsheim and Mayew, William J. and Mohan Venkatachalam (2012), “The Katherine Meizel, New York, NY: Oxford University Press Power of Voice: Managerial Affective States and Future Firm (165–89). Performance,” Journal of Finance, LXVII (1), 1–43. Stone, M. (1978). Cross-validation: A review. Series Statistics, 9(1), Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S., & Dean, J. 127–139. (2013). Distributed representations of words and phrases and Street Jr, Richard L., Robert M. Brady, and William B. Putman their compositionality. Paper presented at the Advances in (1983), “The Influence of Speech Rate Stereotypes and Rate neural information processing systems. Similarity or Listeners’ Evaluations of Speakers,” Journal of Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S., & Dean, J. Language and Social Psychology, 2 (1), 37–56. (2013). Distributed representations of words and phrases and Swartz, Teresa A. (1984), “Relationship between Source Expertise their compositionality. Paper presented at the Advances in and Source Similarity in an Advertising Context,” Journal of neural information processing systems. Advertising, 13 (2), 49–54. Moon, Y. (2002). Personalization and personality: Some effects of Tibshirani, R. (1996). Regression shrinkage and selection via customizing message style based on consumer personality. the lasso. Journal of the Royal Statistical Society. Series B, Journal of Consumer Psychology, 12(4), 313–325. Statistical Methodology, 267–288. Morgan, Robert M and Shelby D Hunt (1994), “The Commitment- Tigue, Cara C., Diana J. Borak, Jillian J. M. O’Connor, Charles Trust Theory of Relationship Marketing,” Journal of Schandl, and David R. Feinberg (2012), “Voice Pitch Marketing, 58 (3), 20–38. Influences Voting Behavior,” Evolution and Human Behavior, Nisbett, Richard E. and Timothy D. Wilson (1977), “The Halo 33 (3), 210–16. Effect: Evidence for Unconscious Alteration of Judgments,” Tormala, Zakary L. and Richard E. Petty (2004), “Source Journal of Personality and Social Psychology, 35 (4), 250–56. Credibility and Attitude Certainty: A Metacognitive Oleszkiewicz, Anna, Katarzyna Pisanski, Kinga Lachowicz- Analysis of Resistance to Persuasion,” Journal of Consumer Tabaczek, and Agnieszka Sorokowska (2017), “Voice-Based Psychology, 14 (4), 427–42. Assessments of Trustworthiness, Competence, and Warmth in Wheeler, S. C., Petty, R. E., & Bizer, G. Y. (2005). Self-schema Blind and Sighted Adults,” Psychonomic Bulletin and Review, matching and attitude change: Situational and dispositional 24 (3), 856–62. determinants of message elaboration. Journal of Consumer Pennebaker, James W. and Anna Graybeal (2001), “Patterns Of Research, 31(4), 787–797. Natural Language Use: Disclosure, Personality, And Social Wiener, Hillary J. D. and Tanya L. Chartrand (2014), “The Effect of Integration,” Current Directions in Psychological Science, Voice Quality on Ad Efficacy,” Psychology and Marketing, 31 10(3), 90-3. (7), 509–17. Pennebaker, James W., Matthias R. Mehl and Kate G. Niederhoffer Wilson, Elizabeth J. and Daniel L. Sherrell (1993), “Source Effects (2003), “Psychological Aspects Of Natural Language Use: in Communication and Persuasion Research: A Meta-Analysis Our Words, Our Selves,” Annual Review of Psychology, 54(1), of Effect Size,”Journal of the Academy of Marketing Science, 547-77. 21 (2), 101–12. Petty, Richard E. and John T. Cacioppo (1986), “The Elaboration Winkielman, Piotr and John T. Cacioppo (2001), “Mind at Ease Likelihood Model Of Persuasion,” Advances in Experimental Puts a Smile on the Face: Psychophysiological Evidence That Social Psychology, 19, 123–94. Processing Facilitation Elicits Positive Affect,”Journal of Phillips, J., & Noble, S. M. (2007). Simply captivating: Personality and Social Psychology, 81 (6), 989–1000. Understanding consumers’ attitudes toward the cinema as an Zhang, P. (1993). Model selection via multifold cross validation. advertising medium. Journal of Advertising, 36(1), 81–94. Annals of Statistics, 21(1), 299–313. Consumer Behavior Meets Machine Learning Chairs: Siham El Kihal, Frankfurt School of Finance & Management, Germany A. Selin Atalay, Frankfurt School of Finance & Management, Germany

Paper #1: Using Natural Language Processing to Investigate the ent. Today, the use of email, social media posts, live chats and blogs Role of Syntactic Structure in Persuasive Communication provides unprecedented opportunities to reach consumers with ver- A. Selin Atalay, Frankfurt School of Finance & Management, bal messages in addition to existing channels such as TV, radio or Germany newspaper ads. Siham El Kihal, Frankfurt School of Finance & Management, Extant literature in marketing suggests that the language used Germany influences the persuasive outcomes in various marketing contexts. Florian Ellsaesser, Frankfurt School of Finance & For instance, language used in news articles reflects the emotions Management, Germany and legitimacy of the article, and influences consumers’ attitudes toward the article as well as their likelihood of reading or sharing the Paper #2: Scalable Content Curation: Learning from Human article (Berger and Milkman 2012; Humphreys and LaTour 2013; Effort Berger, Moe, and Schweidel 2019). Language used in movie scripts Pavel Kireyev, INSEAD, France reflects the theme, scenes, and emotion in the movie and thus pre- Artem Timoshenko, Northwestern Kellogg School of dicts the box office success (Eliashberg, Hui, and Zhang 2007). Lan- Management, USA guage used in songs impacts the songs’ success on the market as the Cathy Yang, HEC Paris, France typicality of the lyrics impacts the popularity of a song (Berger and Paper #3: The Power of Brand Selfies in Consumer-Generated Packard 2018). Elaborateness and the emotionality of tweets about Brand Images political debates influences which topics get consumers’ attention Jochen Hartmann, University of Hamburg, Germany and are shared with others (Berman et al. 2019). Even though a large Mark Heitmann, University of Hamburg, Germany stream of recent literature is dedicated to understanding the role of Christina Schamp, University of Mannheim, Germany language used in marketing contexts (see Berger et al. 2020 for an Oded Netzer, Columbia Business School, USA overview), most of this research is dedicated to the content of the Paper #4: Black-Box Emotion Detection: On the Variability message communicated (i.e., choice of words). Tools for understand- and Predictive Accuracy of Automated Emotion Detection ing what makes the language used in marketing messages persuasive Algorithms independently of the content are still missing. Francesc Busquet, University of St. Gallen, Switzerland In the current research, we seek to understand what makes mar- Christian Hildebrand, University of St. Gallen, Switzerland keting messages persuasive independently of the message content. In that, our focus is on the impact of the syntax of the language used SESSION OVERVIEW (i.e., grammar) on the persuasiveness of the message communicated. In a recent paper, Proserpio and colleagues (2019) highlight Following a natural language processing (NLP) approach, we posit how advances in machine learning research help marketing research- that the syntactic structure of a message influences its persuasiveness ers estimate models that predict extremely well and that are able to (Gibson 1998). We develop and test a novel approach to measuring leverage the power of new data sources such as images, text, audio, syntactic structure in a series of three studies, and identify the spe- and video. At the same time, they raise how machine learning re- cific syntactic elements, referred to as dependencies, which impact search needs a “soul” which should capture competitors’ reactions in the persuasiveness of a marketing message. In our main study, us- complex markets as well as consumer theories to be able to capture ing a dataset consisting of 134 debates with 129,480 sentences, we the complexity of consumer behaviors, their needs and preferences. develop an NLP-based measure of syntactic structure. In this study, The papers in this special session “Consumer Behavior Meets we show that the syntactic structure of the language used in the de- Machine Learning” aim at improving our understanding of consum- bates impacts the attitude change caused by the debates. Using our er behavior and attitudes by combining advanced machine learning measure, we identify the specific syntactic elements that affect the techniques with behavioral theories. The papers contribute to an- persuasiveness of a marketing message. Our measure captures syn- swering the following high-level questions: tactic structure more systematically and better in terms of predicting • How can machine learning enhance our understanding of the persuasiveness of a message than past measures of syntax. consumer attitudes and behaviors? We validate our measure of syntactic structure and the impact • How can behavioral theories help machine learning re- of the specific syntactic elements that affect the persuasiveness of a searchers open the prediction black-box and generate more marketing message in two experiments that follow the main study: interpretable features? First, we conduct a laboratory experiment where we manipulate the syntactic structure of a marketing message using the specific syntac- Using Natural Language Processing to Investigate the tic elements identified in our main study. The context is a persuasive Role of Syntactic Structure in Persuasive Communication message about the benefits of wearing a bike helmet. We show that when the syntactic structure of the message is changed according to EXTENDED ABSTRACT our findings, the persuasiveness of the message is increased. Second, How to use language, that is, how to formulate marketing com- we conduct a field experiment on Facebook, where we improve the munications so that they are persuasive, is a decision problem most syntactic structure in the copy of an advertisement designed by a marketers face. Language, spoken and written alike, is a fundamental marketing agency using the specific syntactic elements identified in element of marketing messages. Language is used in various con- our main study, and measure click through rates for the ad. We show texts such as political debates, product reviews, product descriptions that when the syntactic structure of the ad language is improved us- and brand communications with the goal of persuading the recipi- ing the specific syntactic elements we identified, the persuasiveness

Advances in Consumer Research 831 Volume 48, ©2020 832 / Consumer Behavior Meets Machine Learning of the ad is increased: The click-through rate for the ad increased workload of its employees by 8-25% while maintaining the same significantly by 40% compared to the baseline ad version, which quality of the idea ranking for the sponsor. was designed by the marketing agency. This experiment also demon- We find that voters exhibit persistent differences in their “im- strates how marketers can use our method to improve their market- portance”, or the extent to which their votes help identify finalists. ing messages to make them more persuasive. As a result, a model trained on one set of contests can extend to fu- Our work contributes to marketing research by developing a ture contests even if ideation topics differ across contests. We define novel approach of measuring and modifying syntactic structure, al- a metric for measuring the importance of a voter as the deteriora- lowing for a more precise way of formulating persuasive marketing tion in the quality of the idea ranking when we remove this voter. communications. Additionally, our work contributes to marketing Interestingly, voter importance may differ depending on whether a practice by providing a tool that marketers can use to efficiently im- model is used or not. We call this phenomenon “predictable inac- prove the persuasiveness of their messages as well as any other com- curacy” – a voter may be more predictive of the sponsor’s choice if munication that uses language as a medium. her votes are properly transformed. Moreover, two voters who are similarly important when the platform uses average score to rank Scalable Content Curation: Learning from Human Effort ideas may differ significantly in their importance when a model is used, suggesting that different voters can exhibit different levels of EXTENDED ABSTRACT predictable inaccuracy. Voter “redundancy”, or the extent to which Many firms rely on the crowdsourcing platforms to generate voters are correlated with each other, can also lead to varying levels new product and advertising ideas. For example, Colgate-Palmolive, of predictably inaccuracy. Our results suggest that a firm interested a large consumer products firm, collaborated with a digital creative in identifying its most valuable voters may arrive at different con- agency to generate ideas for one of its most visible advertising cam- clusions depending on whether it assesses voter performance jointly paigns. The firm organized an ideation contest - members ofthe with a statistical model or independently of a model based on the agency’s innovation platform submitted ad ideas, and the winning initial ranking criteria. ideas were used to develop an ad that was shown to over 109 mil- We complement our findings with a survey based on a product lion viewers of the Super Bowl - one of the world’s most popular design contest organized by an open innovation platform called Lego television events. Major advertisers such as P&G and Unilever gen- Ideas. The survey allows us to design simulations to study the generate ideas for new products and advertising campaigns either us- eralizability of our findings, recover ground-truth measures of idea ing internal crowdsourcing platforms or through specialized creative quality that are not available in the field data, and study how voter agencies. Crowdsourced ideas can be more creative than internal characteristics may relate to voter importance. We capture voter het- solutions. The firm may uncover an unexpected extreme-value solu- erogeneity through differences in brand appreciation and generate a tion thanks to a large number of opinions from diverse participants. sponsor score from a subset of the available votes. We find that even Crowds can also identify niche consumer interests which can lead to if voters do not influence sponsor choices, and sponsor preferences valuable innovations that the firm would not have considered oth- are representative of voter preferences, certain types of voters still erwise. emerge as more important than others. A statistical model with text Crowdsourcing campaigns can yield thousands of submissions data yields improvements in the survey setting as well, but there is that firms cannot exhaustively and carefully evaluate. For example, only subtle evidence of predictable inaccuracy in the particular simu- our data provider attracted 74,436 ideas over 153 ideation contests. lation we study because of limited nonlinearities in the relationship The scale of these campaigns raises the important question of how to between voter and sponsor scores. The survey confirms that a statis- identify the best ideas efficiently. Current approaches involve using tical model can exploit varying voter importance and redundancy, either the crowd itself or a set of internal employees to prescreen thereby reducing evaluation effort in this alternative setting, even ideas. However, the crowd-based approach is infeasible if the firm when there are no significant preference differences between voters prefers to keep submissions confidential. In this research, we focus and the sponsor. on the case where the firm uses internal evaluators to maintain sub- Overall, our findings suggest that firms can enhance their con- mission confidentiality and thereby abide with data protection laws tent curation processes, even if they crowdsource ideas on different or withhold the submissions it receives as a competitive advantage. topics over time. In assessing the performance of different voters, In our field setting, each idea was evaluated once by several employ- firms must consider how the voters perform in conjunction with a ees of the crowdsourcing platform in a prescreening stage, which model as opposed to independently. The “predictable inaccuracy” resulted in a total of 186,111 evaluations and consumed about 14 of certain voters, if it is present, can alter their relative importance. workdays per month of human effort in the median month. Our research aims to develop an efficient approach to identify The Power of Brand Selfies in Consumer-Generated the best ideas from crowdsourcing contests using data on the behav- Brand Images ior of internal evaluators, whom we refer to as “voters”, idea characteristics and text, and the selection of winning ideas by the clients, EXTENDED ABSTRACT whom we refer to as “sponsors”. We develop a model to reconcile Every day more than 5 billion images are shared on social me- the difference between the sponsor’s selection of finalist ideas and dia networks such as Facebook, Twitter, or Instagram. Of particular the evaluations of voters who prescreen the ideas. In our field setting, interest to marketers are images that feature brands and consumption the platform sorts ideas based on their average voter score before experiences that contribute approximately 1% of all social media im- presenting them to the sponsor. A model that transforms voter scores ages, resulting in 50 million social media images featured with brand and incorporates idea characteristics and text can yield a better score logos daily. At the same time, one of the biggest trends introduced by to sort ideas, thereby reducing the number of ideas that a sponsor smartphone cameras and social media are selfies. Today, more than must review to identify all finalists by 7-18%, depending on the spe- 383 million images with the hashtag #selfie exist on a single photo- cific evaluation approach. Equivalently, the platform can reduce the sharing platform such as Instagram. The emergence of selfies merits Advances in Consumer Research (Volume 48) / 833 the question of how brands appear in selfie images and how observ- Whether consumer selfies are well suited as a user-generated me- ers respond to this type of social media brand images. dia content is an open question, which we attempt to answer in this In fact, recent marketing practice attempts to capitalize on the research. selfie phenomenon, mainly by actively encouraging consumers to Analyzing consumers’ response to the different types of brand post selfies of their product encounters. Brand images are also of in- images in terms of likes and comments, we find that consumer self- terest to managers when passively listening in to social media posts. ies, in which a person appears with the brand, generate the highest Among other things, companies track brand logo presence on social level of engagement towards the image or the sender in terms of the media to understand social media popularity, rank consumer-gener- number of likes and comments on the image. However, these simple ated images on their social media brand page, or use such images as engagement measures, while encouraging for brands in terms of user part of their own marketing campaigns. potential, may be misleading. Examining the content of the user However, not all brand logo appearances are created equal. comments, using both dictionary-based and machine learning text Some may generate more valuable consumer-brand engagement mining tools, we find that consumer selfies generate fewer self-brand than others. Accordingly, the objective of this research is to inves- mentions and stated purchase intentions of receivers in response to tigate how brands appear in consumer-generated images and to ex- the original image post. These results are consistent with research amine the effectiveness of different types of social media brand im- from traditional advertising and information systems, which indicate ages on generating engagement among consumers. Specifically, we that, on the one hand, images with faces catch more attention than investigate both sender engagement objectives in terms of how brand those without faces (Bakhshi, Shamma and Gilbert 2014; Xiao and images may generate image engagement (i.e., likes or comments) Ding 2014), but, on the other hand, may detract attention away from and brand engagement objectives in terms of how brand images may the brand itself (Erfgen, Zenker and Sattler 2015). generate brand engagement (i.e., brand-related comments such as These results are consistent for both the Twitter and Instagram expressed purchase intent). dataset, suggesting that the effects are not driven by different con- Academic research provides ample evidence for the effective- sumer motives of the respective platforms. We complement these ness of images in advertising (e.g., Hanssens and Weitz 1980; Xiao results with a lab experiment, which allows us to control for the and Ding 2014). Recent studies have also explored the motivations prominence of the brand in the image and collect purchase intent to share content and take photos and how this affects subsequent ratings, as well as to sender behavior (e.g., Barasch, Zauberman and Diehl 2018; Grewal, test the underlying potential psychological mechanism that may Stephen and Coleman 2019). However, little is known about how trigger consumers’ varying reaction to different brand image types. observers respond to brand images in social media. The results of the experiment further suggest that brand selfies have a Using both manual annotations of a sample of images and con- superior impact on perceived purchase intent compared to consumer volutional neural networks (CNN) for automated image classifica- selfies. In addition, it suggests the differential impact of brand selfies tion of nearly half a million brand images related to 185 different is related to easier and more accessible self-reference and mental brands that were posted on Twitter and Instagram, we identify three simulation offered by these image types. types of user-generated brand images that differ in terms of human and facial presence. Consumers post either images of products in Black-Box Emotion Detection: isolation or holding products themselves (i.e., selfies). We find that On the Variability and Predictive Accuracy these brand-related selfies exist in two forms with consumers either of Automated Emotion Detection Algorithms visible (their face) or invisible to the viewer (e.g., first-person point of view of the product). We name the former consumer selfie and the EXTENDED ABSTRACT latter brand selfie to indicate the focus in brand selfies is exclusively Automated emotion detection from facial expressions refers to on the product and differentiate it from consumer selfies where the the use of algorithms that detect facial landmarks in pictures to clas- face of the sender is visible1. This results in the following typology sify people’s discrete emotions (Liu et al. 2014; Wood et al. 2016). of brand images: These automated emotion detection systems classify discrete posi- 1. Brand Selfies: branded products held by an invisible con- tive emotions such as happiness or surprise to negative emotions sumer, such as anger and fear. A computer vision algorithm first identifies 2. Consumer Selfies: visible consumer faces together with a facial landmarks in a picture and then assigns each picture a dis- branded product, crete emotion label based on the features or composition of the exist- 3. Packshots: standalone images of branded products. ing facial landmarks (Fox et al. 2000; Pantic and Rothkrantz 2000) development of an automated system that accomplishes this task is The CNN algorithm accurately classified images into these rather difficult. There are several related problems: detection of an three categories with a hold-out accuracy level of > 80%. In both image segment as a face, extraction of the facial expression informa- datasets, we find that consumer selfies have the lowest fraction of tion, and classification of the expression (e.g., in emotion categories. all types, suggesting that consumers are reluctant to post photos of Companies such as Microsoft, Google, or GfK provide platforms to themselves with the brand on their own accord. At the same time, perform such automated emotion detection with recent industry re- consumer selfies are the image type most often encouraged by corpo- ports suggesting a CAGR of 32.7% and a market size of 25 billion by rate communication campaigns, and face images, akin to consumer 2020, highlighting the importance and dominance of these AI-pow- selfies, are ubiquitous in print advertising (Xiao and Ding 2014). ered technologies for the future of marketing. Despite the increasing availability of automated emotion detection systems, fundamental 1 Note, we distinguish brand and consumer selfies based on the visibility methodological questions arise. Are automated emotion detection of consumer faces. Conceptually, the term selfie suggests the person on the systems valid? Is the same picture classified correctly across emo- image took the photo herself. Yet, in some cases, a third person may have tion detection systems? These questions are important as emotion photographed the sender (e.g., < 3% of brand selfies and < 39% of consumer detection systems are using pre-trained algorithms to classify dis- selfies in our Twitter data). This can be difficult to distinguish empirically for crete emotions from facial expressions in a way that is unknown to both the human eye and an automated image classifier. 834 / Consumer Behavior Meets Machine Learning the user of these systems. Thus, it is likely that the same picture is ative discrete emotions and substantially reduced variation across assigned a different discrete emotion conditional on the type of emo- positive, neutral and also negative emotion labels (σhappy open = 2.71, tion detection system being used. To the best of our knowledge, both σneutral = 1.36, σfear = 5.32, σanger = 4.61). a formal test of this hypothesis is non-existent as is a formal analysis To the best of our knowledge, this is the first systematic study that unravels under which conditions emotion detection algorithms demonstrating the striking variability in automated emotion detec- would increase in predictive accuracy. The current paper fills this tion systems across discrete emotions and we provide two easy to gap by providing a tightly controlled validation study comparing the implement modelling strategies to improve prediction accuracy effectiveness (i.e., in terms of predictive accuracy) of major auto- across automated emotion detection systems. mated emotion detection systems across discrete emotions, dependencies on the modelling technique being used, the stimuli set, and REFERENCES ways to improve predictive accuracy by combining feature sets or Bakhshi, Saeideh, David A. Shamma and Eric Gilbert (2014), using ensemble methods. “Faces Engage Us: Photos with Faces Attract More Likes We used a standardized picture set with objective ground truth, and Comments on Instagram,” in Proceedings of the ACM i.e. knowledge about the discrete emotion that is displayed by actors Conference on Human Factors in Computing Systems, in a standardized picture set. Specifically, we used the Chicago face Toronto, Canada: ACM, 965–974. database (Ma, Correll, and Wittenbrink 2015) to evaluate automated Barasch, Alixandra, Gal Zauberman and Kristin Diehl (2018), emotion detection systems. The Chicago face database contains pho- “How the Intention to Share Can Undermine Enjoyment: tos from 597 male and female targets of varying ethnicity between Photo-Taking Goals and Evaluation of Experiences,” Journal 18 and 40 years under standardized conditions. For a subset of 158 of Consumer Research, 44 (6), 1220–1237. targets, images display either a neutral, angry, fearful, or two positive Berger, Jonah, Ashlee Humphrees, Stephen Ludwig, Wendy emotional states (happy face with either an open or closed mouth). Moe, Oded Netzer, and David Schweidel (2020), “Uniting Each of these images of the 158 subjects displaying various emo- the Tribes: Using Text for Marketing Insight,” Journal of tions was classified using five major automated emotion detection Marketing, 84 (January), forthcoming. systems via an Application Programming Interface (API): Microsoft Berger, Jonah, and Grant Packard (2018), “Are Atypical Things Cognitive Services, Google ML, Sightcorp, Kairos, and the GfK More Popular?” Psychological Science, 29 (7), 1178–84. EmoScan. The data was normalized across emotion detection sys- Berger, Jonah, and Katherine L. Milkman (2012), “What Makes tems to provide meaningful comparisons in the following analyses. Online Content Viral?” Journal of Marketing Research, 49 (2), To set a baseline, we evaluated the effectiveness of the different 192–205. emotion detection systems by predicting each image emotion label Berger, Jonah, Wendy W. Moe, and David A. Schweidel (2019), via the corresponding emotion measure (i.e., predicting happy faces “What Makes Stories More Engaging? Continued Reading in using the happiness feature of each emotion detection system). A Online Content,” Working Paper. multinomial logit model revealed substantial variation across emo- Berman, Ron, Shiri Melumad, Colman Humphrey, and Robert tion detection systems with the highest, cross-validated (CV) predic- Meyer (2019), “A Tale of Two Twitterspheres: Political tion accuracy for Google ML with a test accuracy of 58.94% and a Microblogging During and After the 2016 Primary and test Kappa of 40.25%, compared to the lowest prediction accuracy Presidential Debates,” Journal of Marketing Research, 56(6), for GfK EmoScan with a test sample accuracy of 49.81% and a test 895–917. Kappa of 36.88%. Positive emotions were predicted consistently Eliashberg, Jehoshua, Sam K. Hui, and Z. John Zhang (2007), better compared to the substantial variation across negative emotion “Assessing Box Office Performance Using Movie Scripts: A Kernel-Based Approach,” IEEE Transactions on Knowledge & labels (variability across emotion detection systems: σhappy open = 3.81, Data Engineering, 26 (11), 2639–48. σneutral = 6.89, σfear = 14.98, σanger = 30.9) and overall lower predictive power for negative emotions than for positive and neutral emotions Erfgen, Carsten, Sebastian Zenker and Henrik Sattler (2015), “The Vampire Effect: When Do Celebrity Endorsers Harm Brand (average balanced accuracy for the different emotion classes: happyµ Recall?”, International Journal of Research in Marketing, 32 open = 78.97, µneutral = 79.97, µfear = 65.45, µanger = 44.3). These findings were robust even after controlling for all other emotion features. (2), 155–163. Next, we aimed at further improving prediction accuracy in two Fox, Elaine, Victoria Lester, Riccardo Russo, R. J. Bowles, Alessio ways. First, by increasing the feature space through combining the Pichler, and Kevin Dutton (2000), “Facial expressions entire set of features across all emotion detection systems. Second, of emotion: Are angry faces detected more efficiently?,” by using more flexible model estimation and machine learning tech- Cognition and Emotion. niques. Specifically, recent advances in deep learning might further Gibson, Edward (1998), “Linguistic Complexity: Locality of improve predictive accuracy by increasing the number of layers and Syntactic Dependencies,” Cognition, 68(1), 1–76. / or boosting a model by aiming at predicting misclassified cases. In Grewal, Lauren, Andrew T. Stephen and Nicole V. Coleman (2019), short, increasing the feature space resulted in a greater test accuracy “When Posting About Products on Social Media Backfires: with an averaged accuracy of 85.73% and a Kappa of 81.02%. The Negative Effects of Consumer Identity Signaling Next, we aimed at further improving prediction accuracy by on Product Interest,” Journal of Marketing Research, using more flexible machine learning techniques. Parameters were forthcoming. selected through grid search in the parameter space. Lowest perfor- Hanssens, Dominique M. and Barton A. Weitz (1980), “The mance was achieved by Support Vector Machines with radial basis Effectiveness of Industrial Print Advertisements Across kernel resulting in a test accuracy of 84.41% and a test Kappa of Product Categories,” Journal of Marketing Research, 17 (3), 79.33%. Highest performance was achieved by using Random For- 294–306. ests and LogitBoost achieving 92.02% test accuracy, 89.25% test kappa and 93.60% accuracy, 91.34% kappa, respectively. These changes led to a significant increase in predictive power also for neg- Advances in Consumer Research (Volume 48) / 835

Humphreys, Ashlee, and Kathryn A. LaTour (2013), “Framing the Proserpio, Davide and Hauser, John R. and Liu, Xiao and Amano, Game: Assessing the Impact of Cultural Representations on Tomomichi and Burnap, Alex and Guo, Tong and Lee, Consumer Perceptions of Legitimacy,” Journal of Consumer Dokyun (DK) and Lewis, Randall A. and Misra, Kanishka Research, 40(4), 773–795. and Schwartz, Eric M. and Timoshenko, Artem and Xu, Lilei Liu, Ping, Shizhong Han, Zibo Meng, and Yan Tong (2014), “Facial and Yoganarasimhan, Hema, Soul and Machine (Learning) expression recognition via a boosted deep belief network,” in (September 16, 2019). Available at SSRN: https://ssrn.com/ Proceedings of the IEEE Conference on Computer Vision and abstract=3454294 or http://dx.doi.org/10.2139/ssrn.3454294 Pattern Recognition, 1805–12. Wood, Adrienne, Magdalena Rychlowska, Sebastian Korb, and Ma, Debbie S, Joshua Correll, and Bernd Wittenbrink (2015), Paula Niedenthal (2016), “Fashioning the Face: Sensorimotor “The Chicago face database: A free stimulus set of faces and Simulation Contributes to Facial Expression Recognition,” norming data,” Behavior research methods, 47(4), 1122–35. Trends in Cognitive Sciences. Pantic, Maja, and Leon Ü M. Rothkrantz (2000), “Automatic Xiao, Li and Min Ding (2014), “Just the Faces: Exploring the analysis of facial expressions: The state of the art,” IEEE Effects of Facial Features in Print Advertising,” Marketing Transactions on Pattern Analysis and Machine Intelligence. Science, 33 (3), 338–352. AI Agents in Consumer Service – Tools or Partners? Chairs: Marat Bakpayev, University of Minnesota Duluth, USA Ann Kronrod, University of Massachusetts Lowell, USA

Paper #1: Consumer Interactions with Artificial Intelligence behavioral intentions, as robots shift away from operating as tools to Agents – a Historical Perspective being partners in decision making. Finally, Bakpayev and Kronrod Marat Bakpayev, University of Minnesota Duluth, USA take a linguistic perspective on conversations with AI agents. Look- Russell Belk, York University, Canada ing at the use of figurative language (humor, metaphor) by AI/human service agents and perceptions of conversational cooperativeness, Paper #2: Are Robots in Service of the Environment? The Role the authors find that while human agents are expected to use figura- of Service Robots in Socially Responsible Business Practices tive language because they are perceived as cooperative speakers, AI Marina Puzakova, Lehigh University, USA agents are harmed when they employ figurative language. Amir Grinstein, Northeastern University, USA Taken together, the papers in this session will examine the same Paper #3: The Impact of Humans and Robots on Consumer topic of AI agents as tools or partners but from different perspectives. Behavioral Intentions in Medical Contexts They collectively contribute to the discussion about the role of AI Ilana Shanks, Stony Brook University, USA in consumer service from a historical, CSR, medical, and linguistic Martin Mende, Florida State University, USA perspectives. We, therefore, expect the session to attract a wide vari- Maura L. Scott, Florida State University, USA ety of attendees interested in the integration of artificial machines in Jenny van Doorn, University of Groningen, The Netherlands consumer service in various contexts. Dhruv Grewal, Babson College, USA Paper #4: AI Service Agents, Figurative Language, and Consumer Interactions with Artificial Intelligence Agents Conversational Cooperativeness – a Historical Perspective Marat Bakpayev, University of Minnesota Duluth, USA Ann Kronrod, University of Massachusetts Lowell, USA EXTENDED ABSTRACT “Replicants are like any other machine. They’re either a benefit SESSION OVERVIEW or a hazard. If they’re a benefit, it’s not my problem.” Consumers are increasingly interacting with AI agents, ask- Rick Deckard, Blade Runner. ing for suggestions from chatbots, talking to service robots at hotels Historically, whether in fiction or fact, humans were captivated or restaurants, and telling their AI assistants - be it Alexa, Siri, or by the idea of creating human-like entities (Belk 2016; Kang 2011; Cortana – “I love you” (Belk 2019; Belk, Humayun, and Gopaldas Mayor 2018; Nocks 2008; Riskin 2016). Replicants, as depicted in 2020; Novak and Hoffman 2018; van Doorn et al. 2017; Mende et the movie Blade Runner (1982), are just one example of artificial al. 2019). The nature of these interactions is not yet fully explored. beings that appear identical to human beings. Philip Dick (1968) in Therefore, this session is aimed to bring together works that explore his novel Do Androids Dream of Electric Sheep? – which serves as interactions with AI agents from a multifaceted, multimethod point a primary source for the film – refers to them as “andys” or androids of view. (Greek root “man-like”). Both the novel and the film portray a future We see AI agents from the perspective of slaves, masters, circa 2020; that is, now. friends, partners, or tools (Elder 2017; Breazeal et al. 2004; Gru- While machines in our time are not yet the replicants or an- din 2017; Schweitzer et al. 2019). The consumer research literature droids imagined in Dick’s science-fiction, some of them are remark- mostly examines AI agents as tools, as well as the tendency of people ably human-like and engage us in interactions that are increasingly to humanize/anthropomorphize machines. The papers in this session human in nature. For example, consumers voice commands to their explore a broader range of roles and perspectives that consumers take digital assistants like Alexa; consumers chat with AI friends like toward intelligent machines. In line with the spirit of the conference, Replika; consumers look into the “eyes” of the robot Pepper, who the four papers integrate multiple perspectives and methods, exam- greets them in service environment, and reply “Hello!” ining aspects of interactions with AI agents via a historical lens, a Overall, across multiple studies, consumer researchers highlight socially responsible business lens, a medical lens, and a linguistic anthropomorphic aspects of technology when these nonhumans are lens. Together, these papers illustrate how robots and AI agents take endowed with human-like traits (Epley 2018). But what are the ori- on different roles in the lives of consumers – as tools or as partners. gins and consequences of humanizing AI? While the literature pro- Bakpayev and Belk open the session with a historical overview vides various scenarios and results, there is still a great deal to be of interactions with robots and AI. The authors conceptualize types learned as science fiction becomes science fact. Taking a historical of interactions that people have or would have with machines, and perspective, we look at the nature of our relationships with human- discuss the changing landscape of seeing machines as servants, mas- like machines and the consequences of humanizing them. ters, monsters, or friends. Next, Puzakova and Grinstein focus on The term “robot” comes from a science-fiction stage play. Karel robots in socially responsible business practices. Across three ex- Čapek (1920/2004) wrote R.U.R. (Rossum’s Universal Robots) one- periments, the authors demonstrate and explain the negative effect hundred years ago. In the play, robots and robotesses are artificial of employing service robots (vs. humans) on consumers’ willingness people, forced into labor as secretaries, postmen, and workers. “Ro- to engage with the brand in socially responsible practices, as well as bot” derived from Old Slavic “rab” or slave. The serving purpose consumers’ perceptions of the brand as being socially responsible. was quite prevalent in the early history of artificial life. While not Third, Shanks, Mende, Scott, van Doorn, and Grewal focus on robots referred to as “robots” yet, humans were developing figurines, stat- as partners in a medical context and examine the effect of a robot ues, mechanical knights, stone giants, and automatons that served taking more or less prominent roles in medical environments. This more as tools (Belk 2016; Kang 2011; Mayor 2018; Nocks 2008; research examines the underlying mechanisms that impact consumer Riskin 2016). Researchers have traced other perspectives as well. Advances in Consumer Research 836 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 837 Belk, Humayun, and Gopaldas (2020) explore the history of Artifi- Are Robots in Service of the Environment? The Role of cial Life (AL) and connect it to four cultural myths: Pygmalion (AL Service Robots in Socially Responsible Business Practices as a product of human desire), Golem (AL as a product of human survival), Frankenstein (AL as a product of human curiosity), and EXTENDED ABSTRACT Metropolis (AL as a product of human politics). These myths show Two of the most meaningful societal trends today involve the deeper insights into our fears and fascinations with artificial beings. introduction of robots in services (Mende et al. 2019) and brands’ While there is a degree of tool-oriented practicality in these visions, increased engagement with socially responsible practices (i.e., CSR; there are also deeper orientations toward perfection and immortality. Mishra and Modi 2016). Importantly, the two trends have not been The idea of the creation of human-like machines took a new studied jointly but we argue that they are intertwined and can alter turn with the development of computers. During a nearer period in consumption. Here, we aim to fill this gap and address multiple un- history referred to as the “AI Spring” in the 1940s and 1950s, ma- answered research questions: How the introduction of a service robot chines for the first time seemed intelligent. Drawing on early com- (vs. human) impact: (a) consumer perception of and (b) consumer puting and Norbert Weiner’s cybernetics, Alan Turing proposed a engagement with the brand’s socially responsible practices? What test (the Turing Test) that could determine when human-like AI was is the underlying mechanism that drives consumer response to a ser- able to pass for being human. The field of Human-Computer Interac- vice robot? tion started to develop (Grudin 2017; Haenlein and Kaplan 2019). Research on consumer reactions to service robot providers is an Machines - and technology in general - were stills tools, more like emerging inquiry and has discussed, for example, the impact of ser- extensions of our arms, eyes, and minds. However, as AI went into vice robots on consumers’ social surroundings and service outcomes “Summer” and “Winter” (the 1960s and 1970s), consumers realized (e.g., van Doorn et al. 2017), and consumption patterns (e.g., Mende that their grand expectations for the machines failed. et al. 2019). In turn, the literature on consumers’ reactions to brands’ Similarly, in recent years, with the second “AI Spring” and the CSR practices suggests that consumers believe that business bears threat of a second “AI Winter,” reactions to computers started to responsibility for driving positive social change, and they encour- change. Arguably, we are in the stage of Human-Computer Sym- age firms to improve society and the environment while achieving biosis, where machines are seen more like partners (Breazeal et al. firm goals (Porter and Kramer 2011). Supporting this trend, research 2004; Grudin 2017). Designed to facilitate social interactions, they shows that in general consumers perceive brands’ CSR efforts posi- can be our companions and friends (Elder 2017; Schweitzer et al. tively (Luo and Bhattacharya 2006). Relatedly, some research stud- 2019). Nevertheless, at the end of his career, Weiner warned that ies situations where brands wish to engage consumers in socially like the Golem of the Rabbi of Prague, we might be creating ma- responsible behavior (e.g., recycle, donate; Goldstein et al. 2008, Li- chines that threaten our continued existence. Slaves could become chtenstein et al. 2004) and the conditions under which consumers are masters (Schweitzer et al. 2019). Ideological tensions develop as we more likely follow the brand’s requests (Romani and Grappi 2014). increasingly humanize AI. As depicted in the opening quote from Overall, how would consumers be influenced by service robots in the Blade Runner, further development of AI agents could indeed lead context of brands’ social responsibility? to potential benefits for humankind (Markoff 2015), as well as haz- Relatively to humans, robots are perceived as more efficient, ards, noted by some visionaries like Bill Gates, Stephen Hawking, productive, and generally less warm. Although technological devel- and Elon Musk. Huang and Rust (2018; 2020) specify types of ser- opments imbue robots with increasing automated social presence vice jobs that AI can take on – mechanical, analytical, intuitive, and and humanness (van Doorn et al. 2017), robots are still perceived empathetic – leading to mechanical AI, thinking AI, and feeling AI. only as partial social actors and less warm than humans (Heerink et Indeed, despite significant accomplishments and progress in AI, or al. 2010). Here, we argue that introduction of service robots can lead perhaps due to the progress, research confirms concerns regarding to a greater experience of psychological coldness compared to situ- AI, finding that reactions to humanoid service robots can be discom- ations when customers are served by humans. Given service robots forting (Mende et al. 2019) and uncanny (Kim, Schmitt, and Thal- represent and reflect on the brand, brands that use service robots (vs. mann 2019). It appears that consumers resist automation, especially humans) will be perceived as colder. This process can have unintend- if it threatens their identity (Leung, Paolacci, and Puntoni 2018). ed consequences for consumers’ perceptions of socially responsible In sum, while past perspectives toward human-like machines business practices. That is, prior work establishes the negative rela- focused on ways in which they serve to boost humanity, existential tionship between psychological experience of coldness and prosocial fears about AI seem to be rising in recent years. We worry that al- behaviors (Zhong and Leonardelli 2008). Thus, we postulate that a gorithms threaten our privacy, security, wealth, and opportunity. We “colder” experience will result in consumers’ less positive views on fear that robots will displace us in the workforce and that our ma- a brand’s social responsibility practices and their own willingness to chines will become smarter and stronger than us – in the end, taking engage with the brand on social responsibility. over, and even extinguishing the human species. Based on historical H1: Brands’ use of service robots (relatively to humans) is perspectives and scenarios involving AI agents, the authors consider negatively related to consumer (a) perception of the brand’s social the possible implications of current transformations in consumer in- responsibility, and (b) engagement with the brand’s socially respon- teractions with artificially intelligent agents. Consumers are apt to sible practices. regard these machines as servants, friends, monsters, or masters. As H2: The influence of service robots (relatively to humans) on a result, interactions may be functional or social, calming or fright- consumer reaction is mediated by lower brand warmth perceptions. ening, successful or failed. Such developments contribute to further We test these hypotheses in three studies. Study 1 (N=226, stu- discussions on ethical, moral, legal, and existential questions that are dents) used the context of a hotel bellman service. Participants were critically important for future consumer research. prompted to imagine that upon their check-in, they were greeted by a hotel bellman (by a service robot) that assisted carrying their lug- gage and accompanied them to their rooms. After respondents were assisted by a hotel’s bellman (robot), they further imagined finding a card on their desk describing that the Ario hotel (a fictitious brand) 838 / AI Agents in Consumer Service – Tools or Partners? partnered together with a local charity organization “the Feed the ing a decrease in post-surgical complications and shorter hospital Hungry Foundation”, after which respondents indicated their infer- stays (Kalis, Collier, and Fu 2018; Marr 2018). Previous research has ences about the sincerity of the company’s motives for pursuing the proposed several mechanisms that may impact consumer response CSR activity (Yoon et al. 2006), rated brand warmth (Aaker et al. including anxiety (Broadbent et al. 2011; Mende et al. 2019; Mori 2010), and the service novelty. 1970), social dimensions (van Doorn et al. 2017; Wirtz et al. 2018), The results of study 1 (a one-way (robot vs. human service) and various dimensions of robotic threat (Stein, Liebold, and Ohler ANCOVA with the novelty of service experience as the covariate) 2019; Złotowski, Yogeeswaran, and Bartneck 2017). The use of ad- revealed that experiencing a robot (vs. human) service decreased vanced technology, such as artificial intelligence (AI), in healthcare consumers’ perceptions of the sincerity of the hotel’s motives to en- settings similarly results in adverse consumer reactions due to a be- gage in CSR (Mrobot=4.88 vs. Mhuman=5.54; F(1,223) = 7.45, p = .007). lief that AI is less able to adjust for individual situations (Longoni, The decrease in the sincerity of the motives was driven by lowered Bonezzi, and Morewedge 2019). Moreover, this research examines perceptions of brand warmth (b=–.46, 95% CI[–.77, –.17]). the underlying mechanisms that impact consumer behavioral inten- Next, study 2 (N=206, MTurk participants) replicates the effect tions as robots shift away from operating as tools. An examination of with a different social responsibility domain (environmental con- boundary conditions investigates various effects of risk acceptance, cerns) and examines whether robot (human) service also decreases power, and anthropomorphism on consumer behavioral intentions. consumer engagement with the brand’s socially responsible prac- Nine studies examine the effect of a robot in a more or less tices (i.e., willingness to pay an extra-fee for sustainable breakfast). prominent role in medical environments; two studies involve live in- The results of this study revealed that being assisted by a service teractions with an actual robot. Study 1A examines the effectiveness (human) robot decreased consumers’ own desire to engage in sus- of a human-robot team. Patients at an elderly care facility partici- tainable behavior (Mrobot=3.62 vs. Mhuman=4.28; F(1,203) = 4.47, p < pated in exercise sessions with either a robot with a human physical .036). We also found that consumers perceived the motives as less therapist or a human physical therapist alone. Participants had an sincere for a service robot (Mrobot=4.95 vs. Mhuman=5.57; F(1,203) = unfavorable reaction to the human-robot team (vs. human), rating the 7.29, p < .008). Both of these effects were driven by lower brand exercises as less favorable (p < .01) and reporting lower behavioral warmth (lower desire to engage in sustainable consumption: b=–.23, responses (p = .04). Study 1B replicates these findings. Participants 95% CI [–.50, –.004]); sincerity of the motives: b=–.60, 95% CI imagined going to a doctor for an examination while viewing a pic- [–.94, –.34]). ture of a human or a human-robot medical team. Participants again Study 3 (N=196, MTurk participants) rules out alternative pro- had an unfavorable reaction to the human-robot team, and lower rat- cesses through perceived self-efficacy and uncertainty (Chen 2016; ings of loyalty intentions (p < .001), warmth (p < .001), and compe- van der Wal et al. 2018). Replicating results of studies 1 and 2, the tence (p = .02). findings revealed that being assisted by a service (human) robot led to Study 2 examines the effect of a human-robot team on a con- lower inferences of the motive sincerity (Mrobot=4.69 vs. Mhuman=5.78; sumer’s response to a robot-led team. Participants received a nutri- F(1,193) = 19.28, p < .007) and reduced consumer engagement with tional counseling session from a live, in-person human-robot team a brand’s socially responsible practices (Mrobot=4.01 vs. Mhuman=4.71; with either the human or the robot as the team leader. The results re- F(1,193) = 7.48, p < .007). Next, testing parallel mediators (i.e., vealed that loyalty intentions (p < .001), and perceptions of warmth brand warmth, self-efficacy, uncertainty; model 4; Hayes 2013) re- (p < .001) and competence (p = .03) were lower when the team was vealed a significant indirect path through brand warmth (sincerity of led by a robot (vs. a human). the motives: b=–.73, 95% CI [–1.08, –.46]; consumer engagement Study 3 examines the moderating role of risk acceptance (mea- with CSR: b=–.41, 95% CI [–.69, –.18]), with both self-efficacy and sured) using a video of a human-robot team. The results revealed that uncertainty being nonsignificant. participants with low levels of risk acceptance showed a decrease Overall, this research identifies important consequences of in- in loyalty intentions and ratings of warmth and competence to the troducing robots into service contexts. In particular, three studies robot-led team (vs. human). The effects were attenuated for partici- demonstrate that a robot (vs. human) service lowers consumers’ per- pants high in risk acceptance. ceptions of the sincerity of the motives for the CSR activity and re- Study 4 examines the moderating role of power distance belief duces consumers’ own engagement with the brand’s CSR practices. (PDB). Power distance belief was measured. Participants watched Our work sheds light on the underlying process through reduced at- the same video as in Study 3. The results revealed an increase in tributions of brand warmth, while also ruling out alternative explana- unfavorability to a robot-led team for participants low in PDB and tions through self-efficacy and uncertainty. an attenuation for participants high in PDB. Studies 5A and 5B examine the role of perceived choice on re- The Impact of Humans and Robots on Consumer sponse to a human-robot team. In Study 5A, participants chose the Behavioral Intentions in Medical Contexts robot used in the medical team in the choice condition and were unable to choose the robot used in the control condition. In the control EXTENDED ABSTRACT condition, participants had lower loyalty intentions to the robot-led As the use of robots in the marketplace increases, both the mar- (vs. human-led) team. This was attenuated in the choice condition. keting and robotics literatures have examined how consumers will Study 5B examined the role of choice on upgrading behavior, finding respond. The literature has mostly examined this phenomenon from greater interest in paying to upgrade to a human-human team when the lens of a robot in the position of a tool. However, technologi- the initial team was led by a robot. cal advances are resulting in robots that can begin to assume an in- Studies 6A and 6B generalize the effect to anthropomorphized creased variety of roles in marketing settings. As robots shift away robots. Although the results did not reveal effects of anthropo- from being merely tools that assist consumers, it is important to un- morphism, participants had lower loyalty intentions and ratings of derstand how consumers will respond. warmth and competence to the robot-led team (vs. human-led), con- We examine this phenomenon in a medical context, as research sistent with our previous studies. suggests positive outcomes of the increased use of robots includ- Advances in Consumer Research (Volume 48) / 839

The current research uncovers insights into how consumers re- The purpose of Study 2 was to compare reactions to serious spond to robots as they move away from the role of tools. Our find- or humorous answers by human or artificial agents. 427 MTurk ings have implications for marketing strategy as we demonstrate that participants were randomly assigned to one of 4 conditions in a 2 this shift results in an unfavorable consumer response and strategies (human/artificial agent) by 2 (serious/humorous answer) between- to mitigate this negative response. subjects design. 2-way ANOVA revealed a significant interaction of agent (human/artificial) and answer (humorous/serious) on booking AI Service Agents, Figurative Language, and intention (F(1,423)=7.47, p=.007), and on answer humorousness Conversational Cooperativeness (F(1,423)=8.7, p=.003). Planned contrasts showed that while for human agents there was no significant difference in booking inten- EXTENDED ABSTRACT tion whether the agent was joking (M=5.66) or serious (M=5.78, Firms are improving AI service agents’ conversational capaci- F(1,423)=.35, p=.55), when the agent was artificial, booking inten- ties, striving to make their language more natural and more similar tion was significantly lower when the agent was joking (M=4.77) to that of humans (e.g., Google’s Meena, the open-domain chatbot, than when the agent was serious (M=5.65, F(1,423)=19.9, p<.001). Adiwardana et al. 2020). But how do consumers react to this human- These results support H1 and H2. like conversational ability? Previous research examined reactions to In Study 3, we aimed to measure the mediation of conversa- AI with external similarity to humans (Davenport et al. 2020; van tional cooperativeness expectations. 202 MTurk participants imag- Doorn et al. 2017; Huang and Rust 2020; Mende et al. 2019). To ined interacting with a human(artificial) service agent, for example, date, research has not inquired into the effect of conversational simi- calling a company to book a hotel, or having an online chat to solve larity – one of the most distinctive human capacities – on perceptions a technical problem with a product (e.g., kitchen appliance) and of AI agents, and their effects on service. rated their expectations about the agent’s behavior in the conversa- Addressing this question, we examine reactions to AI service tion. Respondents expected human agents to be more cooperative agents employing human-like conversation. Specifically, we com- (M=5.58) than artificial agents (M=5.09, F(1,200)=7.389, p=.007). pare reactions to the use of figurative language (metaphor, humor) They also had significantly higher expectations from human agents – a distinctly human capacity - by human and AI service agents. than artificial agents on Communication Openness (F(1,200)=6.145, Figurative language requires the speaker to be conversation- p=.014) and General Cooperativeness (F(1,200)=4.288, p=.040). To ally cooperative, making assumptions about the other person’s un- explore the effect of agent type (human/artificial) and cooperative- derstanding (Grice 1975). We suggest that consumers expect human ness expectations on language expectations, a multiple regression service agents to be conversationally cooperative, but not AI service was executed. The overall model was significant (F(3,198)=10.297, agents. These conversational expectations, in turn, attenuate the pos- p<.001; R2=13.5%, adjusted R2=12.2%), suggesting that AI agents itive effect of using figurative language by AI service agents, com- are expected to be less cooperative than human agents. pared to literal language, and compared to human agents. Formally, Finally, Study 4 introduces a boundary condition: in a situa- we suggest that: tion of negative service conversations (e.g., complaints), the differences between human and AI service agents are attenuated. 809 Hypothesis 1: AI service agents who use figurative language MTurk participants were randomly assigned to one of 8 conditions evoke less positive reactions than human service in a 2(human/artificial agent) by 2(metaphorical/literal language) by agents who use figurative language. 2(positive/negative conversation) between-subjects design. Partici- pants read a conversation excerpt where they (the consumer) were Hypothesis 2: AI service agents who use figurative language talking with a human (AI) customer service agent who used literal evoke less positive reactions than when they use language/metaphor in a positive/negative context (choosing a free literal language. t-shirt/replacing a wrong t-shirt). They then indicated a likelihood to shop again, rated the retailer and agent’s cooperativeness. We found a three-way interaction on likelihood to shop again (F(1,801)=8.337, Hypothesis 3: These differences are driven by perceptions of p=004), retailer rating (F(1,801)=10.162, p=.001) and conversation the agents’ conversational cooperativeness: AI cooperativeness (F(1,801)=6.047, p=.014). Contrast analyses within service agents are perceived as significantly less the positive interaction condition showed, as before, that while for conversationally cooperative than human ser- human agents language did not influence behavior intentions, retailer vice agents. evaluation and cooperativeness, for AI agents behavior intentions, and retailer evaluation were significantly lower when they used We explore this effect via four studies. In Study 1, a preliminary metaphor versus literal language However in the negative condition test, we compared reactions to human and artificial service agent hu- this effect disappeared. Mediation analyses showed that agent con- morous response to a customer’s inquiry. 205 MTurk participants versational cooperativeness mediates this effect (B=.289, SE=.064, read a brief dialogue between the customer (themselves) and a hu- t=4.56, p=.0001, 95%CI: [.165,.414]). man/AI hotel booking agent, which ended with the agent making a To sum, research on conversational norms in consumer research joke. Participants then reported intentions to book a room with the suggests that adherence to conversational norms, and specifically us- agent, their comfortableness with and trust towards the agent, and ing figurative language, is positively perceived (e.g., Kronrod and how funny was the joke. T-tests comparing the two groups on the dif- Danziger 2013; Xu and Wyer 2010; Zhang and Schwarz 2012). We ferent variables revealed no significant differences in booking inten- discover that when it comes to AI service agents – this relation is tion (p=.103), but overall participants were more likely to book with moderated. We also find that the negative effect of using figurative the human (M=5.64) rather than the artificial agent (M = 5.33). We language by AI agents is at least partially due to consumers not per- found marginal differences in trust (p=.098), comfort (p=.099) and ceiving them as cooperative conversation partners. Future work can satisfaction with the conversation (p=.083). Across all variables, the further explore AI agents as conversation partners in various con- human agent was rated higher than the artificial agent. sumer service situations and inform marketers on the way to avoid 840 / AI Agents in Consumer Service – Tools or Partners? unwelcome outcomes of the race towards human-like AI conversa- Haenlein, Michael, and Andreas Kaplan (2019), “A Brief History tion. of Artificial Intelligence: On The Past, Present, and Future of Artificial Intelligenc,”California Management Review, 61(4), REFERENCES 5-14. Adiwardana, Daniel, Minh-Thang Luong, David R. So, Jamie Hall, Heerink, Marcel, Ben Kröse, Vanessa Evers, and Bob Wielinga Noah Fiedel, Romal Thoppilan, Zi Yang (2020), “Towards (2010), “Relating Conversational Expressiveness to Social a Human-like Open-domain Chatbot,” arXiv preprint Presence and Acceptance of an Assistive Social Robot,” arXiv:2001.09977. Virtual Reality, 14(1), 77-84. Belanche, Daniel, Luis V. Casaló, Carlos Flavián, and Jeroen Hoffman, Donna L., and Thomas P. Novak (2018), “Consumer and Schepers (2019), “Service Robot Implementation: a Object Experience in the Internet of Things: An Assemblage Theoretical Framework and Research Agenda,” The Service Theory Approach,” Journal of Consumer Research, 44(6), Industries Journal, (2). 1178-1204. Belk, Russell (2016), “Understanding the Robot: Comments on Huang, Ming-Hui, and Roland T. Rust (2018), “Artificial Goudey and Bonnin,” Recherche et Applications en Marketing Intelligence in Service.” Journal of Service Research, 21(2), (English Edition), 31(4), 83-90. 155-172. Belk, Russell (2019), “Machines and Artificial Intelligence,” Huang, Ming-Hui, and Roland T. Rust (2020), “Engaged to Journal of Marketing Behavior, 4(1), 11-30. a Robot? The Role of AI in Service,” Journal of Service Belk, Russell (forthcoming), “Ethical Issues in Service Robotics Research. and Artificial Intelligence,”Services Industries Journal. Kalis, Brian, Matt Collier, and Richard Fu (2018), “10 Promising Belk, Russell, Mariam Humayun, and Ahir Gopaldas (2020), AI Applications in Health Care,” Harvard Business Review, “Artificial Life,”Journal of Macromarketing. http://www.ajronline.org/doi/abs/10.2214/AJR.11.7522. Breazeal, Cynthia, Jesse Gray, Guy Hoffman, and Matt Berlin Kang, Minsoo (2011), Sublime Dreams of Living Machines, (2004), “Social Robots: Beyond Tools to Partners,” In RO- Harvard University Press. MAN 2004. 13th IEEE International Workshop on Robot Kaplan, Andreas, and Michael Haenlein (2019), “Siri, Siri, in my and Human Interactive Communication (IEEE Catalog No. Hand: Who’s the Fairest in the Land? On the Interpretations, 04TH8759), 551-556, IEEE. Illustrations, and Implications of Artificial Intelligence,” Broadbent, Elizabeth, In Lee Yong, Rebecca Q. Stafford, I. Han Business Horizons, 62(1), 15-25. Kuo, and Bruce A. MacDonald (2011), “Mental Schemas of Kim, Seo Young, Bernd H. Schmitt, and Nadia M. Thalmann Robots as More Human-Like Are Associated with Higher (2019), “Eliza in the Uncanny Valley: Anthropomorphizing Blood Pressure and Negative Emotions in a Human-Robot Consumer Robots Increases their Perceived Warmth but Interaction,” International Journal of Social Robotics, 3(3), Decreases Liking,” Marketing Letters, 30(1), 1-12. 291–97. Kronrod, Ann and Shai Danziger (2013), “Wii Will Rock You!” Čapek, Karel (1920/2004), R.U.R. (Rossum’s Universal Robots), The Use and Effect of Figurative Language in Consumer New York, NY: Penguin. Reviews of Hedonic and Utilitarian Consumption,” Journal of Chen, Mei-Fang (2015), “Self-efficacy or Collective Efficacy Consumer Research, 40(4), 726-739. within the Cognitive Theory of Stress Model: Which More Leung, Eugina, Gabriele Paolacci, and Stefano Puntoni (2018), Effectively Explains People’s Self-reported Proenvironmental “Man versus Machine: Resisting Automation in Identity-based Behavior?,” Journal of Environmental Psychology, 42, 66-75.‏ Consumer Behavior,” Journal of Marketing Research, 55(6), Davenport, Thomas, Abhijit Guha, Dhruv Grewal, and Timna 818-831. Bressgott (2020), “How Artificial Intelligence Will Change the Lichtenstein, Donald R., Minette E. Drumwright, and Bridgette M. Future of Marketing,” Journal of the Academy of Marketing Braig (2004), “The Effect of Corporate Social Responsibility Science, 48(1), 24-42. on Customer Donations to Corporate-supported Nonprofits,” van Doorn, Jenny, Martin Mende, Stephanie M. Noble, John Journal of Marketing, 68(4), 16-32.‏ Hulland, Amy L. Ostrom, Dhruv Grewal, and J. Andrew Longoni, Chiara, Andrea Bonezzi, and Carey K. Morewedge Petersen (2017), “Domo Arigato Mr. Roboto: Emergence (2019), “Resistance to Medical Artificial Intelligence,”Journal of Automated Social Presence in Organizational Frontlines of Consumer Research, 46 (4), 629-50. and Customers’ Service Experiences,” Journal of Service Luo, Xueming, and Chitra Bhanu Bhattacharya (2006), “Corporate Research, 20(1), 43-58. Social Responsibility, Customer Satisfaction, and Market Elder, Alexis (2017), Friendship, Robots, and Social Media: False Value,” Journal of Marketing, 70(4), 1-18. Friends and Second Selves, Routledge. Markoff, John (2015), Machines of Loving Grace: The Quest for Epley, Nicholas (2018), “A Mind like Mine: The Exceptionally Common Ground Between Humans and Robots, New York: Ordinary Underpinnings of Anthropomorphism,” Journal of HarperCollins. the Association of Consumer Research, 3(4), 591-98. Marr, Bernard (2018), “How Is AI Used In Healthcare - 5 Powerful Goldstein, Noah J., Robert B. Cialdini, and Vladas Griskevicius Real-World Examples That Show The Latest Advances,” (2008), “A Room With a Viewpoint: Using Social Norms to Forbes, https://www.forbes.com/sites/bernardmarr/2018/07/27/ Motivate Environmental Conservation in Hotels,” Journal of how-is-ai-used-in-healthcare-5-powerful-real-world-examples- Consumer Research, 35(3), 472-482.‏ that-show-the-latest-advances/#62a1c1f75dfb. Grice, Herbert P (1975), “Logic and Conversation,” Speech Acts, Mayor, Adrienne (2018), Gods and Robots: Myths, Machines, and 41-58, Brill. Ancient Dreams of Technology, Princeton University Press. Grudin, Jonathan (2017), “From Tool to Partner: The Evolution of Human-Computer Interaction,” Synthesis Lectures on Human- Centered Interaction, 10(1). Advances in Consumer Research (Volume 48) / 841

Mende, Martin, Maura L. Scott, Jenny van Doorn, Dhruv Grewal, Van der Wal, A. J., Arianne J., Femke van Horen, and Amir and Ilana Shanks (2019), “Service Robots Rising: How Grinstein (2018), “Temporal Myopia in Sustainable Behavior Humanoid Robots Influence Service Experiences and Elicit under Uncertainty,” International Journal of Research in Compensatory Consumer Responses,” Journal of Marketing Marketing, 35, 378–393. Research, 56(4), 535-56. Wirtz, Jochen, Paul G. Patterson, Werner H. Kunz, Thorsten Mishra, Saurabh, and Sachin B. Modi (2016), “Corporate Social Gruber, Vinh Nhat Lu, Stefanie Paluch, and Antje Martins Responsibility and Shareholder Wealth: The Role of Marketing (2018), “Brave New World: Service Robots in the Frontline,” Capability,” Journal of Marketing, 80(1), 26-46.‏ Journal of Service Management, 29(5), 907–31. Mori, Masahiro (1970), “The Uncanny Valley,” Energy, 7(4), Xu, Alison Jing, and Robert S. Wyer Jr. (2010), “Puffery 33–35. in Advertisements: The Effects of Media context, Nocks, Lisa (2007), The Robot: The Life Story of a Technology, Communication Norms, and Consumer Knowledge,” Journal Greenwood Publishing Group. of Consumer Research, 37(2), 329-343. Porter, Michael E., and Mark R. Kramer (2011), “Creating Shared Yoon, Yeosun, Zeynep Gürhan‐Canli, and Norbert Schwarz Value: How to Reinvent Capitalism – and Unleash a Wave of (2006), “The Effect of Corporate Social Responsibility (CSR) Innovation and Growth,” Harvard Business Review, January- Activities on Companies with Bad Reputations,” Journal of February, 1-17. Consumer Psychology, 16(4), 377-390. Riskin, Jessica (2016), The Restless Clock: A History of the Zhang, Y. Charles, and Norbert Schwarz (2012), “How and Centuries-long Argument Over What Makes Living Things Why 1 Year Differs from 365 Days A Conversational Logic Tick. University of Chicago Press. Analysis of Inferences from the Granularity of Quantitative Romani, Simona, and Silvia Grappi (2014), ‘HowCcompanies’ Expressions,” Journal of Consumer Research, 39(2), 248-259. Good Deeds Encourage Consumers to Adopt Pro-social Zhong, Chen-Bo, and Geoffrey J. Leonardelli (2008), “Cold Behavior,” European Journal of Marketing, 48(5/6), 943-963. and Lonely: Does Social Exclusion Literally Feel Cold?,” Schweitzer, Fiona, Russell Belk, Werner Jordan, and Melanie Psychological Science, 19(9), 838-842. Ortner (2019), “Servant, Friend, or Master? The Relationships Złotowski, Jakub, Kumar Yogeeswaran, and Christoph Bartneck Users Build with Voice Controlled Smart Devices,” Journal of (2017), “Can We Control It? Autonomous Robots Threaten Marketing Management, 35(7/8), 693-715. Human Identity, Uniqueness, Safety, and Resources,” Stein, Jan-Philipp, Benny Liebold, and Peter Ohler (2019), “Stay International Journal of Human-Computer Studies, Back, Clever Thing! Linking Situational Control and Human 100(April), 48–54. Uniqueness Concerns to the Aversion Against Autonomous Technology,” Computers in Human Behavior, 95(June), 73–82. The Promise and Perils of Networks and Crowds: New Perspectives on Collective Consumer Beliefs Need Chairs

Paper #1: The Limits of Collective Intelligence: Group The authors of Paper #3 compare the ability of groups versus Dynamics of Binary Choice and Belief Formation information markets to aggregate information effectively. They show Joshua Becker, Northwestern University, USA that groups outperform markets in the absence of conflicts of interest, Douglas Guilbeault, University of Pennsylvania, USA whereas markets outperform groups when conflicts of interest are Ned Smith, Northwestern University, USA present. Moreover, they show that people do not anticipate these results and generally trust groups more than markets, unless they have Paper #2: The Statistical Self-Fulfilling Selection Fallacy: When acquired experience with information markets. Markets Inadvertently Reinforce False Beliefs The authors of propose a new variant of conjoint anal- Yeonjeong Kim, Massachusetts Institute of Technology, USA Paper #4 ysis: Instead of eliciting consumers’ personal choices and preferenc- Christopher Olivola, Carnegie Mellon University, USA es, it asks them to predict those of their peers. They show that such Paper #3: In Groups (vs . Markets) We Trust: Relying on Groups peer-choice conjoint analyses, which tap into consumers’ knowledge or Information Markets to Aggregate Knowledge of their peers, produce more precise market share estimates. Boris Maciejovsky, University of California Riverside, USA In sum, the papers in this session offer important, novel in- David Budescu, Fordham University, USA sights, from multiple perspectives, regarding the promise, but also Paper #4: Peer-Predicted-Preference Conjoint Analysis: perils, of relying on social information transmission and collective Tapping Into Consumers’ Knowledge of their Peers’ Preferences beliefs. This session will appeal to researchers interested in better to Improve Market Share Estimates understanding how consumers converge in their beliefs (e.g., regard- Sonja Radas, The Institute of Economics, Zagreb, Republic of ing products and brands), the extent to which these collective beliefs Croatia should (not) be trusted, and how best to utilize social beliefs. Drazen Prelec, Massachusetts Institute of Technology, USA The Limits of Collective Intelligence SESSION OVERVIEW Group Dynamics of Binary Choice and Belief Formation Consumer behavior, and marketing more generally, is increasingly driven by collective beliefs regarding the popularity of brands, EXTENDED ABSTRACT the quality of products, etc. The growing availability of product reviews and rapid dissemination of opinions about brands via social Introduction media have created more opportunities than ever before, for both Consumer behavior is driven by factual beliefs. For example, consumers and marketers, to aggregate information about products food choices are driven by both true and false information relating and brands. Whereas recent research has emphasized the benefits of certain foods to health benefits, and other foods to health risks. Relat- trusting aggregate opinions (the ‘wisdom of crowds’), this session edly, consumer’s beliefs about the potential environmental impacts provides a more nuanced perspective on collective beliefs, by also of their behavior can also drive purchasing decisions (Van de Velde showcasing their limitations. The four papers in this session reveal et al. 2009). As a result, understanding the factors that determine be- novel insights into the strengths and weaknesses of collective beliefs, liefs in a population—and particularly belief accuracy—is critical to while addressing important questions: How are erroneous beliefs understanding consumer behavior. The study of belief accuracy is formed and why are they maintained? When are collective beliefs a central area of research on collective intelligence. Although con- more (vs. less) accurate? How can we leverage consumers’ social sumer beliefs are heavily shaped by the beliefs of peers in social networks to improve predictions? networks (Katz and Lazarsfeld 1955), which puts populations at the The first two papers examine how consumer social networks risk of effects such as herding and groupthink (Surowiecki 2004), re- and market selection processes can bias consumer beliefs. search on collective intelligence has also found that social influence The authors of Paper #1 demonstrate an important limitation of processes can reliably improve belief accuracy under a wide range of social belief aggregation (e.g., the revision of beliefs through word- conditions (Becker et al. 2017). However, this research has largely of-mouth information transmission). In particular, they show that, studied beliefs as continuous numeric estimates while individual be- whereas social (e.g., word-of-mouth) information transmission can havioral decisions are often driven by binary choice estimates. We improve the accuracy of beliefs regarding continuous numerical val- investigate whether the potential benefits of collective intelligence, ues (“how long will this product last?”), it often fails to improve i.e. social information processing, extend to binary choice behav- beliefs regarding binary choice decisions (“should I purchase?”). iors. We show both theoretically and empirically that the same exact The authors of Paper #2 identify an important, yet previously processes which improve belief accuracy for numeric estimates can unrecognized, marketplace source of erroneous belief maintenance reduce belief accuracy for binary choices. Remarkably, we find that and propagation: the statistical self-fulfilling selection fallacy. They a group can, in a single conversation, simultaneously improve their show how multi-stage market selection processes (e.g., the decisions numeric estimate accuracy even as their binary choice decision based to first purchase, and later return, a product) that are inadvertently bi- on that belief becomes less accurate. ased by erroneous beliefs can produce evidence that seems to support those beliefs, thereby reinforcing them. Critically, this fallacy occurs Theoretical Results even without motivations to maintain those beliefs. We study two agent-based models of social exchange in which The next two papers consider the relative benefits of relying subjects begin with independent beliefs and then repeatedly revise on groups (vs. markets) and social networks to improve predictions. their beliefs after learning the beliefs of peers in a social network. In the first model (“binary-binary”) subjects form binary-choice es-

Advances in Consumer Research 842 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 843 timates and can observe each other’s binary-choice estimates. Prop- suring consumer beliefs could yield fundamentally different pictures osition 1 shows that under very general assumptions, the majority of apparent opinion change. belief will grow over time—regardless of accuracy. In the second model (“binary-continuous”) subjects form con- The Statistical Self-Fulfilling Selection FallacyWhen tinuous numeric estimates and can observe each other’s continuous Markets Inadvertently Reinforce False Beliefs numeric estimates. Our model is identical to the model used in prior empirical (Becker et al. 2017; 2019) and theoretical (Becker et al. EXTENDED ABSTRACT 2017; Golub and Jackson 2010) research on collective intelligence, People have a surprising capacity to maintain, and even rein- i.e. the model studied by DeGroot (1974). Here, however, agents are force, erroneous beliefs concerning the nature of relationships be- expected to convert their final numeric estimate to a binary choice tween variables (e.g., Gilovich 1991; Gilovich, Griffin, and Kahn- via a threshold model. E.g., suppose an agent only wants to buy a eman 2002). In particular, they often overestimate the ability of product that will last at least 5 years before needing to be repaired certain variables to predict outcomes of interest. For example, con- or replaced; they will first form a numeric estimate (“how long will sumers may overestimate the extent to which brand prestige predicts this product last?”) and then make a binary choice estimate (“will it product quality, or overrate the success of particular social strategies last 5 years?”). Proposition 2 shows that changes in resulting binary (e.g., witness the plethora of widely-shared beliefs concerning pre- choice decisions are decoupled from changes in individual beliefs dictors of dating success), and market researchers may over-ascribe as measured by averages. This decoupling means that even if the certain qualities and behaviors to particular segments of the con- average estimate becomes more accurate, binary choice decisions sumer population. Surprisingly, many erroneous beliefs manage to can still become less accurate—i.e., the number of people making propagate among even well intentioned consumers –i.e., absent any the ‘correct’ choice can decrease. We demonstrate theoretically the longstanding, ego-boosting motives to hold such beliefs. Given the conditions under which this seemingly paradoxical result can occur, robustness, prevalence, and widespread impact of such beliefs, an finding that these conditions are consistent with typical properties important goal for researchers should be to not only understand their of empirical data. A scenario in which a numeric estimate becomes origins, but also to elucidate the processes that allow them to survive, more accurate and a vote becomes less accurate. Suppose people are to spread, and (in some cases) to grow stronger over time. estimating the return on an investment, and will choose to purchase We contribute to this important question by providing a novel if the return is greater than 10%, but the (unknown) true return is ac- explanation for the survival, spread, and strengthening of errone- tually below 10%. Each point represents the final belief, showing the ous beliefs concerning the predictive validity of variables. Specifi- trajectory from the initial belief. Although the mean belief becomes cally, we identify a previously overlooked market mechanism –the more accurate, the number of people with the correct vote (i.e. on the statistical self-fulfilling selection fallacy (or ‘triple-S fallacy’)– that correct side of the threshold) decreases. inadvertently produces evidence that appears to confirm an initially hypothesized relationship between two variables (e.g., prestige and Empirical Support quality), or even, in some cases, to inflate it. Critically, this mecha- Study 1 presents a pre-registered virtual lab experiment in nism is separate from (and occurs independently of) previously rec- which subjects are asked to make a binary-choice estimate (e.g., “do ognized sources of biased belief maintenance, such as rich-get-richer more than 50% of people think technology is improving our lives?”) dynamics (the ‘Matthew effect’), the classic self-fulfilling prophecy, before and after observing each other’s binary choice estimates. We and motivated reasoning (e.g., confirmation biases); indeed, the conducted 45 independent trials (15 trials each for 5 unique ques- triple-S fallacy operates even in the absence of these other biasing tions) in which each trial contained 20 subjects answering and re- pressures. acting to each other in real time, for a total of 900 unique subjects. Using a combination of simulations and analyses of actual data, Study 2 re-analyzes data from similar, previously published experi- we demonstrate that the triple-S fallacy can easily occur in market- ments in which subjects formed and reacted to continuous numeric place with multi-stage selection processes --a fairly common, and estimates. We analyze the change in their responses after subjects broad, set of contexts in which ‘candidates’ (e.g., potential product react to each other, and compare the change in the collective intel- purchases, job applicants, financial stock picks, etc.) are initially se- ligence (average estimate) accuracy with changes in decisions along lected based on indirect and incomplete information, and later either hypothetical over/under thresholds. Both Study 1 and Study 2 sup- ‘retained’ or ‘let-go’ after more direct evidence concerning the out- port our theoretical predictions. come of interest (e.g., quality, performance, value, etc.) is observed. Decisions to purchase and return products, for example, exemplify Discussion a multi-stage selection process: products are first purchased (i.e., se- We investigated the effects of social exchange on belief forma- lected) by consumers based on assumed indicators of their quality tion in order to examine whether prior laboratory results on collec- (e.g., brand prestige, price, etc.), and they are later either kept (i.e., tive intelligence will be robust to the kind of decision-making that is ‘retained’) or returned (i.e., ‘let-go’) by those same consumers, de- more likely to determine consumers’ behavior. Our theoretical and pending on their observed quality over a trial period (e.g., a product’s empirical methods followed the same model and principles as prior warranty period). research, and yet showed that instead of reliably improving belief We show that a multi-stage selection process (e.g., first deciding accuracy, social information is likely to just amplify initial majori- whether to purchase a product, then later deciding whether to return ties—regardless of accuracy. Our results offer two important con- it) that is initially biased by erroneous beliefs (e.g., overestimating tributions to the understanding of consumer decision-making. First, how strongly brand prestige predicts product quality) will produce our findings place an important boundary condition on collective evidence that seems to support those initial beliefs, even when ev- intelligence theory, showing that results for continuous numeric esti- eryone is genuinely motivated to hold accurate beliefs concerning mates generally do not apply to binary choice decisions. Second, our the best predictors of the outcome of interest (e.g., product quality). model and analysis highlights important concerns for how to mea- As a result, these beliefs end-up being reinforced, leading to even sure beliefs in a population, showing how different methods of mea- greater reliance on relatively weak, invalid, or even negative predic- 844 / The Promise and Perils of Networks and Crowds: New Perspectives on Collective Consumer Beliefs tors, which produces further deceptively supportive evidence, and crossed the two types of institutions (groups or markets) and the so on. The triple-S fallacy thus creates a self-reinforcing cycle that presence and strength of explicit manipulation inducing conflicts of helps erroneous beliefs propagate and strengthen over time. interest (no manipulation, weak or strong). We also show that the triple-S fallacy can produce four types All groups were asked to choose the best of three applicants of errors concerning the assumed relationship between variables: (labeled A, B, C) for a managerial position. They were given infor- Depending on the situation, it can lead people to: (i) greatly overes- mation about five characteristics of the three applicants and were timate the strength of a weak relationship (effect size inflation), (ii) told that all characteristics carry equal weight in the evaluation. Can- perceive a relationship where none exists (false positive), (iii) fail to didate B was objectively the best; Candidate A was the second best see a negative relationship (miss), or even (iv) perceive a negative applicant and C was the weakest applicant. However, each partici- relationship as being positive (sign-reversal). Moreover, the eviden- pant only received information about three of the five relevant char- tiary distortion produced by the triple-S fallacy can be substantial: acteristics of each applicant. our simulations show that uncorrelated variables (rtrue = 0) can appear During the interactive stage, group members discussed the strongly correlated (robserved >> 0), while negatively correlated vari- candidates, and market participants traded shares corresponding to ables (rtrue < 0) can appear positively correlated (robserved > 0). Interest- the candidates. In the weak (strong) manipulation condition, one ingly, however, the triple-S fallacy mainly works in one direction: to (all three) participant(s) was (were) promised side payments if she inflate observed correlations. By contrast, it does little to suppress (they) could affect the group vote to go for A and C, respectively. (“deflate”) the observation of an existing relationship that nobody This information was private. After the interaction, participants initially believes in. privately picked their preferred candidate. If the majority of their We also show that the triple-S fallacy is robust to various as- group/market (2 or more participants) voted for the best candidate sumptions regarding the nature of the selection process (e.g., wheth- (B), everyone received £4 plus the show-up fee of £5. Otherwise er selection categories are discrete or continuous), the characteris- they just received the show-up fee. In the manipulation conditions, tics of the variables of interest (e.g., whether they follow normal or the participants who were successful in having Candidate A or C uniform distributions), and the relationships between these variables elected received £5 plus the show-up fee. Afterwards, participants (e.g., whether they are negatively or positively correlated). were asked on a 7-point Likert scale how much trust they had that We conclude by discussing strategies that consumers and mar- the group/market accurately aggregated the available information. keters can adopt to mitigate the distorting influence of the triple-S For each group/market we calculated the proportion of (post fallacy. group or market interaction) votes for the best candidate, B. An ANOVA using the votes for applicant B as the dependent variable In Groups (vs . Markets) We Trust Relying on Groups or and institution (groups, markets) and manipulation (control, weak, Information Markets to Aggregate Knowledge strong) as between-subjects factors showed a significant main effect for manipulation (F(4, 140)=11.25, p<.001) and a significant interac- EXTENDED ABSTRACT tion between institution and manipulation (F(4, 140)=3.56, p=.005). Organizations rely on internal forecasts to guide strategic plan- The interaction suggests that groups outperformed markets in the ning decisions and manage day-to-day operations. Often this process absence of manipulations, whereas markets outperformed groups in of pooling and aggregating information involves the use of teams the presence of manipulations. However, people generally trusted and committees, particularly for complex and difficult tasks (Salas, groups more than markets (F(1, 440)=27.33, p<.001; Mgroup=3.52 vs. Cooke, and Rosen 2008). Recently, many organizations, like Ford, Mmarket=3.00). Google (Cowgill and Zitzewitz 2015), Hewlett-Packard (Gillen, In Study 2 we sought to generalize these findings and to bet- Plott, and Shum 2013), and Nokia (Hankins and Lee 2011), started ter understand people’s higher trust in groups. Two hundred and using information markets as an alternative to traditional teams and twenty-four participants were randomly assigned to one of eight be- committees. tween-subjects conditions with the factors interaction setting (group/ The two institutions – groups and markets – differ in many im- market), conflicts of interest (yes/no), and task (hidden profiles/pre- portant ways. First, the setting is markedly different. Ad-hoc groups diction). In all settings participants watched a video of either a group (but also teams and committees) typically operate in face-to-face or a market interaction and were asked to indicate how much trust settings with unregulated communication patterns, whereas markets they had that they would aggregate information accurately. Specifi- are highly structured and are usually implemented electronically on cally, we asked participants to evaluate groups and markets on eight network computers. Equally important, the participants’ motiva- attributes: transparency, benevolence, efficiency, familiarity, -fair tions can vary: The implicit assumption for group decision-making ness, integrity, predictability, and how natural they perceived them. is that members have aligned incentives, implying shared goals and All items were answered on 7-point Likert scales. cooperation. Markets, on the other hand, are competitive and allow The responses to the 8 attributes were highly correlated, sug- for conflicting goals. Thus, traders may wish to withhold private in- gesting the existence of a general trust factor. Hence, we comput- formation that they believe would benefit them during trading. De- ed a composite measure of trust by averaging the 8 items for each spite the surging interest in groups and markets, surprisingly little is participant and used it as the dependent variable in a 2 (interaction known about the conditions under which they perform well and the setting: group vs. market) x 2 (conflicts of interest: yes vs. no) x 2 circumstances under which one outperforms the other. (task: hidden profiles vs. prediction) between-subjects ANOVA. The Following a systematic replicate-and-extend strategy, we pres- results indicate a significant main effect for interaction setting (F(1, ent the results of three experimental studies that demonstrate that 216)=51.14, p<.001, η2=.19), confirming that people trust groups people trust groups more than markets to pool and aggregate infor- more than markets (Mgroup=4.95, SDgroup=0.83, Mmarket=4.03, SDmar- mation effectively, even when their performance is inferior. In Study ket=1.10). 1, N=450 participants were assigned to groups of three, which, in In Study 3, we recruited 358 volunteers from the Good Judg- turn, were assigned to one of six between-subjects conditions. We ment Open Forecasting project (https://www.gjopen.com/) to in- Advances in Consumer Research (Volume 48) / 845 vestigate how prior experience affects peoples’ trust in groups and predictions result from respondents’ internal (i.e., mental) polling markets. The participants watched videos of group and market inter- of their social circle. A personal choice can thus be interpreted as actions and were asked which institution they would prefer, when as- a “degenerate” internal poll based on a “circle of one.” We follow sessing job candidates. We analyzed the choices of those participants this interpretation and adapt it to the binary choice conjoint model, who previously took part in the Good Judgment Forecasting Tour- while allowing for possible interdependencies among social circle nament as part of singular settings only (as individuals, in teams, members. An important implication of our model is that larger social in markets). Their responses differed significantly across the three circles will lead to smaller errors in the estimation of utilities. groups (χ2(2)=10.83, p=.004). Whereas participants who took part in We conducted four experiments (N = 402) to test our models’ the forecasting tournament as individuals or as part of teams gener- prediction that conjoint analyses based on peer-predicted choices ally preferred committees over markets, those participants who took (vs. own-choices) will more accurately predict preference shares. part in markets expressed the opposite preference. All of our experiments were performed with online panels, each The results of the three studies indicate that people generally with a different set of subjects. In each experiment, we compared trust groups more than markets, even though under conflicts of inter- the traditional own-choice-based conjoint analysis with our new est markets outperform groups. Prior experience with markets can peer-predicted-choice approach. Specifically, we evaluated each reverse this pattern. approach’s out-of-sample choice-share predictions. In Experiments 1-3 we use holdout sets, where part-worths estimated from the own- Peer-Predicted-Preference Conjoint Analysis Tapping choice and peer-choice models are utilized to calculate the predicted Into Consumers’ Knowledge of their Peers’ Preferences overall share of an alterative, and those estimates are then compared to Improve Market Share Estimates with the actual share of choices for that alternative. In Experiment 4, instead of holdout sets, we used a validation sample that consisted of EXTENDED ABSTRACT a completely different sample of participants. Typically, market research—including traditional conjoint anal- In line with the predictions of our model, we found that the ysis—begins by asking potential customers for their personal pref- peer-choice conjoint analysis approach achieved significantly higher erences and intentions, and then projects this sample of responses predictive accuracy than the traditional (own-choice) conjoint analy- to the entire market. In this paper, we propose a different approach sis approach. Specifically, the average error of the peer-choice model to conjoint analysis: rather than asking consumers about their own was significantly lower than that of the own-choice model, both in likely choices, we instead ask them about the choices that (they pre- Experiments 1-3 (pooled-error Wilcoxon signed t-test: p = .0061) dict) others will make, thus utilizing consumers’ knowledge about and Experiment 4 (Wilcoxon signed t-test: p = .0004). The errors their social peers. More specifically, we ask respondents to predict of the peer-choice model (vertical / y-axis) against the errors of the the percentage of other respondents who will select each option in own-choice model (horizontal / x-axis), in Experiments 1-3 and Ex- the choice set. We refer to these judgments as “peer-predictions.” periment 4, respectively. The prediction errors of the own-choice Recent research found that people are quite good at predicting oth- model tend to be larger than those of the peer-choice model for most ers’ future choices on the basis of the latter’s observed past choices goods being evaluated (i.e., most data points are on the own-choice (Velez et al. 2016). side of the line representing equality of errors). Some prior studies have considered the role of others’ prefer- We also show that the accuracy of our peer-choice approach is ences in consumer decision-making. For example, Kim et al. (2014) moderated by the level of choice dependence within respondents’ proposed a “group-sourced mechanism” which mimics real-world social circles. In particular, information increases as social circle contexts where the external influence of one’s peers shapes consum- members become mutually more independent, so that smaller in- ers’ choices. Narayan et al. (2011) also examined peer influence, and dependent circles can impart as much information as larger circles found that being exposed to peers’ choices prompted consumers to in which members’ choices are more dependent (i.e., more interre- update their preferences in a Bayesian manner. Although these papers lated). Consequently, our model performs best in situations where considered the influence of peer choices, they nonetheless still esti- social circles are large and diverse. Thus, future research should mated preferences on the basis of consumers’ own/personal choices. investigate if the accuracy of our peer-choice conjoint analysis ap- One rare exception is a study by Dahan et al. (2011), who examined a proach can be further improved by screening for respondents who market mechanism in which consumers traded concepts on the basis have large and diverse social circles. of their perception of others’ preferences, to evaluate new product Finally, we want to stress that, although we compared peer-only concepts. However, Dahan et al.’s method is distinct from conjoint with own-only choices in these studies, our peer-choice method can, analysis. No prior studies, to our knowledge, have considered using in fact, be combined with a traditional (i.e., own-choice) conjoint predictions about others’ preferences in conjoint analysis. In the cur- analysis approach to produce even more accurate estimates of prod- rent research, we depart from traditional conjoint analysis by replac- uct preference shares. That is, we do not preclude the use of personal ing consumers’ own/personal choices with their perceptions of their choice questions in conjoint analysis, but rather advocate for the ad- peers’ choices, as the inputs into conjoint analysis. dition of peer-choice questions, as doing so will yield more precise We begin by introducing a model that modifies the standard overall product preference share predictions while preserving the conjoint analysis approach: rather than considering a consumer’s power of traditional conjoint analysis. own preferences as inputs, it instead utilizes that consumer’s “peer- predicted choice options” —i.e., the alternatives that the consumer believes will be chosen by the majority of his/her peers. The model predicts that using these peer-predicted choices should improve preference share predictions, relative to predictions based on personal choices. The theoretical foundation for our model extrapolates the reasoning of Rothschild and Wolfers (2013), who assumed that peer- 846 / The Promise and Perils of Networks and Crowds: New Perspectives on Collective Consumer Beliefs REFERENCES In Groups (vs . Markets) We Trust Relying on Groups or Information Markets to Aggregate Knowledge The Limits of Collective Intelligence roup Dynamics of Cowgill, Bo, and Eric Zitzewitz (2015), “Corporate Prediction Binary Choice and Belief Formation Markets: Evidence From Google, Ford, and Firm X,” Review Becker, Joshua, Devon Brackbill, and Damon Centola (2017), of Economic Studies, 82 (4), 1309–41. “Network Dynamics of Social Influence in the Wisdom of Gillen, Benjamin J., Charles R. Plott, and Matthew Shum (2013), Crowds,” Proceedings of the National Academy of Sciences, “Inside Intel: Sales Forecasting Using An Information 114 (26), E5070–6. Aggregation Mechanism,” Working Paper, Pasadena: Becker, Joshua, Ethan Porter, and Damon Centola (2019), “The California Institute of Technology. Wisdom of Partisan Crowds,” Proceedings of the National Hankins, Richard, and Alison Lee (2011), “Crowd Sourcing and Academy of Sciences 116 (22), 10717–22. Prediction Markets,” CHI, 11, 1–5. DeGroot, M. H. (1974), “Reaching a Consensus,” Journal of the Salas, Eduardo, Nancy J. Cooke, and Michael A. Rosen (2008), American Statistical Association, 69 (345), 118–21. “On Teams, Teamwork, and Team Performance: Discoveries Golub, Benjamin, and Matthew O. Jackson (2010), “Naive and Developments,” Human Factors, 50 (3), 540–7. Learning in Social Networks and the Wisdom of Crowds,” American Economic Journal: Microeconomics, 2 (1), 112–49. Peer-Predicted-Preference Conjoint Analysis Tapping Katz, Elihu and Paul F. Lazarsfeld (1955), Personal Influence: The Into Consumers’ Knowledge of their Peers’ Preferences Part Played by People in the Flow of Mass Communications, to Improve Market Share Estimates Glencoe, IL: Free Press. Dahan, Ely, Adlar J. Kim, Andrew W. Lo, Tomaso Poggio, and Surowiecki, James (2004), The Wisdom of Crowds, New York, NY: Nicholas Chan (2011), “Securities Trading of Concepts Random House. (STOC),” Journal of Marketing Research, 48 (June), 497–517. Van de Velde, Liesbeth, Wim Verbeke, Michael Popp, Jeroen Kim, Hye-Jin, Young-Hoon Park, Eric T. Bradlow, and Min Ding Buysse, and Guido Van Huylenbroeck (2009), “Perceived (2014), “PIE: A Holistic Preference Concept and Measurement Importance of Fuel Characteristics and its Match with Model,” Journal of Marketing Research, 51 (3), 335–51. Consumer Beliefs About Biofuels in Belgium,” Energy Policy, Narayan, Vishal, Vithala R. Rao, and Carolyne Saunders (2011), 37 (8), 3183–93. “How Peer Influence Affects Attribute Preferences: A Bayesian Updating Mechanism,” Marketing Science, 30 (2), 368–84. The Statistical Self-Fulfilling Selection FallacyWhen Rothschild, David, and Justin Wolfers (2013), “Forecasting Markets Inadvertently Reinforce False Beliefs Elections: Voter Intentions versus Expectations,” NBER Gilovich, Thomas (1991), How We Know What Isn’t So: The Working Paper, https://users.nber.org/~jwolfers/papers/ Fallibility of Human Reason in Everyday Life, New York, NY: VoterExpectations.pdf Free Press. Vélez, Natalia, Yuan Chang Leong, Chelsey Pan, Jamil Zaki, Gilovich, Thomas, Dale Griffin, and Daniel Kahneman, eds. (2002), and Hyowon Gweon (2016), “Learning and Making Novel Heuristics and Biases: The Psychology of Intuitive Judgment, Predictions About Others’ Preferences,” Proceedings of the Cambridge, UK: Cambridge University Press. 38th Annual Conference of the Cognitive Science Society, 1–6. The Value of What’s Hidden: Understanding the Process and Consequences of Hiding Information About the Self and Consumption Chairs: Soo Kim, Cornell University, USA Peggy J. Liu, University of Pittsburgh, USA Kate E. Min, Cornell University, USA

Paper #1: Hiding and Spending: Secret Consumer Behaviors that consumption of such “pseudo-secret” products increases con- and Pro-Relationship Spending sumers’ WOM and actual purchases compared to non-secret prod- Danielle Brick, University of New Hampshire, USA ucts, due to pseudo-secret products’ capacity to enhance consumers’ Kelley Gullo, Duke University, USA perception that they are the hub of the information wheel amongst Gavan Fitzsimons, Duke University, USA their peers. To sum, across three papers in advanced stages (14 studies), this Paper #2: Reminder Avoidance: Why People Hesitate to Disclose session aims to appeal to both consumers and managers who seek to Their Insecurities to Friends understand the process and the consequences of secrets and hidden Soo Kim, Cornell University, USA nuggets of information. Peggy J. Liu, University of Pittsburgh, USA Kate E. Min, Cornell University, USA Hiding and Spending: Secret Consumer Behaviors and Paper #3: Hidden in Plain Sight: Consumer Response to Pseudo- Pro-Relationship Spending Secrets in Marketing Dafna Goor, Harvard University, USA EXTENDED ABSTRACT Anat Keinan, Boston University, USA You had an indulgent lunch, an extra drink with friends, or Nir Halevy, Stanford University, USA bought a pair of shoes, and didn’t tell your partner about this con- Michael I. Norton, Harvard University, USA sumption—it’s your little secret. Are these types of secrets bad for relationships? Although a fair amount of research suggests that secrets SESSION OVERVIEW within relationships are associated with negative consequences, in- With people, there is always more than what meets the eye. Se- cluding greater guilt (Caughlin et al. 2009; Kelly and Yip 2006; Lane crets and hidden information are simply part of human life. However, and Wegner 1995; Simmel 1950), much of this research has focused perhaps due to the difficulty of studying what people wish to keep on emotionally-laden, traumatic, or taboo secrets and ignored the hidden, research on secrets is scarce and generally treats secrets as role of consumer behavior in relationships. In line with research that a stigmatized, negative concept (Slepian, Chun, and Mason 2017), suggests small deceptions can improve relationship outcomes (e.g., with few exceptions (e.g., Rodas and John 2020). white lies: Argo and Shiv 2011), we suggest that secret consumption The three papers in this session seek to allow the audience— may actually have positive outcomes for relationships. Specifically, consumers and managers alike—to develop a more perceptive un- we examine how feelings of guilt from keeping consumption as a derstanding of what motivates people to hide or reveal information, secret from a romantic partner can drive pro-relationship spending. and of the unexpected means that offer them a sense of privacy. It Guilt, a “moral” emotion, motivates people to compensate for a per- also seeks to further destigmatize the concept of secrets and hidden ceived transgression (e.g., Tangney and Dearing 2003). We suggest information by discussing the positive interpersonal consequences as that one way people might respond to guilt from secret consump- well as the marketing impact of secret or pseudo-secret consumption. tion is by investing more spending toward the relationship, such as The first paper (Brick, Gullo, and Fitzsimons) starts off the through gifts. This spending can, in turn, make the partner more sat- session by investigating whether keeping consumer behaviors a se- isfied. cret from one’s partner can have positive outcomes for the relation- We test these predictions across several studies. First, in Study ship. Specifically, this paper investigates whether guilt from keeping 1, we recruited couples at a farmer’s market to test our premise that consumption behaviors hidden can lead to increased pro-relationship people actually keep consumption secrets from their romantic part- spending. Across three studies, the authors find evidence to sup- ners. We asked both members of couples (119 couples, 238 indi- port this prediction and, using dyadic data from both members of a viduals, Mage=41 years) to describe a recent time when they kept a couple, find that greater pro-relationship spending leads to greater consumption secret from their partner and their partner still does not satisfaction with that spending on the part of the partner. know about it. Indeed, 90% of participants were able to provide an The second paper (Kim, Liu, and Min) follows by identifying example, suggesting that this is a common consumer phenomenon. one reason people may hide certain information from close others Further, the majority of responses were about relatively mundane be- (i.e., friends). Five studies show that, although people normally havior. The next two studies test our predictions about how such con- prefer to self-disclose to friends over strangers, they dampen this sumption secrets within the relationship can affect pro-relationship tendency when disclosing their personal insecurities. The authors spending. demonstrate that people may be less forthcoming with friends for In Study 2, we manipulated secret consumption using a scenar- insecurity-disclosure (vs. other disclosures) due to beliefs that they io to tightly test the effects of secret consumption on pro-relationship will be more likely to encounter painful reminders of disclosed con- spending. Specifically, 272 participants (Mage=36 years) from MTurk tents (i.e., personal insecurities) if disclosed to friends than strangers. were asked to imagine that they found more money in their Ama- The third paper then pivots to examining the implications of zon account than they expected, and they decided to buy something secrets on marketing. Specifically, the third paper (Goor, Keinan, Ha- online. Depending on randomized condition, participants were then levy, and Norton) concludes the session by demonstrating consum- asked to imagine either that they told their partner about the purchase ers’ favorable reactions to “pseudo-secret” products, such as secret (control condition) or decided not to tell their partner (secret con- menus and “speakeasy” venues. Across six studies, the authors find dition). Participants indicated how guilty they would feel (7-point Advances in Consumer Research 847 Volume 48, ©2020 848 / The Value of What’s Hidden: Understanding the Process and Consequences of Hiding Information About the Self and Consumption Likert scale). Next, participants were shown two new subscription Reminder Avoidance: Why People Hesitate to Disclose services: one was for a “date night” subscription that they could use Their Insecurities to Friends with their partner and the other was for a personal “hobby” subscription. They were told that they would receive 15 virtual raffle tickets EXTENDED ABSTRACT to use toward winning a $50 gift card for one of the services. The Many of us try to keep our personal insecurities well-buried. amount of raffle tickets put toward the date night subscription served However, they still burden our minds until the burden becomes too as our dependent variable and measure of pro-relationship spending. large to carry alone, such that we yearn to release them. Indeed, a Results revealed that people who imagined keeping a consumption survey of 100 U.S. residents—conducted as part of this research— secret from their partner put significantly more of their raffle tick- reveals that of the 96 people who reported possessing personal inse- ets toward the date box than did people who imagined telling their curities, more than 70% wished they could get their insecurities off partner (p<.05). People in the secret condition reported significantly their chest and more than 60% wished they could disclose them to greater guilt than did people who imagined telling their partner about another person. their purchase (p< .01), and mediation analyses revealed that guilt But, to whom could people disclose their insecurities and hope- mediated the link between secret consumption and pro-relationship fully unburden themselves of the associated psychological pain? One spending. instinctive audience may be friends. Indeed, people generally turn In Study 3, we used dyadic data to examine whether greater more to friends than strangers to self-disclose (Chaikin and Derlega guilt from real secret consumer behavior is associated with greater 1974; Gaebelein 1976; Jourard and Lasakow 1958) and believe that real-life pro-relationship spending, operationalized as spending on if they were to disclose hardships, they would more easily receive the partner for Valentine’s Day. In this study, we also examine the help from friends than strangers (Deri, Stein, and Bohns 2019). Such downstream consequences of secret consumer behaviors and pro- promised support seemingly points to the possibility of people self- relationship spending from the partner’s perspective. Each member disclosing their insecurities—self-aspects that likely require sup- of a romantic couple was recruited by an online sampling panel as port—to friends more than they would for other types of disclosures. part of a larger paid study on relationships. One hundred forty-one However, Rubin (1975)’s “passing strangers” effect introduces individuals successfully completed this study (Mage=53 years). Par- an interesting alternative perspective. In his work, out-of-towners ticipants were asked to recall and describe a recent time that they self-disclosed more intimately to an unacquainted research assistant engaged in a consumption activity and intentionally did not tell their (RA)—a stranger—than local residents. Rubin (1975) speculated partner about it. They were then asked how guilty they felt for the that out-of-towners may have disclosed to the RA their intimate consumption activity and how guilty they felt for not telling their personal information they would not even share with people they partner about it (9-point Likert scales). We next asked participants normally encounter because they could unburden such information to think about previous years, and indicate how much money, time, without the concern of having to face or think about this RA or the and money spent on gifts they spent on Valentine’s Day this year consequences of disclosed information again. compared to previous years (1=a lot less to 7=a lot more; α=.89). We build upon this perspective and posit that there is something Next, we asked participants how satisfied they were with what they disconcerting about disclosing one’s very private personal informa- did for their partner and how indulgent were the gifts/activities they tion, such as personal insecurities, to those one repeatedly interacts gave to their partner. Finally, to assess partner satisfaction with pro- with (e.g., friends). That is, disclosing personal insecurities to friends relationship spending, we asked participants how satisfied they were increases the risk of being reminded of these already reminder-prone with what their partner did for them for Valentine’s Day. Because self-aspects. Friends may encourage the discloser to elaborate on this is dyadic data, we used multi-level modeling with individuals the disclosed insecurities to understand, help, or speculate togeth- nested within couples (Kenny, Kashy, and Cook 2006). In line with er (Rose 2002) or repeatedly bring them up over the course of the the findings from study 2, greater guilt from a recent secret consump- relationship. All these interactions could escalate the risk of being tion activity was significantly associated with greater Valentine’s painfully reminded of the disclosed contents. Hence, we hypothesize Day spending (p<.05). Greater guilt from a recent secret consump- and demonstrate in five pre-registered studies that, due to the risk of tion activity was also significantly associated with greater indulgent painful reminders of disclosed contents, people dampen their usual gifts/activities (p<.05) and greater satisfaction with how Valentine’s tendency to turn more to friends than strangers for self-disclosure Day went this year (p<.05). Using the actor-partner interdependence when disclosing personal insecurities (vs. other information). model (APIM; Kenny et al. 2006), we examined the downstream ef- Study 1 provided initial evidence for our effect in a 2-group fects of secret consumer behaviors on the partner. As actors reported (self-disclosure: insecurity, neutral) between-subjects design. 100 spending more on Valentine’s Day this year relative to other years, friend-pairs (N=200) were recruited. Participants in the insecurity which was significantly related to guilt from keeping their consump- condition wrote about self-aspects that make them feel insecure. tion activity a secret, partners reported significantly greater satisfac- Those in the neutral condition wrote about their daily routine. They tion with how Valentine’s Day went this year (p<.01), suggesting were then told that, as part of this “social interaction” study, they additional positive downstream consequences of secret consumption would need to read what they have written to another person and within relationships. that they could choose between “another participant who is unac- Across three studies, we examine secret consumer behavior in quainted with you” and “the friend you came with” as their disclo- relationships. We find that although keeping a consumption activity sure-audience. As hypothesized, participants disclosing their insecu- from one’s partner can lead to greater guilt, this guilt can have posi- rity (vs. neutral personal information) were significantly less likely tive outcomes for the relationship, specifically in terms of greater to choose their own friend as their audience (insecurity:76.2% vs. pro-relationship spending. Further, using data from couples report- neutral:91.6%), B=-1.22, p=.005. ing on real behavior, we found that greater pro-relationship spending Study 2 (N=608) demonstrated that this effect does not occur was associated with greater satisfaction on the part of the partner, for any negatively-valenced self-disclosure using a 2-group (self- highlighting additional positive downstream consequences of secret disclosure: personal insecurity, societal insecurity) between-subjects consumer behaviors in relationships. design. The personal insecurity condition was the same as in study Advances in Consumer Research (Volume 48) / 849

1. Participants in the societal insecurity condition wrote about some- pitality, food and beverages, fashion and entertainment. Restaurants thing they read in the news that makes them feel insecure about so- ranging from gourmet Michelin-starred to mainstream fast-food ciety (negative but non-personal disclosure). Participants then indi- chains offer secret-menu items, and hidden shops and “speakeasy” cated their preferred disclosure-audience (1=stranger to 7=friend). bars have camouflaged entrances and secret passcodes. Paradoxi- As expected, participants disclosing personal (vs. societal) insecuri- cally, many of these hidden secret items and places have become fa- ties showed significantly lessened preference for a friend as their mous for being secrets, and interestingly, the majority of secret bars audience (Mpersonal=4.92 vs. Msocietal=5.60), F(1, 606)=18.38, p<.001. and shops are located in central public areas - hiding in plain sight. Study 3 (N=202) conceptually replicated study 1 and provided We label these hidden but publicly available products as “pseu- initial support that the dampened friend-preference for insecurity- do-secrets” and demonstrate their positive impact on word-of-mouth disclosure occurs among those more prone to rumination (and thus and actual purchase behavior in the marketplace. We demonstrate more likely to wish to avoid disclosed-content reminders), using a that this effect is mediated by feelings of social centrality, i.e., the 2 (self-disclosure: insecurity, neutral) × rumination-tendency (mea- subjective feeling of being connected to others in the network and sured) between-subjects design. Participants wrote about either their feeling like information hubs, and moderated by how the informa- insecurities or neutral personal information (recent grocery-trip), tion was obtained. Additionally, we explore an important individual completed the Rumination Tendency Scale (Brinker and Dozois characteristic that could moderate consumers’ propensities to dis- 2009), and indicated their disclosure-audience (1=stranger to seminate information in the marketplace: Market Mavens (Feick and 7=friend). Participants disclosing their insecurity (vs. neutral per- Price 1987). sonal information) were again less likely to prefer a friend as their Six field and lab studies test the pseudo-secrecy phenomenon. audience (Minsecurity=4.53 vs. Mneutral=5.10), B = -.57, p = .06. Fur- Exploratory interviews (N=33) suggested that pseudo-secrets thermore, this effect surfaced for those highly prone to rumination: exist in various product categories. Respondents mentioned ex- disclosure topic × rumination-tendency interaction, B=-.70, SE=.28, amples of hidden shops that were nestled in central shopping areas t=-2.51, p=.013 (Johnson-Neyman point=4.87). and secret-menus that were offered by big fast-food chains. They Study 4 (N=404) provided further support for our account that described recurring cues that create a sense of hidden-yet-inviting anticipated risk of painful reminders of disclosed-content decreases products and experiences, such as camouflaged entrances of shops one’s preference for friends over strangers as their disclosure-au- which also provided their actual address and products online. Impor- dience via mediation. We used a 2-group (non-disclosure reason: tantly, respondents noted that knowing about pseudo-secrets made insecurity-trigger, no-opportunity) between-subjects design and them feel “in the know” and like insiders, portraying a phenomenon measured participants’ anticipated pain from reminders. In the inse- that is first and foremost a social enterprise. curity-trigger condition, participants described a personal challenge In a pilot study (N=177), we surveyed shoppers at the rapper they have not disclosed because it triggers insecurities; those in the Kendrick Lamar’s pop-up shop, located in a central area. Although no-opportunity condition described a personal challenge they have the shop did not have a sign or any indication as to what merchandize not disclosed because they did not yet have the opportunity to do so. it was offering, its location, opening dates, and merchandise were Then participants indicated their disclosure-audience (1=stranger to available online and on social media. Perceiving the pop-up shop as 7=friend) and indicated how painful reminders of their stories would more secretive increased consumers’ intentions to buy branded items be (2-item; r=.93). Consistent with our theory, participants pre- (b=.169, t=2.943, p=.004, 95% CI=[.056,.283]). ferred friends less for personal challenge-stories undisclosed due to In Study 1, visitors at a university campus tour (N=2,415) re- insecurity-triggers (vs. no-opportunity), (Minsecurity-trigger=4.24 vs. Mno- ceived a 15% discount coupon for the university store. Since the opportunity=5.14), F(1, 402)=15.94, p<.001. This effect was mediated by store has a small sign at the entrance, we were able to describe it participants’ anticipated pain from reminders, B =-.18, SE=.07, 95% either as a pseudo-secret (“easy to miss”) or as visible and conspicu- CI=[-.34, -.05]. ous (“hard to miss”). Overall, 4.7% of the visitors redeemed tour Study 5 (N=606) provided further evidence that risk of pain- coupons. As predicted, framing the store as pseudo-secret increased ful disclosed-content reminders drives our effect. We used a 2-group actual purchases (redemption rate 6.3%) compared to framing it as (self-disclosure: insecurity, neutral) between-subjects design. The well-known (4.2%; χ2=5.12, p=.024) or control condition (4.1%; self-disclosure procedure was similar to study 3’s. However, the χ2=5.93, p=.015), indicating that pseudo-secrets increased brand in- choice for disclosure-audience was between two equally-support- terest. ive friends differing in one aspect: Friend A who tends to bring up In Study 2, different visitors (N=764) at the same university previously-shared stories and Friend B who does not. Participants tour were asked to help improve future tours by expressing their disclosing insecurities (vs. neutral personal information) were more interest in expanding the tour to focus on the university’s library. likely to choose Friend B, who tends not to bring up previously- Describing the library’s back door as a secret small door (vs. small shared stories (insecurity:78.4% vs. neutral:65.7%), B=.65, p<.001. door) increased visitors’ interest in taking its photo to show their Altogether, our results contribute to the broad consumer psy- friends (F(1,762)=24.22, p<.001). The manipulation did not af- chology literature on how consumers manage secrets by showing fect intentions to show photos of the iconic main door to friends that, when disclosing personal secrets (e.g., insecurities), they turn (F(1,762)=.12, p=.724), indicating that pseudo-secrets (vs. non-se- away from friends because, ironically, even their friendly helping crets) increase WOM intentions about products associated with that gestures could remind them of the painful disclosed-content. information. Studies 3 and 4 examined the mediating role of feelings of cen- Hidden in Plain Sight: Consumer Response to Pseudo- trality. Participants used a social network heat-map to indicate how Secrets in Marketing central they felt. Study 3 had four conditions. 324 adults (67.6% women;

EXTENDED ABSTRACT Mage=35.5) read about a pseudo-secret (vs. not-secret) bar. The bar The marketing and consumption of secret products and experi- entrance was through a back door of (vs. next to) an ice-cream shop. ences has become an emerging cultural phenomenon in retail, hos- In two additional pseudo-secret conditions manipulating the source 850 / The Value of What’s Hidden: Understanding the Process and Consequences of Hiding Information About the Self and Consumption of information, participants imagined that they heard about the bar Caughlin, John P., Allison M. Scott, Laura E. Miller, and Veronica directly from a friend (vs. overheard their friends talking). Pseudo- Hefner (2009), “Putative Secrets: When Information Is secrets increased both interest in going to the bar (p’s<.001) and Supposedly a Secret,” Journal of Social and Personal WOM likelihood (p’s<.003) compared to the non-secrets. Learning Relationships, 26 (5), 713-43. about the pseudo-secret bar directly from a friend further increased Chaikin, Alan L. and Valerian J. Derlega (1974), “Variables purchase intentions (p’s<.044) and WOM intentions (compared to Affecting the Appropriateness of Self-Disclosure,”Journal of non-secret p=.044). Interestingly, learning about the bar indirectly Consulting and Clinical Psychology, 42 (4), 588-93. (overhearing) increased purchase interest (compared to non-secret, Deri, Sebastian, Daniel H. Stein, and Vanessa K. Bohns (2019), p<.001) but decreased WOM intentions (compared to pseudo-secret “With a Little Help from My Friends (and Strangers): control, p=.042), highlighting the important role of feelings of cen- Closeness as a Moderator of the Underestimation-of- trality. Mediation analyses (PROCESS, Model 4) showed significant Compliance Effect,”Journal of Experimental Social effects on purchase interest and WOM via feelings of centrality. Psychology, 82, 6-15. Study 4 had a 2 (pseudo-secret vs. non-secret) × 2 (market- Feick, Lawrence F. and Linda L. Price (1987), “The Market maven) between-subjects design. 439 participants (44.4% female; Maven: A Diffuser of Marketplace Information,”Journal of

Mage=36.6) imagined that a restaurant near them serves an unlimited Marketing, 51 (1), 83-97. number of a secret-menu house burger (vs. house burger) that does Gaebelein, Jacquelyn W. (1976), “Self-Disclosure among Friends, not appear (vs. appears) on the restaurant’s menu. The results re- Acquaintances, and Strangers,” Psychological Reports, 38 (3), vealed two positive effects of pseudo-secrecy and market mavens on 967-70. both purchase interest and WOM intentions (p’s<.001). The brand Jourard, Sidney M. and Paul Lasakow (1958), “Some Factors in information × market maven interaction did not affect purchase in- Self-Disclosure,” Journal of Abnormal and Social Psychology, terest (p=.473). However, a negative effect on WOM (p=.001) indi- 56 (1), 91-98. cated that pseudo-secrets have a stronger effect among non-mavens Kelly, Anita E. and Jonathan J. Yip (2006), “Is Keeping a Secret who scored 5.64 or lower on the 1-to-7 market maven scale, increas- or Being a Secretive Person Linked to Psychological ing their likelihood to spread WOM. A moderated mediation analysis Symptoms?,” Journal of Personality, 74 (5), 1349-70. revealed a significant moderating effect of market mavens on WOM Kenny, David A., Deborah A. Kashy, and William L. Cook (2006), through feelings of centrality (PROCESS, Model 7; a×b=-.0509, “The Analysis of Dyadic Data,” New York: Guilford. SE=.0257, 95% CI=[-.1030,-.0021]). Lane, Julie D. and Daniel M. Wegner (1995), “The Cognitive In addition to the practical implications of understanding how Consequences of Secrecy,” Journal of Personality and Social to design pseudo-secret brand information to generate brand engage- Psychology, 69 (2), 237-53. ment, our research makes important theoretical contributions. First, Legate, Nicole, Cody R. DeHaan, Netta Weinstein, and Richard M it advances research on secret (i.e., private) consumption (Thomas Ryan (2013), “Hurting You Hurts Me Too: The Psychological and Jewell 2019) by demonstrating an opposite phenomenon – prod- Costs of Complying with Ostracism,” Psychological Science, ucts that are designed to offer a secretive experience, but lend con- 24 (4), 583-88. sumers with a gift (i.e., seemingly exclusive knowledge) that they Rodas, Maria A. and Deborah Roedder John (2020), “The Secrecy can pass on to others rather than hide it. It also contributes to re- Effect: Secret Consumption Increases Women’s Product search on social networks (Watts and Dodds 2007) by demonstrat- Evaluations and Choice,” Journal of Consumer Research, ing that consumers’ subjective experience of feeling socially central, forthcoming. rather than their objective centrality, may drive information flow in Rose, Amanda J. (2002), “Co–Rumination in the Friendships of networks. Thus, we suggest that beyond their informational value, Girls and Boys,” Child Development, 73 (6), 1830-43. pseudo-secrets carry an important symbolic social value which can Rubin, Zick (1975), “Disclosing Oneself to a Stranger: Reciprocity resolve an inherent tension between the desire for exclusive products and Its Limits,” Journal of Experimental Social Psychology, and affiliations and the discomfort associated with excluding others 11 (3), 233-60. (Bellezza and Keinan 2014; Legate et al. 2013). Finally, our work Simmel, Georg (1950), “The Secret and the Secret Society,” in The extends research on WOM behavior (Berger 2014) by highlighting Sociology of Georg Simmel, ed. Kurt H. Wolff, New York: the influence of the source of the information on WOM. Free Press of Glencoe, 307-76. Slepian, Michael L., Jinseok S. Chun, and Malia F. Mason (2017), REFERENCES “The Experience of Secrecy,” Journal of Personality and Argo, Jennifer J. and Baba Shiv (2011), “Are White Lies as Social Psychology, 113 (1), 1-33. Innocuous as We Think?,” Journal of Consumer Research, 38 Tangney, June Price and Ronda L. Dearing (2003), Shame and (6), 1093-102. Guilt, New York: Guilford Press. Bellezza, Silvia and Anat Keinan (2014), “Brand Tourists: How Thomas, Veronica L. and Robert D. Jewell (2019), “I Can’t Get Non–Core Users Enhance the Brand Image by Eliciting You out of My Head: The Influence of Secrecy on Consumers’ Pride,” Journal of Consumer Research, 41 (2), 397-417. Self‐Brand Connections,” Journal of Consumer Psychology, Berger, Jonah (2014), “Word of Mouth and Interpersonal 29 (3), 463-71. Communication: A Review and Directions for Future Watts, Duncan J. and Peter Sheridan Dodds (2007), “Influentials, Research,” Journal of Consumer Psychology, 24 (4), 586-607. Networks, and Public Opinion Formation,” Journal of Brinker, Jay K. and David J. A. Dozois (2009), “Ruminative Consumer Research, 34 (4), 441-58. Thought Style and Depressed Mood,” Journal of Clinical Psychology, 65 (1), 1-19. Novel Insights for Enduring Concerns: The Wisdom of Lay Theories Chairs: Gita V. Johar, Columbia University, USA Shilpa Madan, Virginia Tech, USA

Paper #1: Intergenerational Effects of the Unhealthy = Taste for policy makers to increase support for environmental taxes on Intuition on BMI: An Exploration of Food Practices and both sides of the partisan divide. Outcomes The next two papers take the firms’ perspective to ask how lay Barbara Briers, Vlerick Business School, Belgium theories shape consumer choice in emerging domains such as au- Young Eun Huh, Korea Advanced Institute of Science and tonomous products and brand extension failures. De Bellis, Johar, Technology, South Korea and Schweitzer find that one reason behind consumers’ lukewarm Elaine Chan, Nanyang Technological University, Singapore acceptance of highly efficient and convenient autonomous prod- Anirban Mukhopadhyay, Hong Kong University of Science ucts is their lay theory that only effort and hard work yield desired and Technology, China outcomes. To drive consumer adoption, they recommend that firms should emphasize the idea that autonomous products can free up time Paper #2: It Isn’t Easy Being Green: Lay Theories of Wealth for more meaningful pursuits. Applying the lens of lay theories to and Happiness Shape Support for the Environment study the pervasive issue of brand extension failures, Jain et al. un- Shilpa Madan, Virginia Tech, USA cover that consumers’ lay theories about malleability and fixedness Krishna Savani, Nanyang Technological University, Singapore of personality predict the extent to which consumers punish the par- Steve Su, VilleFranche Consulting, France ent brand when an extension failure occurs. Building on entitativity Michael Morris, Columbia University, USA literature, they suggest that breaking the perception of the brand as a Paper #3: No Pain, No Gain: Protestant Work Ethic and the homogenous group can mitigate the risk of extension failure for the Adoption of Autonomous Products parent brand. Emanuel de Bellis, University of St. Gallen, Switzerland This special session integrates several areas of interest high- Gita V. Johar, Columbia University, USA lighted in the call-for-papers (e.g., technology, identity, obesity) un- Nicola Schweitzer, UBS, Switzerland der the umbrella of lay theories to highlight the meaningful role that Paper #4: Brand Extension Failure and Parent Brand Penalty: consumers’ lay theories play in determining substantive consumer The Role of Implicit Theories behavior outcomes. We hope to encourage an engaging conversation Shailendra Pratap Jain, University of Washington, USA on the implications of consumer characteristics such as lay theories Pragya Mathur, City University of New York, USA in influencing consumer behavior and to spark ideas for future re- Matthew Isaac, Seattle University, USA search. Huifang Mao, Iowa State University, USA Durairaj Maheswaran, New York University, USA Intergenerational Effects of the Unhealthy = Taste Intuition on BMI: An Exploration of Food Practices and SESSION OVERVIEW Outcomes Consumers are increasingly grappling with challenging consumption-related issues that impact their health (e.g., processed and EXTENDED ABSTRACT sugar laden foods), their environment (e.g., plastic bags), their un- Childhood obesity is a critical health concern worldwide (WHO derstanding of physical tasks (e.g., autonomous products), and their 2018), and sets the stage for a lifelong struggle with weight and eat- choices in the marketplace (e.g., too many brand extensions). Build- ing. The influence of parents is considered as primary, given that ing on extant literature that finds lay theories as a crucial determinant parents during early childhood act as gatekeepers and role models of real world consumer outcomes (such as obesity, McFerran and around food (McCaffree 2003). Previous research has looked at the Mukhopadhyay 2013), the four papers in this session examine lay influence of parents’ demographics, parents’ nutritional knowledge, theories that shape responses to consequential issues that not only general parenting style, and food parenting practices. This research affect individuals, society, and policymakers, but also have clear investigates how parents’ beliefs about the trade-off between health- implications for brand and marketing strategy. Bringing alive the fulness and tastiness in food (i.e., the Unhealthy = Tasty Intuition, conference theme of “Rendezvous,” this session brings together con- “UTI”; Raghunathan, Naylor, and Hoyer 2005) affects their chil- sumer researchers from three continents to showcase the wisdom of dren’s BMI. consumers’ naive theories in providing novel insights to enduring Taste is generally the most important attribute in food choice and significant issues including obesity, environmental degradation, (Glanz et al. 1998). When parents hold a strong belief that healthy autonomous products, and brand extension failures. food is not tasty, they may consider that their children need an extra The first two papers focus on how their lay theories can nudge incentive to eat food that is healthy but less tasty. Hence, higher UTI consumers to make better (or worse) choices for their well-being. parents may be more inclined to offer their children an extrinsic re- Briers et al. present a unifying framework to underscore the role ward in exchange for eating healthfully. However, this strategy might that parents’ unhealthy = tasty lay belief plays in determining their backfire, as providing external rewards signals that the healthy food children’s BMI. They identify the use of extrinsic rewards as the consumption is a negative experience that needs to be compensated psychological process that drives this relationship, while ruling out (Birch et al. 1982). As a result, the use of external rewards might alternative explanations. Madan et al. uncover a novel antecedent ironically reduce healthy food consumption thereby increasing BMI. to reduce consumers’ resistance to environmental taxes — their lay Thus, we predict that parents’ UTI positively affects their children’s belief about the relationship between wealth and happiness. Drawing BMI, which is mediated by the extent to which parents use an ex- from hedonic adaptation theory, they identify expectations of adjust- trinsic reward to induce healthy eating and the amount of healthy ment as the underlying mechanism and propose recommendations

Advances in Consumer Research 851 Volume 48, ©2020 852 / Novel Insights for Enduring Concerns: The Wisdom of Lay Theories food consumed by their children. We tested this proposition in four or threaten with a reward (UTI= 3.48 versus non-UTI= 3.19; F(1, studies. 205) = 4.61; p = .03) than respondents in the non-UTI condition. For In study 1, 980 MTurk workers who were the main caretaker of the other five strategies, we found no differences. a child (age 6-11) participated in the survey. Respondents indicated Overall, this research contributes to the literature on lay beliefs the child’s age, gender, height, weight, their child’s fruits and vegeta- (McFerran and Mukhupadhyay 2013) and consumer socialization bles consumption and their likelihood to offer their child an external (Ward 1974; John 1999; Moore, Wilkie and Desrochers 2017). This reward for eating healthy. Next, we measured respondents’ UTI with integration allows us to present a unifying framework of the flow of three items (e.g., “Things that are good for me rarely taste good”; influence, parents’ beliefs parents’ actions  children’s behaviors Raghunathan et al. 2005), their own height and weight. Controlling  children’s outcomes, thereby sparking suggestions for potential for parents’ BMI, children’s age and gender, linear regression analy- interventions and future research. sis showed that the more parents believed in UTI, the higher their children’s BMI was (β = .86, SE = .16, t = 5.26, p < .001). Further, It Isn’t Easy Being Green: Lay Theories of Wealth and bootstrap analyses (Hayes 2013, Model 6) showed that the relation- Happiness Shape Support for the Environment ship between the parent’s UTI and the child’s BMI was serially mediated by the use of an extrinsic reward to induce healthy eating and EXTENDED ABSTRACT A strong majority of Americans believe that the government is the amount of fruits and vegetables consumed by the child (a1 x d21 not doing enough to protect the environment but many of them are x b2 indirect effect = .010, bootstrap se = .0053, bias-corrected bootstrap CI [0.0019; 0.0225]). unwilling to support environmentally protective policies (Newport Study 2 aimed to replicate the results of study 1, while control- 2018). Many consumers may believe single use plastic is ruining ling for two other important parental influences on children’s BMI: marine life but not support a plastic bag ban (Raskin et al. 2020). For parent’s education and how often they exercise. Using a similar sur- example, carbon emissions from transportation are a major source vey as in study 1 (N = 938), results reassuringly showed a positive of greenhouse gases that contribute to climate change. One way to effect of parents’ UTI on their child’s BMI (β = .25, SE = .13, t = reduce them is to impose substantial tax on gasoline, which would 1.99, p < .05), serially mediated by the use of external rewards, and reduce gasoline consumption. A key reason for this gap is that many people endorse environmentalist attitudes because it is socially de- the amount of healthy food consumed by the child (a1 x d21 x b2 indirect effect = .0088, bootstrap se = .0056, bias-corrected bootstrap CI sirable to do so, but a large fraction do not follow through with con- [0.0002; 0.022]). sistent actions because it is costly to do so. In study 3, we aimed to find causal evidence for the link between What are the psychological levers for nudging consumers to consumers’ belief in the UTI and the likelihood to use external re- support such environmental consumption taxes (in short ecotaxes)? wards for pushing healthier eating among children. We manipulated Certainly, the personal and societal costs due to the ecotax are rel- UTI by means of an article describing scientific research on healthy evant. People’s views about the significance of these costs might food being tasty (non-UTI) versus less tasty (UTI). A manipulation hinge on their theory about how happiness adapts to changes in check with the same three UTI measures as before showed that it was wealth, which in this case refers to a reduction in discretionary in- successful (α = .89 ; UTI = 2.82 versus non-UTI = 2.44, F(1, 239) come as a result of the ecotax. We propose that individuals differ = 6.95; p = .009). As dependent measure, we presented respondents sharply in their assumptions about how people’s happiness changes with a babysitting scenario in which the 6-year old child refused to with wealth — some assuming that happiness is dependent on wealth eat broccoli. An open-ended question asked the respondents what (“the more you have, the happier you are”) and others assuming that they would do to make the child eat the broccoli. Answers were cod- it largely adapts (“after a certain threshold, changes in wealth do not ed as providing an external reward (1) versus not (0). Results showed bring similar changes in happiness”). no main effect of UTI on the likelihood to choose a rewarding strat- Popular culture and scientific research provide plenty of evi- egy. Further exploratory analyses revealed that among respondents dence for both these points of view. Many believe that wealth brings who had children of their own (i.e., parents), their rewarding strategy well-being, and if not constant joy then at least a higher baseline. As was not affected by UTI (UTI= 43.6% versus non-UTI= 51.6%), pre- Francoise Sagan (Hartley 2015) put it, “Money may not buy happi- sumably because they were likely to use their pre-existing strategies ness, but I’d rather cry in a Jaguar than on a bus.” In a survey, 73% to motivate their children (as shown in studies 1 and 2). Interestingly, Americans forecasted that they would be happier if they had more among those respondents with no children, for whom the babysitting money (Carroll 2006). However, the belief that happiness adapts to scenario presumably was more realistic, external rewards were more changes in wealth is also widely held. Just as pleasant smells become likely to be mentioned in the UTI condition (42.2%) compared to the less pleasurable after continued exposure, people get used to a higher non-UTI condition (21.7%; χ2 = 5.97; p = .02; overall interaction: income—the higher wealth becomes the new normal, not a reason β = 1.29, p = .02). for pleasure (Frederick and Loewenstein 1999). We propose that this Inspired by study 3, study 4 only included respondents who had lay theory is an important antecedent of differences in support for no children. This study used the same manipulation (α = .87; UTI= ecotaxes. The more people believe that happiness adapts to wealth, 2.88 versus non-UTI= 2.36; F(1, 205) = 11.06; p = .001) and baby- the more likely they are to support environmental consumption taxes sitting scenario. As dependent measure, we asked respondents how because they consider ways in which individuals and organizations likely they were to offer the child an exchange in return for eating the would adjust their behavior over time in response to the tax. broccoli. Next, we asked them to rate the likelihood of nine different Eight studies provide converging support for these hypotheses. strategies, of which four were related to providing an external reward We first developed an 11-item scale to measure people’s lay theories (α = .72): reward and threat with food, reward and threat with non- about wealth and happiness. Several studies established the predic- food, praise, scold, distraction, disguise, and making it fun. In line tive validity, convergent, discriminant, and test-retest reliability of with our hypothesis, respondents in the UTI condition were margin- the scale. ally more likely to offer an exchange (UTI= 4.81 versus non-UTI= Study 1a tested if people’s lay beliefs about wealth and happi- 4.28; F(1, 205) = 3.60; p = .06) and significantly more likely to offer ness influenced their support for close to 100% increase in gasoline Advances in Consumer Research (Volume 48) / 853 tax. Four hundred five American citizens on MTurk completed the the completion of cumbersome tasks can be associated with satisfac- study. In line with the hypothesis, people who believed that hap- tion and positive feelings. Do consumers account for these feelings piness adapts to changes in wealth were more likely to support the and if so, do they impact the adoption of these new technologies? policy (β = .22, t(404) = 2.048, p = .041) Consumer research has demonstrated that consumers experi- Studies 1b-1d (N = 198, 199, 205) replicated the findings in ence autonomous products in ambivalent ways. Autonomous prod- Study 1a with a wider variety of consumption taxes such as plastic ucts are perceived as increasingly beneficial but risky (Rijsdijk and bag tax, a disposable cup charge, and a processed foods tax. We also Hultink 2003; 2009). A recent focus has been on the role of product controlled for other alternative predictors of people’s support for the autonomy, which has been associated with consumer disempower- environment and/or taxes such as environmental concern, political ment and reduced feelings of control (André et al. 2018; Schmitt orientation, personal cost incurred due to the tax, materialism, and 2019; Wertenbroch et al. 2019). Autonomous products can also chal- income. Consumers’ lay beliefs about wealth and happiness predict- lenge identity-based consumption, as they limit consumers’ ability ed support for ecotaxes even when controlling for these constructs. to attribute the outcomes of consumption to themselves (Leung, Studied 2a and 2b provided causal evidence for our hypothesis. Paolacci, and Puntoni 2018). Finally, completing a cumbersome task Participants (N = 326, MTurk) either read an article advocating that (e.g., a chore such as cleaning that is less central to one’s identity) is happiness adapts to changes in wealth or that happiness changes with positively associated with meaning in life (Baumeister et al. 2013). changes in wealth. They then indicated their willingness to support According to our theorizing, the lay theory that effort and hard a multiple consumption taxes on gasoline, plastic, disposable cups, work yield desired outcomes—reflected in everyday idioms such as and processed foods. Consistent with our hypothesis, participants “no pain, no gain” and “nothing ventured, nothing gained”—holds who read the adaptation article were more supportive of the con- an important key to understanding autonomous product adoption. sumption tax (MWealthAdapt = 4.37, MWealthDependence = 4.00, t(324) = -1.99, This notion is ingrained in many cultures and denoted by the prot- p = .048). Further, the indirect effect through the perception that the estant work ethic (PWE; Cheng, Mukhopadhyay, and Schrift 2017). society will adjust its consumption habits in response to the tax was We posit that consumers high in PWE should be less likely to adopt significant [B =.32, SE =.12, 95%CI [.09,.56]). autonomous products, as these products go against their core belief Study 3 provided additional evidence for adjustment as the that hard work is needed to be successful. Given that the PWE is underlying mechanism through mediation by moderation. Specifi- linked to feelings of meaningfulness (Pogson et al. 2003; Schnell cally, participants were told that funds generated from the tax would and Becker 2006), we expect that consumers high in PWE derive either be used to help the society adjust to a higher tax regime or more meaning from manual tasks, which reduces their willingness to not (e.g., an air ticket tax being used to develop the train network in delegate manual tasks to autonomous products. the state versus being used to upgrade vending machines). We found A series of four studies, conducted in the field, online, and in that wealth-adaptation theorists supported the tax when the proceeds the lab, shows that high (vs. low) PWE is related to lower adop- were used to facilitate societal adjustment (B = .71, SE = .29, t(464) tion of autonomous products. We provide evidence for this effect by = 2.41, p = .016), but not when they were not used to support societal measuring and manipulating PWE across four autonomous product adjustment (B = -.35, SE = .24, t(470) = -1.44, p = .15). domains. Based on the idea that consumers’ meaning from manual Finally, Study 4 sought to provide behavioral evidence through tasks drives the effect, we derive an intervention for firms to reverse an incentive compatible design. We first measured participants’ lay the effect. Specifically, consumers high in PWE tend to choose au- beliefs about wealth and happiness. After completing several unre- tonomous over manual products when a meaningful task substitute is lated tasks, participants were told that 1 out of every 100 participants provided by highlighting that the use of autonomous products frees would be getting a $10 bonus. They were then given the opportunity up time that can be used for more meaningful activities. to divide this bonus between themselves and a lobbying organization Study 1 explores the relationship between PWE and autono- working to get ecotaxes mentioned above implemented. We found mous product adoption at the country level. We collected aggregated that a stronger belief in wealth adaptation theory was associated with field data on consumers’ readiness for automation (Economist Intel- greater amounts pledged for the lobbying organization (B = .71, 95% ligence Unit 2018) and their acceptance of autonomous cars (KPMG CI [.29, 1.13], SE = .19, t(262) = 3.35, p = .001). This was signifi- 2018), which we correlated with consumers’ PWE (Cheng et al. cant even after controlling for age, gender, education, income, social 2017) across major markets. Results showed that PWE is negatively class, and political orientation. associated with both readiness for automation (b = −14.53, t(19) = In sum, the present research identifies a novel construct—peo- 3.18, p = .005) and acceptance of autonomous cars (b = −5.49, t(16) ple’s lay theories about wealth and happiness—and shows that it in- = 3.18, p = .006), even when controlling for GDP per capita and fluences consumers’ support for environmental taxes. We contribute number of motor vehicles. to the science on lay theories and the antecedents of environmentally Study 2 examines the relationship between PWE and autono- conscious behavior. More importantly, we contribute to the limited mous product adoption at the individual level. We conducted this research in consumer behavior that focuses on curbing consumption. study with 381 consumers from MTurk who participated in ostensibly two unrelated studies. In the first part, we measured consumers’ No Pain, No Gain: Protestant Work Ethic and the PWE (Schrift et al. 2016) along with other measures. In the second Adoption of Autonomous Products part, consumers chose an autonomous versus manual product in either of three domains (cooking, cleaning, or lawn mowing). Control- EXTENDED ABSTRACT ling for desire for control, meaning in life, and domain expertise, Autonomous products have entered our everyday lives. From we found a negative relationship between PWE and the choice of an cleaning the floor to mowing the lawn, consumers are increasingly autonomous product (b = −.30, z = 2.07, p = .04), with the effect be- able to delegate manual tasks to this novel type of technology. Au- ing mediated by consumers’ meaning from manual tasks. tonomous products allow consumers to free themselves from mun- Study 3 investigates the causal relationship between PWE and dane chores and promise unprecedented levels of efficiency and pro- the adoption of autonomous products. We randomly assigned 105 ductivity. At the same time, anecdotal evidence suggests that even consumers from MTurk to an either high or low PWE condition, 854 / Novel Insights for Enduring Concerns: The Wisdom of Lay Theories employing a quotes-ranking task to induce a temporary state of PWE texts unrelated to brand extensions, prior research has shown that (Cheng et al. 2017). Subsequently, we asked consumers to choose individuals’ implicit theory orientation determines their reaction to between an autonomous versus manual cleaning device. Whereas failure (Aronson, Fried, and Good 2002). Because they believe in 62.3% of consumers chose the autonomous product in the low PWE the pliability of individual characteristics, incremental theorists per- condition, only 42.3% did so in the high PWE condition (χ2(1, 105) ceive failure as representative of an opportunity to learn and improve = 4.19, p = .04). These results were robust when controlling for iden- (Hong et al., 1999). In contrast, entity theorists attribute failure to a tity relevance of the task (Leung et al. 2019). lack of ability because as they see it, individuals cannot change and Study 4 aims at reversing the previously shown effect by proso, failure reflects undesirable traits and features of an individual viding a meaningful task substitute to consumers. We conducted a (Cury et al. 2008). Therefore, entity theorists are more likely to pe- lab study with 174 students to explore the effectiveness of this inter- nalize themselves and others when they learn about or encounter vention. Participants completed the quotes-ranking task of Study 3, failure (Haselhuhn et al. 2010). before being randomly assigned to one of two conditions. Whereas Recent research has also shown that compared to incremental the control condition entailed the same product description as in theorists, entity theorists are more likely to view social groups as Study 2, the meaningful condition entailed one extra attribute in the coherent and stable units and thus perceive the group to have greater autonomous product description (i.e., “allows you to spend time on entitativity (Rydell et al. 2007). We expect implicit theories to exert more meaningful things”). Results showed a significant interaction similar influence on consumers’ perceptions of a brand’s entitativ- between PWE and meaningful task substitute (b = .34, z = 2.15, p ity levels, such that entity theorists are more likely to view prod- = .03). In the control condition, more consumers chose the autono- ucts under the same brand name as constituting a meaningful and mous product in the low PWE condition (52.5%) than in the high entitative group (Wheeler and Omair 2016). Following this line of PWE condition (44.2%). In the meaningful condition, however, research (e.g., McConnell et al. 1997), we posit that when a brand fewer consumers chose the autonomous product in the low PWE is perceived as more entitative, consumers will incorporate and inte- condition (38.9%) than in the high PWE condition (63.6%), provid- grate extension failure information into their evaluation of the parent ing an effective intervention for firms to increase the adoption of brand. As a result, the negative associations of a failed extension are autonomous products. more likely to be transferred to the parent brand and produce greater We presented a series of field and experimental studies demon- attitude change (i.e., a steeper penalty for the parent brand). strating that PWE seems to inhibit the adoption of autonomous prod- In sum, we propose that when evaluating a parent brand, entity ucts, because high PWE consumers place more emphasis on mean- (vs. incremental) theorists will view the brand as a cohesive group ing from manual tasks. In other words, consumers that place more (vs. a loose collection of products). Because of these differences in emphasis on the “no pain, no gain” lay theory choose autonomous their perceptions and beliefs about entitativity, we further posit that products less often. The findings contribute to our understanding of entity theorists will be more likely to incorporate brand extension the adoption of new technologies and the psychological forces that failure into parent brand judgments, leading them to revise their govern this effect. evaluation of the parent brand. That is, we predict that brand dilu- tion (i.e., a negative feedback effect) is more likely for entity (vs. Brand Extension Failure and Parent Brand Penalty: The incremental) theorists, and this effect is expected to occur through Role of Implicit Theories perceived brand entitativity. We test these predictions in a series of five studies. In study EXTENDED ABSTRACT 1A, we measured implicit theory as a dispositional trait and found Companies routinely attempt to leverage the equity of an es- that entity (vs. incremental) theorists penalize a real parent brand tablished parent brand by introducing products with the same brand (i.e., Virgin) more after the failure of a brand extension (i.e., smart name in one or more categories (Ourusoff et al. 1992). Although re- watches). An OLS regression of the six-item parent brand penalty lying on a strong brand can sometimes be an effective strategy when measure (α = .63) on the implicit theory measure revealed that par- launching new offerings (Aaker and Keller 1990), over 80% of brand ticipants’ implicit theory orientation significantly influenced their extensions fail (e.g., Torelli and Ahluwalia 2011). Given this extraor- intention to penalize the parent brand (β = -.21, SE = .10, t(103) = dinary failure rate, it is imperative for companies to understand the -2.09, p = .039). In study 1B, we replicated this result using another consequences of extension failures on established brands. Although real brand (General Electric) and an established manipulation of im- previous research has produced valuable insights as to whether such plicit theory. In studies 2 and 3, we provided evidence for our follow feedback effects are influenced by features of the parent brand and model by showing that the greater parent brand penalty conferred the extension, relatively less work has examined the role of consum- by entity (vs. incremental) theorists is mediated by perceptions of a er characteristics. To fill this gap, the authors of the present research parent brand’s entitativity. investigate whether consumers’ implicit theory orientation, which is In study 2, we measured participants’ implicit theory and brand their perspective on whether personality traits are malleable versus entitativity beliefs and found that stronger entity theory orientation fixed, influences the severity of negative feedback effects following led to higher penalties for the parent brand (General Electric) after an extension failure. extension failure (β = -.25, SE = .10, t(498) = -2.42, p = .016). We Extant literature has documented that entity theorists, who be- also found that an entity (vs. incremental) theory orientation led to lieve in the fixedness of personality traits, are often less likely to a higher perception of brand entitativity (β = -.26, SE = .05, t(498) change their existing beliefs than incremental theorists, who believe = -4.77, p < .001) and that greater entitativity was associated with that personality traits are malleable (Dweck, Chiu, and Hong 1995). higher penalty (β = .20, SE = .08, t(498) = 2.47, p = .014). A me- As a result of this difference, one might expect that individuals who diation analysis revealed that the effect of implicit theory on parent endorse an entity (vs. incremental) theory orientation to be less likely brand penalty was mediated through perceived entitativity (β = -.04, to change parent brand evaluations after an extension failure. In con- SE = .02, 95% confidence interval [CI] = -.096, -.002). In study 3, we trast, we predict that entity (vs. incremental) theorists will be more replicated this finding even though we manipulated rather than mea- punitive when they learn about an extension failure. Albeit in con- sured implicit theory and used a fictitious fashion and apparel brand Advances in Consumer Research (Volume 48) / 855 named Epic that had launched a failed shoe extension. In study 4, Haselhuhn, Michael P., Maurice E. Schweitzer, and Alison M. we extended these findings by demonstrating an asymmetry in how Wood (2010), “How implicit beliefs influence trust recovery,” entity theorists respond to brand extension failure versus extension Psychological Science, 21 (5), 645-648. success, again using the General Electric brand. We found that while Hayes, Andrew F. (2013), An Introduction to Mediation, entity theorists penalize brands when they encounter brand extension Moderation, and Conditional Process Analysis: A Regression- failure, they do not change their parent brand evaluations when the Based Approach, New York: Guilford Press extension is successful. Heather, Hartley (2015), “Apéritif: Speeding with Françoise This work contributes to the literature on implicit theory, brand Sagan,” https://tinhouse.com/aperitif-speeding-with-francoise- extensions, and implicates perceived brand entitativity as a determi- sagan/. nant of the effects of failed extensions. John, Deborah Roedder (1999), “Consumer Socialization of Children: A Retrospective Look at Twenty-five Years of REFERENCES Research,” Journal of Consumer Research, 26 (3), 183-213. Aaker, David A., and Kevin Lane Keller (1990), “Consumer KPMG (2018), “Autonomous Vehicles Readiness Index: Assessing evaluations of brand extensions,” Journal of Marketing, 54 Countries’ Openness and Preparedness for Autonomous (1), 27-41. Vehicles,” https://assets.kpmg/content/dam/kpmg/nl/pdf/2018/ André, Quentin, Ziv Carmon, Klaus Wertenbroch, Alia Crum, sector/automotive/autonomous-vehicles-readiness-index.pdf. Douglas Frank, William Goldstein, et al. (2018), “Consumer Leung, Eugina, Gabriele Paolacci, and Stefano Puntoni (2018), Choice and Autonomy in the Age of Artificial Intelligence and Man Versus Machine: Resisting Automation in Identity-Based Big Data,” Customer Needs and Solutions, 5 (1-2), 28-37. Consumer Behavior,” Journal of Marketing Research, 55 (6), Aronson, Joshua, Carrie B. Fried, and Catherine Good (2002), 818-831. “Reducing the effects of stereotype threat on African American McCaffree, Jim (2003), “Position of the American Dietetic college students by shaping theories of intelligence,” Journal Association: Integration of Medical Nutrition Therapy of Experimental Social Psychology, 38 (2), 113-125. and Pharmacotherapy,” Journal of the American Dietetic Baumeister, Roy F., Kathleen D. Vohs, Jennifer L. Aaker, and Association, 103 (10), 1363-70. Emily N. Garbinsky (2013), “Some Key Differences Between McConnell, Allen R., Steven J. Sherman, and David L. Hamilton a Happy Life and a Meaningful Life,” The Journal of Positive (1997), “Target entitativity: Implications for information Psychology, 8 (6), 505-516. processing about individual and group targets,” Journal of Birch, Leann L., David Birch, Diane W. Marlin, and Laurie Kramer Personality and Social Psychology, 72 (4), 750. (1982), “Effects of Instrumental Consumption on Children’s McFerran, Brent, and Anirban Mukhopadhyay (2013), “Lay Food Preference,” Appetite, 3 (2), 125-34. Theories of Obesity Predict Actual Body Mass,” Psychological Carroll, Joseph (2006), “American Workers Would be Happier Science, 24 (8), 1428-36. in Same Job With Higher Pay,” http://news.gallup.com/ Moore, Elizabeth S., William L. Wilkie, and Debra M. Desrochers poll/24016/american-workers-would-happier-same-job-higher- (2017), “All in the Family? Parental Roles in the Epidemic of pay.aspx. Childhood Obesity,” Journal of Consumer Research, 43 (5), Cheng, Yimin, Anirban Mukhopadhyay, and Rom Y. Schrift (2017), 824–59. “Do Costly Options Lead to Better Outcomes? How the Newport, Frank (2018), “Americans Want Government to Do Protestant Work Ethic Influences the Cost-Benefit Heuristic More on Environment,” https://news.gallup.com/poll/232007/ in Goal Pursuit,” Journal of Marketing Research, 54 (4), 636- americans-want-government-more-environment.aspx 649. Ourusoff, Alexandra, Michael Ozanian, Paul B. Brown, and Jason Cury, François, David Da Fonseca, Ista Zahn, and Andrew Elliot, Starr (1992), “What’s in a name? What the world’s top brands “Implicit theories and IQ test performance: A sequential are worth,” Financial World, 161 (17), 32-49. mediational analysis,” Journal of Experimental Social Pogson, Corrie E., Alana B. Cober, Dennis Doverspike, and James Psychology, 44 (3), 783-791. R. Rogers (2003), “Differences in Self-Reported Work Ethic Dweck, Carol S., Chi-yue Chiu, and Ying-yi Hong (1995), “Implicit Across Three Career Stages,” Journal of Vocational Behavior, theories and their role in judgments and reactions: A word 62 (1), 189-201. from two perspectives,” Psychological Inquiry, 6 (4), 267-285. Raghunathan, Rajagopal, Rebecca Walker Naylor, and Wayne Economist Intelligence Unit (2018), “The Automation Readiness D. Hoyer (2006), “The Unhealthy=Tasty Intuition and Its Index: Who Is Ready for the Coming Wave of Automation?” Effects on Taste Inferences, Enjoyment, and Choice of Food http://www.automationreadiness.eiu.com/. Products,” Journal of Marketing, 70 (4), 170-84. Frederick, Shane and George Loewenstein (1999), “Hedonic Raskin, Sam, Bensimon, Olivia, and Weissman, Ruth (2020), adaptation,” in Well-being: Foundations of Hedonic “New Yorkers want to bag the new plastic ban,”https://nypost. Psychology, eds. Daniel Kahneman, Ed Diener and Norbert com/2020/02/29/new-yorkers-want-to-bag-the-new-plastic- Schwarz, New York, NY: Russell Sage Foundation, 302-29. ban/ Glanz, Karen, Michael Basil, Edward Maibach, Jeanne Goldberg, Rijsdijk, Serge A. and Erik Jan Hultink (2003), “‘Honey, Have You and D. A. N. Snyder (1998), “Why Americans Eat What They Seen Our Hamster?’ Consumer Evaluations of Autonomous Do: Taste, Nutrition, Cost, Convenience, and Weight Control Domestic Products,” Journal of Product Innovation Concerns as Influences on Food Consumption,”Journal of the Management, 20 (3), 204-216. American Dietetic Association, 98 (10), 1118-26. Rijsdijk, Serge A. and Erik Jan Hultink (2009), “How Today’s Hammond, Ross A (2010), “Social Influence and Obesity,” Current Consumers Perceive Tomorrow’s Smart Products,” Journal of Opinion in Endocrinology, Diabetes and Obesity, 17 (5), 467- Product Innovation Management, 26 (1), 24-42. 71. 856 / Novel Insights for Enduring Concerns: The Wisdom of Lay Theories

Rydell, Robert J., Kurt Hugenberg, Devin Ray, and Diane M. Torelli, Carlos J., and Rohini Ahluwalia (2012), “Extending Mackie. “Implicit theories about groups and stereotyping: The culturally symbolic brands: A blessing or a curse?,” Journal of role of group entitativity.” Personality and Social Psychology Consumer Research, 38 (5), 933-947. Bulletin, 33 (4), 549-558. Ward, Scott (1974), “Consumer Socialization,” Journal of Schmitt, Bernd (2019), “From Atoms to Bits and Back: A Research Consumer Research, 1 (2), 1-14. Curation on Digital Technology and Agenda for Future Wertenbroch, Klaus, Rom Y. Schrift, Joseph W. Alba, Alixandra Research,” Journal of Consumer Research, 46 (4), 825-832. Barasch, Amit Bhattacharjee, Markus Giesler, et al. (2019), Schnell, Tatjana and Peter Becker (2006), “Personality and “Autonomy in Consumer Choice,” Marketing Letters, Meaning in Life,” Personality and Individual Differences, 41 forthcoming. (1), 117-129. Wheeler, S. Christian, and Akhtar Omair (2016), “Potential growth Schrift, Rom Y., Ran Kivetz, and Oded Netzer (2016), areas for implicit theories research,” Journal of Consumer “Complicating Decisions: The Work Ethic Heuristic and the Psychology, 26 (1), 137-141. Construction of Effortful Decisions,” Journal of Experimental World Health Organization (2018), “Obesity and Overweight,” 16 Psychology: General, 145 (7), 807-829. February 2018, (accessed November 15, 2018) [available at http://www.who.int/mediacentre/factsheets/fs311/en/]. Feeling Unique and Grounded: Psychological Drivers and Consequences of Customer Engagement in Product Creation Chairs: Rocío Alarcón López, University of Murcia, Spain Anne-Kathrin Klesse, Erasmus University, Rotterdam School of Management, The Netherlands Martin Schreier, WU Vienna, Austria

Paper #1: You’re One in a Million: Strict Uniqueness of Mass- uct (i.e., consumers feel more proud), because sharing one’s choices Customized Products with others provides consumers with the possibility to express their Pia Burghartz, University of St. Gallen, Switzerland unique identity. Emanuel de Bellis, University of St. Gallen, Switzerland Rather than examining drivers of consumer engagement, the re- Franziska Krause, EBS University of Business and Law, maining two papers discuss consequences of engaging in the product Germany creation process. The third paper by Kaiser, Schreier and Janisze- Nikolaus Franke, WU Vienna, Austria wski reveals that feelings of uniqueness prompted by customizing or Gerald Häubl, University of Alberta, Canada producing one’s own product give rise to entitled behavior. This may have positive (such as donating more money to charity) or negative Paper #2: When I Share, I Feel Unique: Sharing one’s outcomes (such as demanding higher salaries). Relatedly, the fourth Customization Experience Increases Feelings of Pride paper by documents that engaging in self-production Rocío Alarcón López, University of Murcia, Spain Monnier et al . leads to feelings of groundedness (i.e. feelings of deep connected- Anne-Kathrin Klesse, Erasmus University, Rotterdam School ness to one’s social, historical, and physical environment). Addition- of Management, The Netherlands ally, the authors explore ways in which companies can prompt vicari- Salvador Ruiz de Maya, University of Murcia, Spain ous feelings of self-production and, in doing so, increase feelings of Inés López López, University of Murcia, Spain groundedness. Paper #3: Me, me, me: Customization, Uniqueness, and Entitled Taken together, the four papers in this session answer the fol- Behavior lowing questions: What are psychological benefits of engaging Ulrike Kaiser, WU Vienna, Austria customers in the product creation process? How can companies fa- Martin Schreier, WU Vienna, Austria cilitate these benefits? Importantly, the papers utilize multi-method Chris Janiszewski, University of Florida, USA approaches, combining data from online, lab, and field studies to Paper #4: How Self-Production Increases Feelings of explore the robustness and generalizability of the proposed effects. Groundedness Broadly, this session speaks to ACR’s call for sessions on the topic Arnaud Monnier, Cornell University, USA of identity by highlighting how customer engagement in the product Isabel Eichinger, WU Vienna, Austria creation process can allow consumers express their unique identity Martin Schreier, WU Vienna, Austria and help them feel reconnected with themselves. Thus, we believe Stijn van Osselaer, Cornell University, USA that this session should not only be of interest to scholars working on customization and self-production, but also to those interested SESSION OVERVIEW in identity-related topics, such as self-expression, uniqueness, and Instead of buying readily available products off the shelf, con- groundedness. sumers frequently engage in the product creation process and customize offerings or self-produce products. For instance, consumers You’re One in a Million: Strict Uniqueness of Mass- may customize a birthday card by choosing their own colors, fonts, Customized Products and symbols, or create the card completely from scratch. Engaging consumers in the product creation process has several positive con- EXTENDED ABSTRACT sequences for companies, such as more positive product evaluations In this research we investigate extensions of mass customiza- (Bendapudi and Leone 2003), higher individual demand (Franke, tion (MC) systems that highlight the uniqueness of a configured Schreier, and Kaiser 2010) and greater willingness to pay (Norton, product. It examines the behavioral consequences of informing con- Mochon, and Ariely 2012), as well as for consumers, such as feel- sumers that they are the first ever to have created a particular product ings of competence (Dahl and Moreau 2007) and accomplishment configuration (strict uniqueness feedback) and in addition assuring (Norton et al. 2012). that this product configuration will remain unique (strict uniqueness This special session is intended to deepen our knowledge on blocking). the consequences for consumers by focusing on two specific benefits Past research has identified preference fit (Dellaert and of customer engagement: feeling unique and grounded. Whereas all Stremersch 2005) and process-related benefits (Troye and Supphel- four papers share the emphasis on psychological benefits, they differ len 2012) as sources of consumer value in MC systems. Our findings in the specific perspective they take. The first two papers focus on shed light on the role of product uniqueness (Franke and Schreier need for uniqueness as a salient motivation to customize products 2008) as a value-generating force in MC systems. We conceptualize and examine strategies that companies can utilize to boost feelings of a strictly unique product as a product that is literally one of a kind— uniqueness. The first paper by Burghartz et al . examines providing a combination of features that exists only once. This expands our uniqueness feedback and the option to block customers’ creation as understanding of the construct of product uniqueness, and of how strategies to increase consumers’ valuation of the customized prod- consumers respond to it, beyond products that are merely rare but uct. Similarly, the second paper by Alarcón et al . documents that not strictly unique (Lynn and Harris 1997). Whereas prior research sharing one’s customization experience with others (e.g., in a person- focused on non-automated social feedback, our findings reveal that al message) increases the subjective value of the customized prod- automated feedback on strict product uniqueness can be a significant driver of consumer value (Franke and Hader 2014). Advances in Consumer Research 857 Volume 48, ©2020 858 / Feeling Unique and Grounded: Psychological Drivers and Consequences of Customer Engagement in Product Creation

We propose that feedback on the strict uniqueness of a product cant effect of strict uniqueness on WTP. We found a significant effect

(“Your product configuration has never been created before”) cre- of strict uniqueness on WTP (Munique_blocking = US$50.02, Munique_feed- ates value for consumers and that being informed that the strictly back = US$43.94, Mnonunique_feedback = US$35.47; F(2, 169) = 10.46, p unique product configuration was blocked (“Your configuration will < .001). Planned contrasts showed a significant difference in WTP be blocked for future customers”) creates additional value. between the non-uniqueness and uniqueness feedback (p=.024) and the uniqueness feedback and blocking condition (p=.002). Hypothesis 1a: Feedback on strict product uniqueness increases The results also demonstrate that the effect of uniqueness feed- consumers’ valuation of their customized prod- back and uniqueness blocking on WTP is mediated by perceived ucts. exclusivity. We conducted two mediation analyses with a categorical predictor (PROCESS Model 4; Hayes 2013). The results show that uniqueness feedback, blocking and non-uniqueness feedback Hypothesis 1b: Blocking strict product uniqueness further in- indirectly influence WTP through perceived exclusivity. The effect creases consumers’ valuation of their custom- between uniqueness feedback and non-uniqueness feedback and be- ized products. tween uniqueness feedback and blocking on WTP is positive and Moreover, this research theorizes that perceived exclusivity— significant (a = 1.86, p < .001; a = 1.05, p < .001); in support of the i.e., the perception that a product is not available to others—can be indirect effect (ab < .001), the bias-corrected 95% confidence inter- a key driver of consumers’ valuation of a product. We propose that val based on 1,000 boot-strapped samples (marginally) excludes 0 strict uniqueness feedback and (to a larger extent) strict uniqueness ([.08, .34]; [-.00, .19]. blocking should trigger perceptions of exclusivity which in turn Experiment 3 examines whether making the (large) number of should create value for consumers. sneaker configuration possibilities transparent to consumers attenuates the effect of uniqueness feedback and uniqueness blocking on Hypothesis 2: Strict uniqueness feedback and strict unique- consumers’ product valuation (H3). The findings demonstrate that ness blocking increase consumers’ valuation of the effect of uniqueness feedback and uniqueness blocking on WTP their customized products via the perception of is moderated by solution space transparency. A one-way mixed de- greater exclusivity. sign ANCOVA with solution space transparency as between-subjects factor, strict uniqueness as within-subject factor, WTP as dependent We further propose that consumers’ valuation of strict variable, and brand attitude, status consumption, age, gender, and uniqueness is attenuated when the (large) number of possible prod- income as covariates showed a significant interaction between so- uct configurations in MC systems is made transparent to them. By lution space transparency and strict uniqueness (Mnontransparent&no_feed- making the solution space of MC systems transparent to consumers back = €78.94, Mtransparent&o_feedback = €81.82, Mnontransparent&unique_feedback = the subjective value they attribute to their configured product likely €86.90, Mtransparent&unique_feedback = €89.13, Mnontransparent&blocking = €100.02, decreases. Mtransparent&blocking = €96.73; F(2, 926) = 4.62, p = .010). While planned contrasts showed significant effects in the non-transparent (p=.005, Hypothesis 3: The effect of strict uniqueness feedback and p<.001), they showed non-significant effects in the transparent con- strict uniqueness blocking on consumers’ valu- ditions (p=.11, p=.14). ation of their customized products is attenuated The reported findings have important practical implications in when the number of possible product configura- that they identify an inexpensive way for firms to create consumer tions is made transparent. value by capitalizing on the fact that MC systems naturally promote consumers’ creation of products that are strictly unique. It appears We conducted three experiments across different samples and to be a promising and easy-to-implement way to enhance the value product domains. We also employed different study designs and consumers obtain from configuring products. types of dependent variables to test our hypotheses. Experiment 1 examines whether strict uniqueness feedback has a positive effect When I Share, I Feel Unique: Sharing one’s on the subjective value that consumers attribute to customized cereal Customization Experience Increases Feelings of Pride in a field setting (measured as conversion rate; H1a). Analyzing the conversion rates across conditions indicates that uniqueness feed- EXTENDED ABSTRACT back on the customized product creates value for customers. Specifi- Nowadays, product customization is omnipresent. Enabled by cally, we found a significant difference in conversion rates between technological advances, companies (e.g., Nike, Nissan, Gucci) re- the uniqueness feedback and the no feedback condition of 2.24% cently began to explicitly encourage customers to share their cus- 2 (χ (1, 33,001) = 16.50, p < .001). This increase in conversion ap- tomization experience with others. For instance, Ralph Lauren offers proximates to €1,145,601 in additional yearly revenue for the cereal customers the possibility to share their customization experience producer. with friends via email immediately after they completed the custom- Experiment 2 examines whether strict uniqueness blocking, ization process. Prompting customers to share their experience is in addition to strict uniqueness feedback, can drive the subjective uniquely suited to customization processes because consumers feel value consumers attribute to their customized sunglasses (measured proud of their creations (Franke et al. 2010; Gandhi et al. 2014). as WTP; H1b). Further, it tests whether the additional value consum- We explore whether the mere act of sharing one’s customiza- ers attribute to their configured product is mediated by perceived tion experience with others can boost affective reactions (i.e., pride) exclusivity (H2). The findings indicate that both uniqueness feed- toward the customized product. When engaging in product custom- back and uniqueness blocking significantly increased WTP using ization, customers typically make choices that allow them to express a between-subject design with non-uniqueness feedback as a third their unique identity (D’Angelo et al. 2019; Ratner and Kahn 2002). treatment condition. A one-way ANCOVA, with strict uniqueness as We argue that sharing these choices with others (e.g., in a personal factor, WTP as dependent variable, and brand attitude, status con- message) provides customers with the possibility to express their sumption, age, gender, and income as covariates showed a signifi- Advances in Consumer Research (Volume 48) / 859 unique identity even before they obtain the customized product. than those who keep it to themselves (M = 5.69; F(1,321) = 0.16 , Since most consumers have a strong need to differentiate themselves p = .69). from others (see Ariely and Levav 2000), being able to express one’s Study 4 tests whether the effect is pronounced for consumers unique identity should strengthen customers’ feelings of pride to- with high need for uniqueness (NFU) but attenuated (or reversed) for ward the customized product. those with low NFU. 215 females customized their own earrings by One pilot study and four experiments test our predictions. For choosing different hoops and charms; afterward participants either the pilot study, we collected data in a KitKat pop-up store that of- shared their customization experience with a friend or shared an un- fered customers the possibility to customize their own chocolate by related experience (i.e., their Summer holiday). Finally, everyone in- completing a number of steps (base chocolate, toppings, package dicated their feelings of pride and responded to the NFU measure by design). Once participants (n = 121) handed in their customization Chernev et al. 2011 (three items; e.g., “I would like to be perceived choices and before they received the chocolate, they participated in a as different from the general population”; 0 = strongly disagree, 10 = brief survey. Whereas some participants first shared their customiza- strongly agree). We find a significant interaction effect (β=.37, t(211) tion experience (i.e., their choices and how they felt about them; n = 3.27, p = .001). Sharing one’s customization experience results = 801) and then reported how tasty they expected their chocolate to in higher feelings of pride for participants high in NFU (i.e., values be (0 = not tasty at all, 10 = very tasty), other participants (n = 41) ≥ 7.99) but prompts lower feelings of pride for participants low in immediately indicated their taste perceptions. We find that sharing NFU (i.e., values ≤ 4.51). one’s customization experience increased the subjective value of the This work contributes to existing research on product custom- chocolate (M = 9.19) compared to not sharing it (M = 8.59, F(1,119) ization by highlighting an easy-to-implement strategy (i.e., encour- = 7.43, p = .007). aging social sharing of customization experiences) that companies Then, in study 1, participants (n = 171, prolific) customized can utilize to further boost feelings of pride, which are essential pre- a muesli by choosing a base, ingredients, a design for the muesli dictors for repurchasing intentions (Septianto et al. 2019) and self- box, and adding a personalized message. Afterward half of the par- brand connections (Williams et al. 2018). ticipants prepared a message to a friend to share their experience by explaining what choices they made and how they felt while custom- Me, me, me: Customization, Uniqueness, and Entitled izing the muesli. The other half engaged in an unrelated word-puzzle Behavior task (n = 87). We measured participants’ feelings of pride toward the customized muesli using a two-item scale: “I am proud of my cus- EXTENDED ABSTRACT tomized muesli”, and “My customized muesli makes me feel proud” Psychological entitlement – that is, a person’s “pervasive sense (0 = strongly disagree, 10 = strongly agree2). Sharing one’s custom- that one deserves more and is entitled to more than others” (Camp- bell et al. 2004, p. 31) – has become a major topic of interest to both ization experience increased feelings of pride (M = 7.27) compared to not sharing it (M = 6.39, F(1,169) = 8.29 , p = .005). In study 2 (n the general public and scholars. Entitled individuals believe they = 173, prolific), we utilized the same muesli paradigm but compared are special, and they expect others to treat them this way (Zitek and sharing one’s customization experience to an introspection condition Vincent 2014). Entitlement does not involve any reciprocity (e.g., (i.e., participants were instructed to think about their customization “I deserve a higher salary because I performed well”), but is expe- experience in private). We replicate our effect so that sharing one’s rienced regardless of one’s effort or performance relative to others (Campbell et al. 2004). customization experience results in greater feelings of pride (M = Most research to date has studied entitlement as an individual 7.57) than merely thinking about it (M = 6.78; F(1,171) = 7.23 , p = .008). This provides direct evidence that social sharing rather than differences characteristic. There is a large body of research on dif- simply deliberating on the customization experience prompts feel- ferent facets of entitlement, how it is related to other traits (e.g., nar- ings of pride. cissism, self-esteem, etc.), and on generational differences (Twenge In studies 3 and 4, we test theoretically derived boundary condi- and Campbell 2009). There is also an emerging body of research tions. First, we predict that sharing one’s customization experience on behavioral consequences. For example, entitlement is associated only prompts greater feelings of pride if the customization task al- with selfish choices, unethical behavior, unforgiveness, or lack of lows for expressing one’s identity but not if it is functional in nature perspective taking (Campbell et al. 2004, Exline et al. 2004, Vincent (Kaiser et al. 2017). Students in the laboratory (n = 325) customized and Kouchaki 2016). While most research has focused on negative their own fitness tracker; participants were randomly assigned to a outcomes, Zitek and Vincent (2014) are the exception. They show self-expressive customization task (e.g., choosing the strap color) or that a boost in entitlement can increase creativity, mediated by a a functional task (e.g., choosing the size for the wrist brand). Af- higher need for uniqueness associated with entitlement. terward participants either shared their customization experience In this paper, we focus on the less studied antecedents of entitle- by sending a message to a friend or engaged in a word-puzzle task ment. In particular, we are interested in exploring causal triggers that and then reported pride towards the fitness tracker. The results docu- lie outside the individual. To our knowledge, we are the first to dem- ment a marginally significant interaction effect (F(1,321) = 3.68 , p onstrate that uniqueness cues that are ubiquitous in the marketplace = .055). As expected, simple contrasts highlight that sharing one’s (think of customized products, limited editions, or advertising claims that tell consumers how unique and special they are) lead to entitled customization experience (M = 5.97) increases feelings of pride com- behaviors. We report five experiments that establish the link between pared to not sharing it (M = 5.12; F(1,321) = 5.17 , p = .024) for the self-expressive customization. However, consumers who share their perceived uniqueness and entitled behavior. functional customization experience (M = 5.55) feel equally proud In study 1 (n = 280, lab, undergraduates) we introduced a clothing style quiz. After completing the quiz, participants received feed- 1 Note that some participants (N = 41) shared their experience in back that either stated that their style was “in many respects unique writing whereas others (N= 39) shared it orally. Since there was no and special” or “similar to others”. Next, we measured salary de- difference between these groups, we merged the two conditions. manded to complete a job as a known proxy for entitled behavior. 2 We used the same dependent measure in all subsequent Participants were told that they could work on an extra task in return studies. 860 / Feeling Unique and Grounded: Psychological Drivers and Consequences of Customer Engagement in Product Creation for payment. We described the task in more detail (it was about a Binkley 2008; Brooks 2018; Husemann and Eckhardt 2019; Steiner blog entry), and participants were then asked to indicate the mini- 2005). We propose and show that engaging in self-production gener- mum payment to complete the job. Their choices had real economic ates feelings of groundedness, defined as feelings of deep connected- consequences. Findings show that the uniqueness feedback resulted ness to one’s physical, social, and historical environment. in higher average salary requests (MUnique = 16.35 euro vs. MControl = Previous research has established that self-production increases 14.28 euro, p = .01). product valuation (Franke, Schreier and Kaiser 2010; Norton, Mo- In study 2 (n = 182, Amazon Mechanical Turk) we used a simi- chon and Ariely 2012). This literature has also delineated some of lar quiz to manipulate uniqueness, but instead of salary request, we the psychological implications of engaging in self-production. For measured prosocial behavior as a proxy for entitled behavior. After individuals, these include feelings of competence (Dahl and Moreau participants received the uniqueness feedback (“your style is unique” 2007; Mochon, Norton, and Ariely 2012), feelings of accomplish- vs. “similar to others”), they were told that we would randomly pick ment (Franke et al. 2010; Norton et al. 2012), and task enjoyment one participant to get a bonus payment of $ 25. Participants were (Dahl and Moreau 2007). At the product level, implications include asked to indicate how much of the bonus payment they would like the transfer of positive self-views to the product, through a process to receive on their MTurk account, and how much they would like to of associative self-anchoring (Gawronski, Bodenhausen, and Becker donate to Doctors Without Borders (conditional on winning). Again, 2007; Troye and Supphellen 2012). their choice had real economic consequences. Results show that the Across lab and field studies conducted on two continents and uniqueness feedback increased donations (MUnique = $ 8.89 vs. MControl utilizing diverse product categories, this research documents a nov- = $ 6.01, p = .03). We argue that this is because their generosity el psychological benefit of engaging in self-production: increased served as a means to express that they are unique and special (Lee, feelings of groundedness. We further demonstrate how evoking the Gregg, and Park 2013). realm of self-production enables brands to generate vicarious feel- In study 3a (n = 97, MTurk) and 3b (n = 201, MTurk), we repli- ings of self-production, and subsequently increase groundedness. cate the effect on donation, using two different uniqueness manipula- In a first preregistered experiment, we tested the basic effect of tions. In study 3a, we used a writing prime. Participants were asked self-production versus mere consumption of a product on feelings to write a short text about “what makes you unique and special” vs. of groundedness. We randomly assigned 293 US students to one of “similar to others”. The dependent variable was the same as in study two conditions (self-production vs. control). In the self-production

2. Results show that perceived uniqueness increases donations (MU- condition, participants were instructed to make apple juice with local nique = $ 8.89 vs. MControl = $ 6.01, p = .03). Study 3b was a scenario apples. The procedure involved selecting, cutting and juicing the ap- study in which participants learned about a uniquely designed watch. ples. In the control condition, participants were simply provided with They were either asked to imagine wearing the watch or seeing an- the same apple juice made from local apples. All participants tasted other person wearing it. Next, we assessed their intention to donate the apple juice and answered our three-item groundedness scale (α to Doctors Without Borders (in this study, we used intentions instead = .88). The three items include two verbal statements (“I feel deep- of real behavior). Again, in the experimental condition, donation in- rooted”, “I feel well-grounded”) and a pictorial item, “How well tentions were higher (MUnique = $ 6.90 vs. MControl = $ 5.44, p = .08). does the following picture describe your feelings at this moment?”). Study 4 (n = 222, lab, undergraduates) is a 2 x 2 experiment that We found a main effect of self-production on feelings of grounded- employed a real customization task (self-customize vs. control). The ness (F(1,291) = 10.38, p =.001). Participants felt significantly more second factor was whether participants donated to charity in private grounded in the self-production condition (MSelf-production = 4.74, MControl or in public. Students participated in a lip balm workshop and either = 4.25). received a self-customized product or a standard product. After an- Study 2 tested whether merely consuming a product that evokes swering some filler questions, they were thanked for their participa- the realm of self-production can also elicit vicarious feelings of self- tion and received their monetary compensation. We paid 8 euro in production and, in turn, feelings of groundedness. We hypothesized coins. Participants were invited to donate all or some of the money to that artisanal products that resemble self-made goods would increase a charity. The donation box was either placed in a cubicle (private) or feelings of groundedness compared to industrial products, but less next to the experimenter (public). Findings show that after receiving so than self-producing. We further posited that this occurs through the self-customized product, participants donated more when the do- vicarious feelings of self-production. To explore this, we randomly nation was done in private (MUnique = 1.19 euro vs. MControl = 0.65 euro, assigned 251 European students in a lab to one of three conditions. p = .01), but not when they donated in public (MUnique = 1.12 euro vs. Participants in the first two conditions were asked to inspect a real

MControl = 1.25 euro, n.s.). In the private setting, being more generous bar of soap that was either industrial or artisanal (i.e., locally made served as means to express their uniqueness. In the public setting, by hand from organic ingredients and wrapped in a craft packag- this effect is attenuated due to social desirability effects. ing). They were asked to imagine having just purchased the bar of Taken together, five experiments show that marketplace-in- soap. Participants in the third condition were asked to imagine hav- duced feelings of uniqueness (e.g., through customization) lead to ing made a bar of soap themselves. All participants answered our entitled behaviors. We also show that entitlement is not only associ- groundedness scale (α = .91), as well as feelings of self-production ated with self-centered and selfish behaviors (e.g., demand higher (three items, e.g., “In a certain way, it almost feels as if I were the salary), but that it can also lead to favorable behaviors (e.g., proso- producer of the soap”, α = .97). We found a significant treatment cial behavior). effect on feelings of groundedness(F (2,248) = 21.30, p < .001). Par- ticipants who inspected the artisanal soap reported higher feelings

How Self-Production Increases Feelings of Groundedness of groundedness (MArtisanal = 4.32) compared to those who inspected

the industrial soap (MIndustrial = 3.35, p < .001). Participants in the EXTENDED ABSTRACT self-production condition reported even more pronounced feelings In an increasingly fast-paced, urbanized, and digitized world, of groundedness (vs. the artisanal condition, MSelf-production = 4.77, p = individuals often feel uprooted from their environment. Accordingly, .037). We also found a significant treatment effect on feelings of self- many frequently experience a need for reconnection (Asimov 2017; production (F(2,248) = 106.71, p < .001). Participants who inspected Advances in Consumer Research (Volume 48) / 861 the artisanal (vs. industrial) soap reported higher feelings of self- D’Angelo, Jennifer K., Kristin Diehl, and Lisa A. Cavanaugh production (MArtisanal= 2.92, MIndustrial = 1.63, p < .001). Participants in (2019), “Lead by Example? Custom-Made Examples Created the self-production condition had even higher feelings of self-pro- by Close Others Lead Consumers to Make Dissimilar duction (vs. the artisanal condition, MSelf-production= 5.05, p < .001). The Choices,” Journal of Consumer Research, 46 (4), 750-773. groundedness effect was mediated by feelings of self-production for Dahl, Darren W. and C. Page Moreau (2007), “Thinking Inside both the artisanal versus industrial (indirect effect = .64, CI95%: .39, the Box: Why Consumers Enjoy Constrained Creative .94) and the self-production versus artisanal comparison (indirect ef- Experiences,” Journal of Marketing Research, 44 (3), 357- fect = .71, CI95%: .44, 1.06). 369. Study 3 aimed to replicate our findings in a field setting, using Dellaert, Benedict G. C. and Stefan Stremersch (2005), “Marketing yet another product category (i.e., birthday cards). Additionally, we Mass-Customized Products: Striking a Balance Between examined plausible alternative explanations such as mood and per- Utility and Complexity,” Journal of Marketing Research, 42 ceived quality. The study was conducted in a large museum in a Eu- (2), 219–27. ropean city (n = 269 museum visitors). In one condition, participants Exline, Julie Juola, Roy F. Baumeister, Brad J. Bushman, and W. were given a blank card and color pencils, and asked to self-produce Keith Campbell, and Eli J. Finkel (2004), “Too Proud to Let a birthday card. Two additional conditions exposed participants to an Go: Narcissistic Entitlement as a Barrier to Forgiveness,” either industrial or artisanal birthday card. All participants read that Journal of Personality and Social Psychology, 87 (6), 894- other museum visitors would receive these cards. Beyond feelings 912. of groundedness (α = .91) and feelings of self-production (α = .97), Franke, Nikolaus and Christopher Hader (2014), “Mass or we captured alternative accounts: time on the task, perceived quality, only “Niche customization”? Why we Should Interpret mood, thoughts about close others and thoughts about birthdays. We Configuration Toolkits as Learning Instruments,”Journal of found a main effect of condition on feelings of groundedness that Product Innovation Management, 31 (6), 1214–34. parallels that of Study 2 (Mself-production = 4.71, MArtisanal = 3.98, MIndustrial = Franke, Nikolaus and Martin Schreier (2008), “Product Uniqueness

3.08, F(2,252) = 28.86, p < .001 ). The same pattern emerged on feel- as a Driver of Customer Utility in Mass Customization,” ings of self-production. We found a main effect of condition (Mself- Marketing Letters, 19 (2), 93–107. production = 6.11, MArtisanal = 3.24, MIndustrial = 1.42, F(2,254) = 222.31, p Franke, Nikolaus, Martin Schreier, and Ulrike Kaiser (2010), < .001). The groundedness effect was again mediated by feelings of “The “I Designed it Myself” Effect in Mass Customization,” self-production for both the artisanal versus industrial (indirect effect Management Science, 56 (1), 125-140.

= .55, CI95%: .31, .87), and the self-production versus artisanal com- Gandhi, Anshuk, Carmen Magar, and Roger Roberts (2014), parison (indirect effect = .40, CI95%: .06, .79). Regarding alternative “How Technology Can Drive the Next Wave of Mass accounts, mood was higher in the self-production condition, but did Customization,” Business Technology Office, 1-8. not vary between the two others. Perceived quality significantly dif- Gawronski, Bertram, Galen V. Bodenhausen, and Andrew P. fered across conditions, but the mediation held when adding quality Becker (2007), “I Like it, Because I Like Myself: Associative as covariate. No other measure varied across conditions. Self-Anchoring and Post-Decisional Change of Implicit Our research demonstrates a robust effect of self-production Evaluations,” Journal of Experimental Social Psychology, 43 on feelings of groundedness, and illuminates how brands can also (2), 221-232. beget such feelings. We thus identify groundedness as a distinct and Hayes, Andrew F. (2013), Methodology in the Social Sciences. critical psychological need that consumers fulfil when engaging in Introduction to Mediation, Moderation, and Conditional self-production. Process Analysis. A Regression-Based Approach. New York, USA: Guilford Press. REFERENCES Husemann, Katharina C. and Giana M. Eckhardt (2019), Ariely, Dan and Jonathan Levav (2000), “Sequential Choice in “Consumer Deceleration,” Journal of Consumer Research, 45 Group Settings: Taking the Road Less Travelled and Less (6), 1142-1163. Enjoyed,” Journal of Consumer Research, 27 (3), 279-290. Kaiser, Ulrike, Martin Schreier, and Chris Janiszewski, (2017), Asimov, Eric (2017), “Much Excitement on the Oregon Trail,” New “The Self-Expressive Customization of a Product can Improve York Times, September 20, D4. Performance,” Journal of Marketing Research, 54 (5), 816- Bendapudi Neeli and Robert P. Leonne (2003), “Psychological 831 Implications of Customer Participation in Co-Production,” Lee, Seung, Aiden Gregg, and Seong Park (2013), “The Person in Journal of Marketing, 67 (1), 14-28. the Purchase: Narcissistic Consumers Prefer Products That Binkley, Sam (2008), “Liquid Consumption: Anti-Consumerism Positively Distinguish Them,” Journal of Personality and and the Fetishized De-Fetishization of Commodities,” Cultural Social Psychology, 105 (2), 335-352. Studies, 22 (5), 599-623. Lynn, Michael and Judy Harris (1997), “The Desire for Unique Brooks, Arthur C. (2018), “Loneliness Is Tearing America Apart,” Consumer Products: A New Individual Differences Scale,” New York Times, November 24, A25. Psychology & Marketing, 14 (6), 601–16. Campbell, W. Keith, Angelica M. Bonacci, Jeremy Shelton, Julie Mochon, Daniel, Michael I. Norton, and Dan Ariely (2012), J. Exline, and Brad J. Bushman, (2004), “Psychological “Bolstering and Restoring Feelings of Competence via the Entitlement: Interpersonal Consequences and Validation of a IKEA Effect,”International Journal of Research in Marketing, Self-Report Measure,” Journal of Personality Assessment, 83, 29 (4), 363-369. 29–45. Moreau, C. Page, Leff Bonney, and Kelly B. Herd (2011), “It’s Chernev, Alexander, Rebecca Hamilton, and David Gal (2011), the Thought (and the Effort) that Counts: How Customizing “Competing for Consumer Identity: Limits to Self-expression for Others Differs from Customizing for Oneself,”Journal of and the Perils of Lifestyle Branding,” Journal of Marketing, Marketing, 75 (5), 120-133. 75 (3), 66-82. 862 / Feeling Unique and Grounded: Psychological Drivers and Consequences of Customer Engagement in Product Creation

Norton, Michael I., Daniel Mochon, and Dan Ariely (2012), Twenge, Jean M. and W. Keith Campbell (2009), The Narcissism “The IKEA Effect: When Labor Leads to Love,”Journal of Epidemic: Living in the Age of Entitlement. New York: Free Consumer Psychology, 22 (3), 453-460. Press. Ratner, Rebecca and Barbara Kahn (2002), “The Impact of Private Vincent, Lynne C. and Maryam Kouchaki (2015), “Creative, Rare, versus Public Consumption on Variety-Seeking Behaviour,” Entitled, and Dishonest: How Commonality of Creativity Journal of Consumer Research, 29 (2), 246-257. in One’s Group Decreases an Individual’s Entitlement and Septianto, Felix, Jake An, Tung Moi Chiew, Widya Paramita, and Dishonesty,” Academy of Management Journal, 59 (4), 1451- Istiharini Tanudharma (2019), “The Similar versus Divergent 1473. Effects of Pride and Happiness on the Effectiveness of Loyalty Williams, Patti, Nicole Verrochi Coleman, Andrea C. Morales, and Programs,” Journal of Business Research, 99, 12-22. Ludovica Cesareo (2018), “Connections to Brands that Help Steiner, Rudolf (2005), What is Biodynamics: A Way to Heal and Others versus Help the Self: The Impact of Incidental Awe Revitalize the Earth, Fair Oaks, CA: Steiner. and Pride on Consumer Eelationships with Social-Benefit and Troye, Sigurd Villads and Magne Supphellen (2012), “Consumer Luxury Brands,” Journal of the Association for Consumer Participation in Coproduction: “I Made it Myself” Effects on Research, 3 (2), 202-215. Consumers’ Sensory Perceptions and Evaluations of Outcome Zitek, Emily M. and Lynne C. Vincent (2015), “Deserve and and Input Product,” Journal of Marketing, 76 (2), 33–46. Diverge: Feeling Entitled Makes People More Creative,” Journal of Experimental Social Psychology, 56, 242–248. Antecedents of and Remedies to the Spread of False Information in the Social Media Era Chair: Gizem Ceylan, University of Southern California, USA

Paper #1: Look What I am Re-Sharing: How Self-Presentation accuracy by other salient motives when choosing what to share. As Goals Impact What Consumers Spread on Social Networks a remedy to the problem, they suggest that subtly inducing people to Gizem Ceylan, University of Southern California, USA think about the concept of accuracy increases the quality of the news Norbert Schwarz, University of Southern California, USA they share. Ding and Johar propose a method to leverage the input of the Paper #2: Impression Management in the Echo-Chamber: How general population (crowdsourcing), algorithm (supervised learn- Self-Censorship Biases Evidence-Sharing ing), and experts (third-party checkers) to detect false information. Ike Silver, University of Pennsylvania, USA They suggest that asking readers to compare the opinions and view- Deborah Small, University of Pennsylvania, USA points of two articles (as opposed to asking them to directly evalu- Geoff Goodwin, University of Pennsylvania, USA ate the source or the articles) can be a way to overcome motivated Paper #3: Understanding and Reducing the Spread of reasoning among people with different ideologies. Misinformation Online Taken together, these papers document novel and timely in- Gordon Pennycook, University of Regina, Canada sights into important implications of spread of false information. All Ziv Epstein, Massachusetts Institute of Technology, USA papers are at an advanced stage of development with multiple studies Mohsen Mosleh, Massachusetts Institute of Technology, USA completed. We expect this session to generate strong interest among Antonio A. Arechar, Massachusetts Institute of Technology, researchers studying false information, impression management, so- USA cial influence and social judgment, as well as social media and word- Dean Eckles, Massachusetts Institute of Technology, USA of-mouth more broadly. David G. Rand, Massachusetts Institute of Technology, USA Paper #4: Factchecking Matters: The Value of Crowdsourcing Look What I am Re-Sharing: How Self-Presentation for Enhanced Accuracy Judgments Goals Impact What Consumers Re-Transmit on Social Yu Ding, Columbia University, USA Networks Gita V. Johar, Columbia University, USA EXTENDED ABSTRACT SESSION OVERVIEW Information spreads fast on social platforms. Consumers re- There is worldwide concern over false news and the possibility transmit news on Facebook, re-share others’ photos on Instagram, that it can influence political, economic, and social well-being. From re-tweet the quotes from their beloved companies or political parties 2006 to 2017, false news reached more people and spread faster than (e.g., Chevalier and Mayzlin 2006; Trusov, Bucklin, and Pauwels truth (Vosoughi, Roy, and Aral 2018). While false information seems 2009). These decisions seem to be made in seconds. Consumers seem to diffuse faster, when asked, consumers overwhelmingly say that it to be quicker to share than to read the information. A recent Twit- is important for them to share accurate content only (Pennycook et ter study found that 59 percent of all links shared went unclicked, al. 2019). The presentations in this session examine this puzzling and presumably unread (Gabielkov, Ramachandran, Chaintreau, and phenomenon and identify potential antecedents of and remedies to Legout 2016). Instead of attending to central elements (i.e., the qual- the spread of false information on social networks. In this session, ity of information), consumers seem to attend to peripheral cues, we explore whether (presentations 1) and how (presentation 2) con- such as the source of the information. We suggest that this quick sumers’ self-presentation goals impact what they spread on social evaluation process may be moderated by consumers’ salient self- networks. Further, we explore potential remedies that can improve presentation goals. the quality of the information consumers share on social media both Prior research showed that consumers evaluate content differ- from the sharer (i.e., consumer; presentation 3) as well as the source ently depending on their salient motivation in their social network. (i.e., news media; presentation 4) perspective. For instance, they seem to share more self-presenting content (vs. Ceylan and Schwarz explore spreading information (e.g., useful content) when sharing with multiple others (vs. one other; news) on social media as a strategic form of self-disclosure. They Barasch and Berger 2014). Our research examines the impact of self- find that individuals with the motivation to fit in share news only if it presentation goals on sharing decisions and shows that consumers comes from a high-credibility source or is also shared by many oth- who want to fit in prefer to share popular (vs. unpopular) informa- ers (high-consensus). However, those motivated to stand out are less tion from well-known (vs. unknown) and credible (vs. less credible) sensitive to source credibility and consensus, and may be more likely sources, whereas consumers who want to stand out are willing to to contribute to the spread of false information. share regardless of the source and popularity of the content. Despite Silver, Small, and Goodwin examine a motivated sharing ef- these differences in sharing, both groups of consumers find- infor fect, suggesting that consumers prefer to omit attitude-incongruent, mation more truthful when it comes from a well-know and familiar factual information from conversations with others. The authors source, indicating that those who want to stand out are willing to predict and find that political allies provide one another with biased share information they may not consider truthful. information, strategically omitting factual and relevant evidence if Study 1 . Following a 3 (Self-Presentation Goal: Fit-in vs. it casts doubt on their commitment to a shared cause. This effect is Stand-out vs. Control) x 2 (Credibility: High vs. Low) x 2 (Consen- magnified when consumers share information in public versus in pri- sus: High vs. Low) mixed design, participants (N=152; undergradu- vate, implicating impression management concerns specifically. ates) were randomly assigned to one of the self-presentation goal Pennycook et al . directly examine the relationship between conditions. They rated four news headlines that were combined with accuracy judgments and sharing intentions. They propose that most either a high (e.g., New York Time) vs low (e.g., Denver Guardian) people do not want to spread misinformation but are distracted from credibility source. The news had ostensibly been shared by 790,000

Advances in Consumer Research 863 Volume 48, ©2020 864 / Antecedents of and Remedies to the Spread of False Information in the Social Media Era vs. 79 people, resulting in high vs. low perceived social consensus. for a relationship between uniqueness need and interest value of the We assessed participants’ willingness to re-share the information on content. their social network and its perceived truthfulness. Fifty-one partici- In combination, our findings highlight that consumers’ self- pants were removed; 3 of them completed less than 10% of the sur- presentation motives play a key role in their re-sharing decisions in vey and the remainder failed the attention check. online networks. People with high need for uniqueness may be less A repeated measures ANOVA indicated an interaction of goal, sensitive to credibility cues and share information coming from any credibility and consensus, F(2, 98)= 5.09, p = .01. As expected, par- source, which may expose them to spread of misinformation. Ongo- ticipants who had the goal to fit in reported higher willingness to ing studies examine which aspects of information consumers attend share under high-credibility/high-consensus conditions (M=6.70, to when they want to stand-out. Given the spread of misinformation SE=1.95, N=21) than under low-credibility/low-consensus condi- stemming from less-known and less credible sources, it is important tions (M=4.73, SE=2.44, N=21); mixed conditions (low-credibility/ to understand the interplay of consumer motivation and news char- high-consensus or high-credibility/low-consensus) fell in between acteristics. (M=5.76 SE=1.83, N=21). In stark contrast, neither source credibility nor consensus information had a significant influence when partic- Impression Management in the Echo-Chamber: How ipants had the goal to stand out (high/high, M=5.35, SD=2.33, N=27; Self-Censorship Biases Evidence-Sharing low/low, M=5.12, SD=2.38, N=27; mixed, M=5.30, SD=2.10, N=27). These differences in reported willingness to share could not be traced EXTENDED ABSTRACT to differences in truth perception. Independent of self-presentation The free exchange of ideas on the internet has led to the forma- goal, participants found high-credibility/high-consensus news more tion of ideological echo chambers, in which attitude-congruent infor- truthful (M=5.46, SD=1.18, N=98) than low-credibility/low-consen- mation circulates while attitude-incongruent information is filtered sus news (M=4.20, SD=1.34, N=98), with the mixed conditions fall- out (Del Vicario et al. 2016). Indeed, people preferentially seek and ing in between (M=4.95, SD=1.04, N=98), F(1,118)=69.65, p<.001. consume information that aligns with their pre-existing moral and In sum, consumers who have the goal to fit in seem to attend political attitudes about important issues like gun control, climate to the likely veracity of information before they share it, which is change, or immigration. In explaining these dynamics, past research not the case for consumers who want to stand out. Study 2 provides has highlighted the roles of prior attitudes and motivated reasoning, a conceptual replication of this finding and explores the role of the which together cause consumers to selectively attend to, believe, and interest value of the news (boring vs. interesting). remember attitude-congruent information and to avoid, discredit, Study 2 . Next, we tested (N = 340; undergraduates) whether our and forget attitude-incongruent information (Golman, Haggman, results hold when we measured people’s need for uniqueness using and Lowenstein 2017; Kahan 2012; Kunda 1990). Above and be- the scale from Snyder and Fromkin (1980). Since sharing news from yond motivated reasoning, we report a motivated sharing effect: low credibility is particularly problematic, we paired news head- Consumers prefer to omit attitude-incongruent, factual information lines (same with Study 1) with only low credibility sources (e.g., from conversations with others. Denver Guardian). We expected and found that as people’s need for Specifically, we propose that when sharing information with uniqueness increases, they become more likely to share news head- like-minded others, people will selectively omit relevant evidence lines coming from low credibility sources (ß=0.7, t=3.8, p<.001). As that might cast doubt on their loyalty to shared causes. For example, expected, need for uniqueness was not associated with truth judg- we predict that a liberal sharing results from a report on the welfare ment (t<1). Our correlational study provided consistent findings with effects of cutting taxes will strategically omit findings which speak study 1 when people’s uniqueness needs were measured. to the effectiveness of stereotypically conservative tax policies. By Study 3 . We wondered whether people with uniqueness need contrast, a conservative will omit from the same report findings that would take interest value of the content into account more so than speak to the effectiveness of stereotypically liberal tax policies. Al- other people. Following a 3 (Self-Presentation Goal: Fit-in vs. Stand- though both sides purport to be concerned with the actual effects out vs. Control) x 2 (Credibility: High vs. Low) x 2 (Content: In- of cutting taxes, individual actors on both sides are also concerned teresting vs. Boring) between-subjects (N=170; undergraduates) about appearing appropriately committed to their respective political design, participants were randomly assigned to one of the self-pre- movements. sentation goal conditions. They rated four news headlines that were This phenomenon – a bias towards sharing attitude-congruent either interesting (pretested, M=5.14, SD=1.55) or boring (M=3.65, vs. attitude-incongruent facts – represents a form of self-censorship, SD=1.86) and combined with either a high (e.g., New York Time) or rooted in concerns about impression management (Schlenker 1980). low (e.g., Denver Guardian) credibility source. Critically, this bias exerts influence above and beyond processes of Replicating the pattern of Study 1, willingness to share depend- motivated reasoning: Even when information is believed to be faced on motivation and source credibility, F(2,165)=2.95, p=.06, for tual and relevant, worries about the inferences others might make if the interaction. Participants with a fit-in goal reported greater willing- one shares that information will lead consumers to censor it from ness to share when the news came from a high-credibility (M=4.06, conversation. SD=2.28, N=29) rather than low-credibility source (M=2.78, In, study 1 (pre-registered, n=300) participants read a real news SD=1.60, N=25), F(1,52)=5.50, p=.02, for the simple effect. In con- article which outlined facts both supporting and undermining the trast, source credibility did not influence willingness to share when idea that President Trump deserves credit for recent growth in the participants had a stand-out goal, F(1,54)=.05, p=.82; the latter US economy. After reading, participants were presented sequentially condition did not differ from the control condition, F(1,58)=1.52, with six specific facts taken from the passage. Three of these facts p=.22. In addition, all participants were more willing to share inter- supported Trump’s performance (e.g., ‘Early last week, the stock esting news (M=4.30, SD=2.14, N=75) than boring news (M=3.43, market hit a new all-time high, something that has happened mul- SD=2.09, N=90), F(1,168)=6.97, p=.009, independent of their self- tiple times under President Trump’); while the other three under- presentation goal, F(2,165)=.95, p=.39, for the interaction. With this mined Trump’s performance (e.g., ‘Job growth has been gradually study, we confirmed our main hypothesis but did not find support slowing since Donald Trump’s inauguration’). We classified each of Advances in Consumer Research (Volume 48) / 865 these items as attitude-congruent or attitude-incongruent on the basis Understanding and Reducing the Spread of of participants’ answers to an initial question about their broader atti- Misinformation Online tudes towards Trump’s presidency. Participants were asked to imagine having a conversation with a friend or peer who shared their gen- EXTENDED ABSTRACT eral views about the president and to report how likely they would be The spread of misinformation – including blatantly false po- to bring up each fact in conversation (7-point likelihood scale). They litical “fake news”’ – on social media has become a major focus of also rated the extent to which they believed each fact to be true/accu- public debate and academic study in recent years [1]. Here we inves- rate and the extent to which they believed each to be relevant to un- tigate this willingness to share seemingly unbelievable content, and derstanding Trump’s job performance overall. Controlling for these ask what it can tell us about how to design interventions to reduce the variables in our analyses allows us to isolate self-censorship from spread of misinformation. other motivated reasoning processes. In line with our predictions, In a series of survey experiments, we provide evidence that the we found that, controlling for their perceptions of accuracy and rel- problem is not that people simply cannot tell what news is true ver- evance, participants were much less likely to share attitude-incon- sus false, or that people can tell but simply do not care. Instead, it gruent (M=3.41, 95% CI = [3.26, 3.55]) vs. attitude-congruent facts seems that the sharing of misinformation is rooted in a distraction- (M=4.79, 95% CI = [4.65, 4.95]; t(1557.7)=16.5, p<.001). This ef- based failure to consider the accuracy of claims. As a result, nudging fect held for both Trump-supporters and Trump-detractors (ps<.001). participants to think about the concept of accuracy by asking them Study 2 (preregistered, n=519) aimed to replicate these results to rate the accuracy of a single non-political headline substantially in a new context and to tie them specifically to impression manage- reduces their self-reported willingness to consider sharing false (but ment concerns. Study 2 followed a similar procedure to Study 1 with not true) headlines. the following differences. In Study 2, participants read a two-sided To corroborate these findings in an ecologically valid domain, article about gun control and made judgments about six facts: three and evaluate the efficacy of potential social media interventions to supporting and three undermining the case for increased gun regula- reduce the spread of misinformation, we conduct a digital field ex- tion in the US. We again classified these facts as attitude-congru- periment on Twitter. ent or attitude-incongruent. This time, participants were asked to Specifically, we sent users direct messages from a set ofbot imagine summarizing what they had read either for their own fu- accounts asking them to rate the accuracy of a single non-political ture use (private condition) or to share with and inform a political headline. Our subject pool consisted of N=5,379 Twitter users who ally (public condition). Next, they reported how likely they would had previously (re)tweeted links to Breitbart or Infowars, two lead- be to include each fact in their summary (7-point likelihood scale). ing misinformation sites. To allow for causal inference, we used a We again found that, above and beyond perceptions of accuracy and stepped-wedge (randomized roll-out) design in which users were relevance, participants indicated a lower likelihood of including randomly assigned to a date on which to receive the treatment mes- attitude-incongruent facts (M=4.41, 95% CI = [4.33, 4.49] in their sage. Within each 24-hour time-window, we then compared the links summaries than attitude-congruent ones (M=5.14, 95% CI =[5.06, shared by users who received the treatment message at the beginning 5.22]; t(2738.1)=13.3, p<.001). Importantly, we also detected an in- of that time window to the links shared by all the users who had not teraction, whereby reluctance to include attitude-incongruent facts yet been messaged (who thereby represented the control condition). was significantly stronger for a summary that would be made public We then combined estimates across dates to arrive at an overall treat- than one which would be kept private (t(2590.7)=5.1, p<.001). That ment effect (This randomized roll-out approach is required, instead the attitude-incongruence effect was larger in public than in private of a more traditional approach where half of the subjects receive implicates social-signaling concerns specifically, above and beyond treatment at the same time while the other half act as the control, self-signaling or mere dislike of dealing with attitude-incongruent because the rate limits imposed by Twitter forced us to only send a information per se. Again, all results held for both pro-gun and anti- small number of messages per account per day). gun participants (ps<.001). To quantify the quality of the news shared in any given tweet, Consumers typically expect those who share their moral at- we used a previously published list of 60 news websites whose truth- titudes to be trustworthy sources of information (Haidt 2012). By worthiness was rated by professional fact-checkers [2]. contrast, we predict and find that political allies provide one another As predicted, we find that the intervention leads to a 1.9% in- with biased information, strategically omitting factual and relevant crease (p=.009) in the average quality of new sites shared (5% in- evidence if it casts doubt on their commitment to a shared cause. crease when estimating the treatment effect just on users who ac- This self-censorship effect can explain, at least in part, the preva- tually tweet - i.e. excluding “never taker” user-days). Given the lence of political echo-chambers among groups of like-minded con- complexity of the experimental design and tweet data, there are a sumers: People prefer not to share facts that challenge their image as multitude of reasonable approaches for assessing whether our inter- loyal group members. vention successfully increased the quality of news sharing. Thus we More broadly, our results suggest that consumers hoping to computed effect size estimates using 98 different analysis approach- gather all of the facts on complex and consequential issues must con- es. Considering the analyses in aggregate provides strong evidence tend not only with the presence of irrelevant and factually inaccurate that, indeed, the accuracy message significantly increased the aver- information in their environment (i.e., fake news; Bago, Rand, and age quality of news sources subsequently shared by the users in our Pennycook 2020) but also with the absence of relevant, factual in- experiment. For the large majority of analytic approaches, the information, particularly when sourcing evidence from political allies. crease is statistically significant (p<0.05 in over 80% specifications). Finally, a domain-level analysis provides a more detailed picture of the effect of the intervention. Our results highlight an overlooked avenue by which social media fosters the spread of misinformation. Rather than (or in addition to) the often-discussed phenomenon of echo chambers and filter-bubbles, social media platforms may actually discourage peo- 866 / Antecedents of and Remedies to the Spread of False Information in the Social Media Era ple from reflecting on accuracy. These platforms are designed to en- publishers. Readers and experts separately rated each publisher on courage users to rapidly scroll and spontaneously engage with their its political ideology biases on a 5-point scale from -2 to 2 (-2 = newsfeeds, and mix serious news content with emotionally engaging “left”, -1 = “left-center”, 0 = “least biased”, 1 = “right-center”, 2 = content where accuracy is not a relevant feature (e.g., photos of ba- “right”). We compared the mean of reader voting poll with expert bies, videos of cats knocking things off tables for no good reason). ratings on media political ideology biases. Results revealed that gen- Social media platforms also provide immediate quantified social eral readers rated the media more left biased than the experts did, feedback (e.g., number of likes, shares, etc.) on users’ posts and are a t(99) = 3.30, p = .001, d = 0.35. The results also yielded a significant space which users come to relax rather than engage in critical think- effect of the variance of reader’s rating on the reader-expert differing. These factors imply that social media platforms may, by design, ence such that the larger the variance among reader’s rating, the big- tilt users away from considering accuracy when making sharing de- ger the discrepancy between reader’s rating and expert’s rating (b cisions. But this need not be the case. Our treatment translates eas- = 0.34, Robust SE = 0.13, t (97) = 2.62, p = .010). In other words, ily into interventions that social media platforms could employ to readers’ ratings of media appear to be affected by their prior beliefs increase users’ focus on accuracy. For example, platforms could pe- and are not consistent with expert opinion. riodically ask users to rate the accuracy of randomly selected head- Study 2 examines our hypothesis that similarity judgment lines (e.g. “to help inform algorithms”) - thus reminding them about questions can be effective in judging accuracy of news articles. Par- accuracy in a subtle way that should avoid reactance. The platforms ticipants were instructed to read three news articles that either sup- also have the resources to optimize the presentation and details of ported the idea of climate change or disagreed with climate change the messaging, likely leading to effect sizes much larger than what arguments. Half of the participants were asked to make evaluation we observed here. Such an approach could potentially reduce the judgments (e.g., how trustworthy the argument was in each article); amount of misinformation circulating online without relying on a the other half of the participants were asked to make similarity judg- centralized institution to certify truth and censor falsehood. ments (e.g., how similar are the arguments between two articles). Results revealed that responses to evaluation questions are largely Factchecking Matters: The Value of Crowdsourcing for affected by readers’ political ideology with correlations varying from Enhanced Accuracy Judgments 0.73 to 0.83. For example, more liberal participants were more likely to rat an article that is against climate change false, b = -.37, p = .002; EXTENDED ABSTRACT but rate an article that shows evidences of climate change true, b = Misleading information on news media is of key interest to in- .45, p < .001. Importantly, similarity judgment ratings were not af- dividual readers, publication managers, and policy makers. In recent fected by political ideology, ps > .88. That is, both conservative and years, both mass media and social media are rated as more biased liberal participants rated the similarities between articles in the same than ever before (Pew Research 2018; Garimella and Weber 2017). way regardless of their political ideology. The relationship between In fact, online news media were found the most politically polarized question type and participant’s cognitive ability showed different in the U.S. compared to other countries (Reuters 2017). patterns. Although political ideology does not affect the Cognitive However, most research and industry reports have relied on in- Reflection Task (CRT) score (Thomson and Oppenheimer 2016), b dividual readers’ self-reported beliefs and opinions to measure media = -.06, p = .34, participants with higher CRT score were better able trustworthiness (Pennycook and Rand 2019; Reuters 2017). For in- to distinguish articles with opposing arguments, b = -.56, p = .002. stance, readers are usually asked to answer questions such as “Does However, the CRT score does not affect the relationship between [certain news media company] provide accurate and reliable news?” political ideology and response to evaluation questions. Study 3 rep- or “How trustworthy is [certain news media company]?” There are licated the findings in Study 2 with a similar experimental design and two critical issues with this approach which we highlight in our first another set of news articles. study. First, the target of these questions is news media (rather than In the last ongoing study, we are incorporating expert ratings, the article content) which is problematic because the names of news reader ratings, semantic markers, and the similarity matrix generated media can lead to biased responses based on prior beliefs (Knight from text analysis tools to build a tool that can help make accurate Foundation and Gallup 2017). As suggested by research on moti- accuracy judgments. We also examine the extent to which crowd- vated reasoning (Kunda 1990), the conclusions drawn from readers’ sourced similarity judgments can add over and above machine learn- belief measures are highly driven by readers’ ideology and largely ing tools to judgments of article accuracy. This project utilized tools affected by which readers are sampled (Mercier and Sperber 2011; that span disciplines and is timely, relevant, and critical. Kahan 2012; Van Bavel and Pereira 2018). So far, neither researchers nor industry practitioners have found a reliable measure of article REFERENCES accuracy. In this research, we propose a new approach to utilizing Bago, Bence, David G. Rand, and Gordon Pennycook (2020), crowdsourcing to uncover misinformation on news media at the “Fake news, fast and slow: Deliberation reduces belief in level of individual articles. false (but not true) news headlines,” Journal of Experimental We propose a form of cognition-based judgment, similar- Psychology: General. ity judgment, which asks readers to compare how the opinions and Barasch, Alixandra, and Jonah Berger (2014), “Broadcasting and viewpoints of two articles are different from or similar to each other. Narrowcasting: How Audience Size Affects What People Because our proposed similarity judgment is constructed rather than Share.” Journal of Marketing Research, 51 (3), 286-299. retrieved (Tversky 1977), such judgment is less likely to be biased Carroll, J. D., and Green, P. E. (1997). Psychometric methods in by readers’ ideology and more likely to reach high consensus among marketing research: Part II, multidimensional scaling. Journal different reader populations, regardless of their prior beliefs or mo- of Marketing research, 34(2), 193-204. tivations. Chevalier, Judith A., and Dina Mayzlin (2006), “The Effect of Study 1 examines the problems with using evaluation ques- Word of Mouth on Sales: Online Book Reviews.” Journal of tions to rate accuracy at the media outlet level. This study involved Marketing Research, 43 (3), 345-354. 611,188 online reader votes on 100 mainstream English news media Advances in Consumer Research (Volume 48) / 867

David MJ Lazer et al. “The science of fake news”. In: Science Pennycook, G., and Rand, D. G. (2019). Fighting misinformation 359.6380 (2018), pp. 1094– 1096. on social media using crowdsourced judgments of news source Del Vicario, Michela, et al. “The spreading of misinformation quality. Proceedings of the National Academy of Sciences, online.” Proceedings of the National Academy of Sciences 116(7), 2521-2526. 113.3 (2016): 554-559. Pew Research Center. (2018). Publics globally want unbiased Gabielkov, Maksym, Arthi Ramachandran, Augustin Chaintreau, news coverage but are divided on whether their news media and Arnaud Legout (2016), “Social Clicks: What and Who deliver. Retrieved from https://www.pewresearch.org/ Gets Read on Twitter?,” ACM SIGMETRICS Performance global/2018/01/11/publics-globally-want-unbiased-news- Evaluation Review, 44 (1), 179-192. coverage-but-are-divided-on-whether-their-news-media- Garimella, V. R. K., and Weber, I. (2017, May). A long-term deliver/ analysis of polarization on Twitter. In Eleventh International Reuters. (2017). Reuters institute digital news report 2017. AAAI Conference on Web and Social Media. Retrieved from https://reutersinstitute.politics.ox.ac.uk/sites/ Golman, Russell, David Hagmann, and George Loewenstein default/files/Digital%20News%20Report%202017%20web_0. (2017), “Information Avoidance,” Journal of Economic pdf Literature, 55(1), 96-135. Schlenker, Barry R. (1980), Impression management, Brooks/Cole Gordon Pennycook and David G Rand. “Fighting misinformation Publishing Company, Monterey, CA. on social media using crowdsourced judgments of news source Trusov, Michael, Randolph E. Bucklin, and Koen Pauwels (2009), quality”. In: Proceedings of the National Academy of Sciences “Effects of Word-of-Mouth versus Traditional Marketing: 116.7 (2019), pp. 2521–2526. Findings from an Internet Social Networking Site.” Journal of Green, P. E., Wind, Y., and Carroll, J. D. (1973). Multiattribute Marketing, 73 (5), 90-102. decisions in marketing: A measurement approach (Vol. 973). Tversky, A. (1977). Features of similarity. Psychological review, Hinsdale, IL: Dryden Press. 84(4), 327. Haidt, Jonathan (2012), The Righteous Mind: Why Good People are Thomson, K. S., and Oppenheimer, D. M. (2016). Investigating an Divided by Politics and Religion. Vintage books. alternate form of the cognitive reflection test.Judgment and Kahan, Daniel M. (2012), “Ideology, motivated reasoning, and Decision making, 11(1), 99. cognitive reflection: An experimental study,” Judgment and Van Bavel, J. J., and Pereira, A. (2018). The partisan brain: An Decision-Making, 8, 407-424. Identity-based model of political belief. Trends in cognitive Kunda, Ziva (1990), “The case for motivated reasoning,” sciences, 22(3), 213-224. Psychological Bulletin, 108(3), 480. Vosoughi, Soroush, Deb Roy, and Sinan Aral (2018). “The spread Mercier, H., and Sperber, D. (2011). Why do humans reason? of true and false news online.” Science 359 (63801), 146-1151. Arguments for an argumentative theory. Behavioral and brain sciences, 34(2), 57-74. Netzer, O., Feldman, R., Goldenberg, J., and Fresko, M. (2012). Mine your own business: Market-structure surveillance through text mining. Marketing Science, 31(3), 521-543. The Sense of ‘MINENESS’: Exploring the Role of Ownership in New Consumer Domains Chair: Ipek Demirdag, University of California, Los Angeles, USA

Paper #1: Creating Ownership Where Ownership Does Not ership status. They find that the owners’ feeling of being distanced Exist: Psychological Ownership Increases Enjoyment in Sharing from the brand decrease the owners’ preference for the brand. Never- Economy theless, there is a positive effect of upgraded product for non-owners Ipek Demirdag, University of California, Los Angeles, USA because without an anchor for self-referencing, non-owners are only Suzanne Shu, University of California, Los Angeles, USA positively influenced by the brand’s competence signaled by its advanced products. Paper #2: Virtual Touch Facilitates Psychological Ownership of The final paper (Christensen and Shu) explores ownership felt Products in Virtual Reality towards extraordinary goods where ownership is collective, extend- Andrea Webb Luangrath, University of Iowa, USA ing the self by connecting owners to history and important traditions. Joann Peck, University of Wisconsin – Madison, USA They establish a gap between a seller’s willingness to accept from William Hedgcock, University of Minnesota, USA family versus from strangers, which is greater for inherited goods Yixiang Xu, University of California, Berkeley, USA than for purchased goods. In addition, the paper identifies heritage Paper #3: “My Brand is Moving Away from Me”: When value, that is, the degree to which an object extends the self by con- Upgrade Products Leave Consumers Behind necting the owner to history and traditions. They demonstrate that Wonsuk Jung, University of Wisconsin-Madison, USA this heritage value predicts the gap in willingness to accept from Joann Peck, University of Wisconsin-Madison, USA family versus strangers. Mauricio Palmeira, University of South Florida, USA Ownership is a key corollary of consumption, and this session Kyeongheui Kim, Sungkyunkwan University, South Korea spotlights its extensive applicability to consumer behavior. The di- Paper #4: Heritage Loss: How Connection to the Past Shapes verse approaches taken by each paper and the varying aspects of Consumer Valuation in the Present ownership investigated will attract a broad range of ACR members Kate Christensen, University of California, Los Angeles, USA interested in ownership, endowment effect, sharing economy, virtual Suzanne Shu, University of California, Los Angeles, USA reality experience.

SESSION OVERVIEW Creating Ownership Where Ownership Does Not Exist: Buying and owning material and experiential products is in- Psychological Ownership Increases Enjoyment in dispensable for consumer judgment and decision making. Notwith- Sharing Economy standing that initial research on ownership has concentrated on legal ownership’s effect on valuation (e.g., Kahneman, Knetsch, and EXTENDED ABSTRACT Thaler, 1990), recent research on psychological ownership has ex- Since the 1950s, hyper-consumption, and ownership has defined amined the effects of regarding something as “mine” in wide variety economic security. A 1955 LIFE magazine article celebrated the idea of areas from digital technology to financial decision making (Peck of throwing away items after single-use and advertised several of & Shu, 2018; Pierce, Kostova, & Dirks 2001). The present session these items, e.g., single-use barbeque grill, disposa-pan that “elimi- gives prominence to the diverse impacts of ownership on consumer nates scouring of pots after cooking”, and feeding bowl for pets that behavior. ends the “washing-up chore” (“Throwaway Living,” 1955). Nowa- The emergence of the sharing economy motivates the first pa- days, we increasingly have a “reduce, reuse, and recycle” mentality. per (Demirdag and Shu) to consider the psychological ownership This notable shift away from the “own and dispose” society has laid of these new shared goods and services. The authors document that foundations for the sharing economy. Consumers have come to the sharing economy products like Uber rides or Airbnb stays produced a realization that legal ownership is not mandatory for consumption. lower sense of psychological ownership than legal car or home own- They can access goods and services they cannot afford to purchase ership. By separately manipulating psychological ownership through or simply do not own at the time of the consumption decision. To an intimate knowledge of the target and variety, the authors increase increasing extent, value relies on technology and consumer experi- satisfaction with these shared services. Moreover, they demonstrate ence, rather than ownership. that collective (vs. individual) experiences reduced tip and satisfac- We examine a facet of sharing economy that has not been stud- tion through reduced psychological ownership, suggesting that expe- ied yet: psychological ownership, which is the sense that something riences in certain sharing economies are inherently solitary. is “mine” (Pierce, Kostova, & Dirks, 2001). The sharing economy The second paper (Luangrath, Peck, Hedgcock, and Xu) ex- presents novel ways of accessing goods and services without legal amines the effect of vicarious touch. The authors study whether ownership. We suggest that consumption in this domain reduces psy- observing someone else touch a product increase the psychological chological ownership compared to consumption based around legal ownership of a virtual limb and, hence, increase the psychological ownership. This is problematic, as it can lead to the tragedy of the ownership of products. They find evidence for this effect, labeled commons. For instance, shared scooters are vandalized, tossed off the vicarious haptic effect. They find that vicarious touch leads to the buildings, set on fire, and thrown into the ocean (Newberry, 2018). psychological ownership of a product via a sense of body ownership Our studies seek to increase psychological ownership, which could of a virtual limb. This mediation holds under egocentric (from the prevent these types of disastrous behavior. perspective of oneself) but not allocentric (from the perspective of In Study 1 (N = 321), we found that ownership of a hypotheti- others) orientation. cal Uber ride is significantly correlated with the consumers’ willing- The third paper (Jung, Peck, Palmeira, and Kim) investigates ness to pay for the ride (r = .14, p < .05), tip for the Uber driver (r an effect of introductions of upgraded products on consumers’ brand = .11, p < .05), and ride satisfaction (r = .19, p < .01). Furthermore, preference and examine how the effect may differ depending on own- through planned comparisons we showed that real-life car owners

Advances in Consumer Research 868 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 869 have a greater sense of ride ownership than real-life car leasers (F (1, sumers lack the incentive to take care of the goods they access in the 605) = 169.2, p < .001) and that leasers have a greater sense of ride sharing economy, as they do not have to think about the long-term ownership than Uber riders (F(1, 605) = 166.5, p < .001). Similarly, impact of their actions during the short period of time they consume we found that real-life homeowners have a greater psychological these goods. An enhanced sense of psychological ownership could ownership of their houses than renters (F (1, 450) = 257.7, p < .001) prevent these ruinous consequences and increase satisfaction with and that renters have a greater psychological ownership than AirB- the experiences a sharing economy platform provides. nBers (F (1, 450) = 20.04, p < .001). In sum, Study 1 demonstrated that psychological ownership of goods and experiences is lost in the Virtual Touch Facilitates Psychological Ownership of sharing economies, at least in ride-sharing and space-sharing ser- Products in Virtual Reality vices, but that higher psychological ownership in sharing economies indicate greater satisfaction with experiences in sharing economies. EXTENDED ABSTRACT Study 2 (N = 132) was a field study in the domain of a scooter- Online product videos and virtual reality (VR) interfaces en- sharing company, Bird. We found that psychological ownership of liven consumer experiences and are a source of product information. a scooter ride was correlated positively with ride satisfaction (r = Intriguingly, these virtual environments often lack a semblance of .34, p < .01), likeliness of future rides (r = .40, p < .01), telling oth- “self” in the experience, oftentimes using a cursor rather than a vir- ers about Bird (r = .27, p < .01), and general trust towards Bird (r = tual hand. In this research, we investigate vicarious touch, conceptu- .45, p < .01). Although these results are correlational, our field study alized as the observation of a virtual hand using a haptic exploratory lays the groundwork for field interventions that augment psychologi- procedure with a product in an online or virtual environment. Does cal ownership, with the goal of increasing enjoyment derived from the vicarious experience of virtual touch affect consumers’ percep- goods and services in the sharing economy. tions of psychological ownership? That is, does observing someone In Study 3 (N = 641), we differential effect of collective (vs. else touch make a consumer feel as if they, themselves, are touch- individual) experiences. We found that sharing an Uber ride with ing the product? Previous research demonstrates that people can feel a friend led to lower ride ownership (F(1, 639) = 4.591, p = 0.03). transported to a virtual space by blurring the lines of what is consid- Moreover, a mediation analysis with 5,000 bootstrapped samples ered one’s own body (Slater et al. 2008). A sense of body ownership showed an indirect-only mediation for the effect of collective (vs. is the feeling that the body that I inhabit is ‘my own’ and a part of individual) experience on tip through ride ownership, indirect effect ‘me’ (Tsakiris 2010). Psychological ownership, or the feeling that = -.016, 95% CI: [-.035, -.002]. A similar process was revealed for something is ‘MINE!’ (Pierce, Kostova and Dirks 2001), is often the effect of collective (vs. individual) experience on satisfaction associated with greater product valuation (Peck, Barger, and Webb through another mediation analysis with 5,000 bootstrapped sam- 2013; Peck and Shu 2009). We conjecture that vicarious touch in- ples, indirect effect = -.023, 95% CI: [-.049, -.004]. These results increases body ownership of a virtual limb, which then increases psy- dicate that sharing an Uber ride with a friend reduces psychological chological ownership felt of a product, termed the vicarious haptic ownership, and therefore, enjoyment, suggesting that experiences in effect. ride-sharing are inherently solitary. Study 1 tests the main proposition that body ownership drives In Study 4 (N = 645), we aimed to increase psychological own- the effect of vicarious touch on psychological ownership. - Partici ership. One of the routes of psychological ownership is intimate pants (N=208) were either shown a video of a hand pointing toward knowledge of the target (Pierce, Kostova, & Dirks, 2001). We found a wristwatch or picking up and hefting the wristwatch. Psychological that providing information about the Uber driver (such as their rat- ownership (e.g., “I feel a personal ownership of the watch”) and body ing, hometown, compliments they have received, languages they ownership (e.g., “While viewing the video: ‘I felt as if the hand in speak, number of years they have been an Uber driver, and the num- the video was my hand”) were measured. An ANOVA reveals a sig- ber of trips they have had) increased tip amount (F(1, 638) = 3.96, p nificant effect of vicarious touch on body ownership (MNoTouch=1.78, 2 = 0.05), WTP (F(1, 564) = 5.34, p = 0.02), satisfaction (F(1, 643) = MTouch=2.84, F(1,194)=31.96, p<.001, η =.14) and psychologi-

5.26, p = .02). Furthermore, this effect was mediated by ride owner- cal ownership (MNoTouch=1.89, MTouch=2.23, F(1,194)=3.86, p=.05, ship. A mediation analysis with 5,000 bootstrapped samples uncov- η2=.02). A mediational model demonstrates that body ownership ered an indirect effect of information on tip through ride ownership, mediates the effect of vicarious touch on psychological ownership indirect effect = .018, 95% CI: [.004, .037]. Similarly, another me- (.3670; CI95%=[.2326, .5109]). These results reveal the vicarious hap- diation analysis with 5,000 bootstrapped samples demonstrated an tic effect. indirect effect of information on satisfaction through ride ownership, Given that vicarious touch facilitates a sense of body ownership indirect effect = .017, 95% CI: [.004, .042]. Thus, information about for the displayed arms, we expect to attenuate this effect by altering the driver increased ride enjoyment (specificallytip and satisfaction) the orientation of touch. Authorship processing suggests we leverage through ride ownership. cues from the environment to discern whether an event has occurred In Study 5 (N = 640), we increased satisfaction through en- as a result of oneself (Wegner and Sparrow 2004). Study 2 was a hanced psychological ownership by bringing forth variety where the 2 (no touch/touch) x 2 (egocentric/allocentric orientation). MTurk goods are typically identical (e.g., CitiBikes, Lime Bikes, Bird scoot- participants (N=209) view a video displaying a hand either resting on ers). In a bike-sharing scenario, variety increased the ride enjoyment the table or feeling the texture of the product. This was shown either (indirect effect = .055, 95% CI = [.017, .103]) and willingness to ride from the perspective of the self or from the other. Results reveal a in the future (indirect effect = .083, 95% CI = [.024, .149]) through significant interaction between touch and orientation on psychologi- enhanced psychological ownership of the ride. cal ownership (F(1,185)=6.43, p=.012, η2=.03) such that perceptions In conclusion, psychological ownership of experiences is lower of psychological ownership were significantly higher when viewing in sharing economies than it is in legal-ownership based consump- touch from an egocentric (vs. allocentric) orientation (MEgoTouch=3.13, 2 tion. Through our lab and field studies, we demonstrated that the MAlloTouch=2.32, F(1,185)=4.28, p=.04, η =.02) while there was no higher the psychological ownership, the higher the enjoyment of the significant difference in orientation with no touch (MEgoNoTouch =2.50, 2 experiences. The implications of this research are important. Con- MAlloNoTouch =3.08, F(1,185)=2.29, p=.13, η =.01). A significant mod- 870 / The Sense of ‘MINENESS’: Exploring the Role of Ownership in New Consumer Domains erated mediational model indicates that body ownership mediates success. Despite the increasing prevalence of upgrades, however, the relationship between vicarious touch and psychological owner- only limited research has investigated its implication for consumers’ ship under conditions of egocentric orientation but not allocentric attitudes toward upgrades. orientation (-.5205; CI95%=[-1.0193, -.0310]). In this research, we aim to broaden an understanding of the In study 3, we test whether the vicarious haptic effect holds in emerging but limited body of research on product upgrades. Specifi- virtual reality. Further we examine if having one’s own hands occu- cally, we investigate an effect of introductions of upgraded products pied (i.e., holding onto balls) interferes with feelings of body owner- on consumers’ brand preference and examine how the effect may ship. We created a 360° virtual reality retail store. The design was differ depending on ownership status. We propose that the release of a 2 (vicarious touch/cursor) x 2 (hands occupied/hands free). Un- an upgraded product has distinct psychological effects on consum- dergraduates (N=255) viewed a VR retail store on Oculus headsets. ers who are current owners of the brand, and that this psychological Participants responded to a variety of additional measures includ- process would subsequently influence how much they prefer their ing willingness to pay, product evaluation, and purchase intention. current brand. Contrary to common intuition, we argue that the in- Results of an ANOVA reveal a significant effect of vicarious touch troduction of an upgrade could have an undesirable effect on cur- on body ownership (F(1,241)=20.31, p<.001, η2=.08). Holding onto rent owners (but not for non-owners). We propose that since owners something (thereby having one’s hands occupied) did not dampen perceive the product they currently own as a representation of “me” the effect of vicarious touch on body ownership (F(1,239)=.346, (Weiss and Johar 2013, 2016) and perceive the new upgraded prod- p=.56). Body ownership mediates the effect of vicarious touch uct as an embodiment of “the brand” (John, Loken, and Joiner 1998), on psychological ownership (.1642; CI95%=[.0855, .2590]), pur- the upgrade product would make the owners feel as if the brand has chase intention (.1320; CI95%=[.0596, .2239]), product evaluation moved away from them just as the new product has advanced away

(.1093; CI95%=[.0508, .1828]), as well as willingness to pay (.3320; from its previous versions. Feeling distant from a brand results in

CI95%=[.1164, .6245]). a less positive consumer-brand relationship, which in turn decrease A human need for stimulation or activation is one of the fun- brand commitment (Park, Eisingerich, and Park 2013). In line with damental motives that is considered the reason for the existence of the findings, we propose that the owners’ feeling of being distanced the psychological state of ownership (Pierce and Jussila 2011; Pierce from the brand would decrease the owners’ preference for the brand. and Peck 2018). Heart rate is one physiological manifestation of this For non-owners, however, we expect to observe a positive effect of individual difference in need for stimulation (Mathias and Stanford upgraded product because without an anchor for self-referencing, 2003). We expect that those highly stimulated by VR would feel non-owners are only positively influenced by the brand’s -compe stronger effects of vicarious touch on psychological ownership. Un- tence signaled by its advanced products (Heath, DelVecchio, and dergraduates (N=144) viewed a VR retail store while having their McCarthy 2011). heart rate measured on a Biopac MP36. Study design was similar to Four studies support our predictions. Throughout studies, we study 3 with the addition of a no touch condition in which partici- measured participants’ purchase intention and their general attitudes pants saw a hand but it did not touch the product. Viewing the virtual toward the brand as dependent variables. For purchase intention, we touch, as opposed to no touch and cursor, enhanced feelings of body measured consumers’ willingness to choose the brands for purchas- ownership (F(2,141)=6.55, p=.002, η2=.09). Mediational analyses ing another product because upgraders are often hindered by mental replicate the effects from study 3. There is a significant interaction cost of retiring their current products that are still functional (Okada with vicarious touch and heart rate on psychological ownership (β 2006). =-.065, t(123) = -2.25, p=.026). For highly stimulated individuals Study 1 examined actual brand owners’ and non-owners’ re- (+1 sd heart rate change), the effect of vicarious touch on psycho- sponses to the release of upgraded products, and tests the underly- logical ownership is positive and significant (t(123)=-2.17, p=.03) ing process. To this end, we recruited 344 owners and non-owners while for those low in stimulation (-1 sd heart rate change) the rela- of Apple 10th generation iPhone and assessed their responses to the tionship between vicarious touch and psychological ownership is not presence of newer generations of iPhones. To manipulate the pres- significant (t(123)=-1.096, p=.28). ence of upgraded products, we manipulated the saliency of the up- We document the power of virtual touch. Our research demon- grades to make them look either more or less apparent for consum- strates that vicarious touch increases a sense of psychological owners. The owners and non-owners were randomly assigned to either a ership of a product, which occurs due to a sense of body ownership control or an upgrade condition and were asked to rate their purchase of the virtual hands. We find that not only should a virtual hand be intention (computer monitor), brand attitude, and perceived rela- present, but it should be engaging haptically with the product. We tional closeness toward Apple as a process measure using IOS scale further show that this effect is attenuated when observing touch from (Aron, Aron, and Smollan 1992). The result revealed a significant the perspective of another (i.e., allocentric orientation) as opposed 2 (upgrade) X 2 (ownership status) interaction (p < 0.01). Partici- to oneself (i.e., egocentric orientation). The effect is robust even if pants in the owner (non-owner) condition were less (more) willing the observer’s hand is otherwise occupied. Finally, using the physi- to choose Apple when purchasing another product if the upgraded ological measure of elevation of heart rate, we demonstrate that for products were made more salient compared to when they were made those individuals who become highly stimulated in VR, the effect of less salient. The results for brand attitude (α = 0.96) followed the vicarious touch on psychological ownership is strengthened. parallel pattern. Moderated mediation analysis suggested that the indirect effect of upgraded products through IOS varies as a func- “My Brand is Moving Away from Me”: tion of ownership status (p = 0.08); the conditional indirect effect of When Upgrade Products Leave Consumers Behind upgraded products was significant for owners (p < 0.05) but not for non-owners (p > 0.1). EXTENDED ABSTRACT Study 2 replicates the findings from Study 1 in a more con- In today’s advanced economies, continual product enhancement trolled setting with hypothetical scenarios to eliminate prior brand through product upgrades is deemed a key component of a firm’s knowledge. As in Study1, results on purchase intention (microwave) Advances in Consumer Research (Volume 48) / 871 revealed the anticipated 2 (upgrade) X 2 (ownership status) interac- guochao style in China, consumers have increasingly chosen to buy tion (p < 0.01), showing that owners (non-owners) were less (more) goods that connect them to a shared past (Achim 2019; Duveau and willing to choose the brand when the brand released upgraded prod- Dumenil 2018; Killam 2019; Luo 2019; Rapp 2019). ucts. The results for brand attitude (α = 0.94) followed the parallel In this project we explore how a connection to this shared past pattern. and the moderated mediation analysis replicated the finding affects consumer valuations in the present. This research on heritage from Study 1 (p < 0.05). connection builds on the sharing literature, which has introduced the Study 3 put the focus on owners and further examined the un- concept of “ours” and called for work on how the aggregate extended derlying process using a moderation-of-process approach by ma- self – “us” – affects consumer practices (Belk 2010, 2017; Curasi, nipulating the extent to which the focal brand is manipulated as Price, and Arnould 2004; Epp and Price 2008; Lastovicka and Fer- psychologically close or neutral to one’s self. Study 3 also tested nandez 2005). The current work also extends research on the en- an alternative explanation that upgrades makes owners unhappy be- dowment effect by exploring how collective ownership of the shared cause they see a frequent release of upgrades as an attempt by brands past can impact sellers’ WTA (Brough and Isaac 2012; Dommer and to entice consumers to unnecessarily spending. To test this account, Swaminathan 2013; Morewedge, Shu, Gilbert, and Wilson 2009; we incorporated an additional condition in which the interval be- Shu and Peck 2011; Weiss and Johar 2013)a phenomenon called the tween the new upgrades is manipulated to be longer. 362 students endowment effect. Loss aversion has typically accounted for the en- participated and were randomly assigned to one of six conditions in dowment effect, but an alternative explanation suggests that owner- a 2 (brand-self distance: neutral vs. close) X 3 (upgrade: no-upgrade ship creates an association between the item and the self, and this vs. upgrade vs. upgrade-after-three-years) between-subjects design. possession-self link increases the value of the good. To test the own- The result showed that the brand-self distance moderates the pre- ership account, this research examines three moderators that theory dicted pattern (2, -1, -1 contrast) across the three upgrade conditions suggests should affect the possession-self link and consequently the (p < .01). As predicted, when the brand-self distance was perceived endowment effect: self-threat, identity associations of a good, and close (neutral), participants were no less (less) likely to choose their gender. After a social self-threat, the endowment effect is strength- current brand in both the upgrade and the upgrade-after-three-years ened for in-group goods among both men and women but is elimi- condition than in the no-upgrade condition. nated for out-group goods among men (but not women. We predict Study 4 complements our experimental results with evidence that the same factor – heritage connection - increases valuations for from a real-world dataset containing vehicle ownership data of both sellers and buyers—but it has the opposite effect that an econo- 49,998 households across the United States. Using a multilevel mod- mist would predict on selling prices. The buyer’s heritage connection el incorporating 30 different automobile brands and 294 different car reduces the seller’s WTA. models nested within the brands, we find that as the number of newer We define heritage connection as a good’s capacity to bridge generations of car models that current households own increases in the gap between the consumer and their heritage by helping the con- the market, the households become more likely to switch their brand sumer remember where they come from, giving them a sense of his- rather than to stay loyal for their next car purchase. tory, and connecting them with the people who came before them. In We believe that our research may offer substantive implications four studies, we find a main effect such that sellers set lower prices in today’s ever innovating economies. We suggest that while manag- for heritage goods when selling to buyers with heritage connection, ers intend to leverage brand equity among consumers through prod- relative to buyers without heritage connection (i.e., the heritage dis- uct enhancement, the presence of new upgrades may in fact come count). with unintended consequences. Therefore, adequate strategy may be In Study 1, we sampled California residents on Mturk (N = needed that pays special attention to current consumers. 400). The study ran online during the 250th anniversary of Monterey, a city in California that is home to California’s first constitution, Heritage Lost: How Connection to the Past Shapes California’s first scenic highway, and is the first place where Cali- Consumer Valuation in the Present fornians raised the American flag. Monterey also has an aquarium, and visitors come to see it. We informed participants that one person EXTENDED ABSTRACT taking the study would receive a Monterey 250th anniversary hat, Heritage – a connection to a shared past – is a fundamental part and that it would be theirs to keep. We endowed participants with of human experience, and consumers’ interest in connecting to the the anniversary hat, and we measured participants actual WTA for people who came before them has been growing (Belk 1990; Csik- this heritage good using the Becker DeGroot Marschak test (Becker, szentmihalyi and Rochberg-Halton 1981)researchers, and students Degroot, and Marschak 1964) As predicted, sellers’ WTA from a discover, use, and build upon a wide range of content in a trusted control buyer (visitor to Monterey) was higher than their WTA from digital archive. We use information technology and tools to increase the heritage buyer (resident of Monterey), t(1,364) = 2.32, p = .021. productivity and facilitate new forms of scholarship. For more in- Sellers had a lower WTA for heritage buyers despite indicating that formation about JSTOR, please contact [email protected]. Psychol- heritage buyers would value the good more (p < .001). Further, the ogy and Things The Manchester seminars on Technology as Skilled effect of heritage held after adding controls for buyers’ usage (p = Practice brought together perspectives from a number of disciplines, .030). Thus, Study 1 provided preliminary evidence for a heritage including anthropology, history, and psychology. As an intentionally discount. peripheral psychologist, I thought twice about attending an interdis- Study 2 tested heritage in a new domain using real goods that ciplinary meeting willing to include psychologists. Academic psy- people have in their homes. Participants (N = 547) were randomly chology throughout its brief history has been remarkably anxious assigned to answer questions about either an inherited or a purchased about dem arcating its own exclusive subject — the \”subject\” — item in their home that would normally cost less than $1000. Par- and has perpetuated, even ins titutionalized, the traditional dualism ticipants indicated their WTA from both a close family member and within modern science of the objective and subjective. Psychology a stranger. As predicted, the simple effect of stranger vs. family for (along with the other human sciences. From the popularity of heri- heritage products was larger than the simple effect of stranger vs. tage travel on Airbnb to the trendiness of heritage fabrics in England, family for non-heritage products. The interaction of buyer’s identity 872 / The Sense of ‘MINENESS’: Exploring the Role of Ownership in New Consumer Domains and product type (heritage vs. no heritage) was significant, all p’s < Csikszentmihalyi, M., & Rochberg-Halton, E. (1981). The meaning .001, and sellers’ heritage connection to a good moderated the heri- of things. The meaning of things. Cambridge University Press. tage discount with more connected sellers giving larger discounts, https://doi.org/10.1017/cbo9781139167611 F(1, 545) = 46.20, p < .001. Specifically, we found that the greater Curasi, C. F., Price, L. L., & Arnould, E. J. (2004). How the seller’s heritage connection, the larger the heritage discount con- Individuals’ Cherished Possessions Become Families’ nected buyers received, B = 1.405, SE =.175, t(545) = 8.04, p < .001. Inalienable Wealth. Journal of Consumer Research, 31(3), At this point, it seemed like holding the item constant and ma- 609–622. nipulating heritage connection might provide an even cleaner test of Duveau, J., & Dumenil, S. (2018). Opinion: Heritage has Become our hypothesis, and we did this in Study 3. We asked participants (N Hip With Chinese Millennials | Jing Daily. Jing Daily. = 400) to imagine that they were going through storage and found a Retrieved from https://jingdaily.com/opinion-heritage-has- watch from their great grandfather. In a between subjects design, we become-hip-with-chinese-millennials/ asked them to consider selling this heritage good to a control buyer Epp, A. M., & Price, L. L. (2008). Family Identity: A Framework or a heritage buyer, and we specified usage: both buyers would use of Identity Interplay in Consumption Practices. Journal of the watch as they went about their daily work. Sellers set a lower Consumer Research, 35(1), 50–70. WTA for heritage buyers than control buyers, t(386) = 2.50, p = Heath, Timothy B., Devon DelVecchio, and Michael S. McCarthy .013. There was no difference in appropriateness of usage between (2011), “The asymmetric effects of extending brands to lower conditions, p = .714. We also measured seller’s heritage connection, and higher quality,” Journal of Marketing, 75(4), 3-20. buyer’s perceived heritage connection, and we calculated heritage John, Deborah Roedder, Barbara Loken, and Christopher Joiner loss (i.e., the difference between a seller’s heritage connection and a (1988), “The negative impact of extensions: can flagship buyer’s heritage connection). Overall, Study 3 provided strong evi- products be diluted?,” Journal of Marketing, 62(1), 19-32. dence that heritage loss mediates the effect of buyer’s identity on the Killam, A. (2019). Check-In to the Trend: Heritage Travel on the seller’s WTA (95%, CI = $127.25, $218.35). Next, we sought to vali- Rise. Retrieved November 30, 2019, from https://news.airbnb. date these findings in an incentive compatible context. com/check-in-to-the-trend-heritage-travel-on-the-rise/ In Study 4, we sampled the population on the campus of a large Lastovicka, J. L., & Fernandez, K. V. (2005). Three Paths to American research university (N = 388), and endowed participants Disposition: The Movement of Meaningful Possessions to with a good symbolizing 100 years of the university’s history. In a Strangers. Journal of Consumer Research, 31(4), 813–823. between subjects, fully incentive compatible task, participants indi- Dommer, S., & Swaminathan, V. (2013). Explaining the cated their WTA from a heritage buyer (student at their university) or Endowment Effect through Ownership: The Role of Identity, their WTA from a control buyer (stranger with no university connec- Gender, and Self-Threat. Journal of Consumer Research, 39. tion). As predicted, participants’ WTA from a control buyer ($2.65) Luo, J. (2019). Can ‘Made in China’ be cool? Yes, if the West was higher than their WTA from a heritage buyer ($2.20, F(1,385) thinks so | Goldthread. Retrieved December 10, 2019, from = 6.54, p = .011). Further, we found this effect was mediated by https://www.goldthread2.com/culture/made-in-china-guochao- heritage loss. hipster/article/3027942 In conclusion, this research suggests that the shared past af- Mathias, C. W. & Stanford, M. S. (2003). Impulsiveness and fects consumer valuations in the present. In four studies, we quantify arousal: Heart rate under conditions of rest and challenge in the value of this heritage connection, and these findings have sig- healthy males. Personality and Individual Differences, 35(2), nificance for marketing practice. Our research has applications to 355-371. marketers of products that appeal to consumers’ desire to connect to Morewedge, C. K., Shu, L. L., Gilbert, D. T., & Wilson, T. D. their heritage across time as well as markets that involve resale (e.g., (2009). Bad riddance or good rubbish? Ownership and not loss housing and collectibles). aversion causes the endowment effect. Newberry, L. (2018, August 10). Fed-up locals are setting electric REFERENCES scooters on fire and burying them at sea. Los Angeles Times. Achim, A.-L. (2019). Market Trends | Successful “Heritage” Retrieved from https://www.latimes.com/local/lanow/la-me- Marketing Resonates with Consumers in China. Retrieved ln-bird-scooter-vandalism-20180809-story.html December 8, 2019, from https://jingdaily.com/successful- Okada, Erica Mina. (2001), “Trade-ins, mental accounting, heritage-marketing-resonates-with-consumers-in-china/ and product replacement decisions,” Journal of Consumer Aron, Arthur, Elaine N. Aron, and Danny Smollan (1992), Research, 27(4), 433-446. “Inclusion of other in the self scale and the structure of ——— (2006), “Upgrades and new purchases,” Journal of interpersonal closeness,” Journal of personality and social Marketing, 70(4), 92-102. psychology, 63(4), 596. Park, C. Whan, Andreas B. Eisingerich, and Jason Whan Park Becker, G. M., Degroot, M. H., & Marschak, J. (1964). Measuring (2013), “Attachment–aversion (AA) model of customer–brand utility by a single-response sequential method. Behavioral relationships,” Journal of consumer psychology, 23(2), 229- Science, 9(3), 226–232. 248. Belk, R. (2010). Sharing. Journal of Consumer Research, 36(5), Peck, J., Barger V. A., & Webb A. (2013). In search of a surrogate 715–734. for touch: The effect of haptic imagery on perceived ———. (2017). Sharing without caring. Cambridge Journal of ownership. Journal of Consumer Psychology, 23 (2), 189-196. Regions, Economy and Society, 10(2), 249–261. https://doi. Peck, J. & Shu S. (2009). The effect of mere touch on perceived org/10.1093/cjres/rsw045 ownership. Journal of Consumer Research, 36 (3), 434-447. ———. (1990). The Role of Possessions in Constructing and Pierce, J. L., Kostova, T. & Dirks, K.T. (2001). Toward a theory Maintaining a Sense of Past. ACR North American Advances, of psychological ownership in organizations. Academy of NA-17. Management Review, 26 (2), 298-310. Advances in Consumer Research (Volume 48) / 873

Rapp, J. (2019). Heritage rediscovered. Retrieved from https:// Wegner, D. M., & Sparrow, B. (2004). Authorship processing. In www.jwtintelligence.com/2019/01/heritage-rediscovered/ M. S. Gazzaniga (Ed.), The cognitive neurosciences (pp. 1201- Shu, S. B., & Peck, J. (2011). Psychological ownership and 1209). Cambridge, MA, US: MIT Press. affective reaction: Emotional attachment process variables Weiss, L., & Johar, G. V. (2013). Egocentric Categorization and and the endowment effect.Journal of Consumer Psychology, Product Judgment: Seeing Your Traits in What You Own (and 21(4), 439–452. Their Opposite in What You Don’t). Journal of Consumer Slater, M, Marcos, D.P., Ehrsson, H., & Sanchez-Vives, M. V. Research, 40(1), 185–201. (2008). Towards a digital body: The virtual arm illusion. ——— (2016), “Products as self-evaluation standards: When Frontiers in Human Neuroscience, 2, 6. owned and unowned products have opposite effects on self- Strahilevitz, M. A., & Loewenstein, G. (1998). The Effect of judgment,” Journal of Consumer Research, 42(6), 915-930. Ownership History on the Valuation of Objects. Journal of Consumer Research, 25(3), 276–289. Tsakiris, Manos (2010). My body in the brain: A neurocognitive model of body-ownership. Neuropsychologia, 48 (3), 703-712. Monetizing the Megaphone: How Consumers and Firms Use Platforms to Conquer the Attention Economy Chair: Zeynep Arsel, Concordia University, USA Discussant: Ashlee Humphreys, Northwestern University, USA

Paper #1: Problematizing the Megaphone: The Very Difficult The final paper explores how platform content is shaped by Journey to Becoming an Influencer memes that are also controlled and seeded by firms. The authors Thanh Nguyen, Concordia University, Canada show how memes are actively initiated, surveilled, and leveraged by Pierre-Yann Dolbec, Concordia University, Canada firms. They show how companies use sponsored hashtags to entice consumers to engage with brand content through three types of call Paper #2: Aesthetics of Food: The Role of Visual Framing to action, where paid influencers take thelead, followed by everyday Strategies for Influence Building on Instagram (unpaid) consumers amplifying the message. Shuhan Yang, Rochester Institute of Technology, USA Together, the papers in this session tackle the following ques- Jonathan Schroeder, Rochester Institute of Technology, USA tions: Duygu Akdevelioglu, Rochester Institute of Technology, USA 1. How do consumers position themselves as influencers in Paper #3: Orchestrating Pet Influencers: Rhetorical and Visual the attention economy? Strategies in Creating Mediated Platform Content 2. How do consumers invest in and deploy resources to Ghalia Shamayleh, Concordia University, Canada strengthen their positioning? Zeynep Arsel, Concordia University, Canada 3. How do consumers use platforms’ expressive capacities to Paper #4: Marketing Memeification on TikTok: Initiating and create engagement? Leveraging Consumer Creativity for Commercial Means 4. How do consumers interact with firms to convert attention Jenna Drenten, Loyola University Chicago, USA to value and revenues? Pilar Castro González, Universidad Loyola Andalucía, Spain The session provides a meeting point for scholars using different methods (quantitative and qualitative) and complementary SESSION OVERVIEW perspectives to understand social media. It investigates the role of Platforms have allowed ordinary consumers to monetize their different actors that participate in the influencer business: pets, firms, creative and expressive pursuits. This session brings together four and consumers—as both creators and audiences. All papers are in ad- papers that inquire about the tricky, increasingly complex, and multi- vanced stages with empirical work completed. Papers also use mixed actor world of influencer business and provide conceptual and em- methods which will attract broader attention. Due to the significance pirical contributions to understand social media consumption. and relevance of the substantive context, we expect a range of ACR Monetization of influence provides fruitful areas of inquiry for members to show interest in the session. researchers. Among these, we uncover 1) entrepreneurial pursuits of ordinary consumers to become influencers, 2) the dialectical tensions Problematizing the Megaphone: The Very Difficult between promotion and expression in online content, 3) interactions Journey to Becoming an Influencer and relations between the audience, influencers, and firms, 4) codification and framing of content to meet audience and firm- expec EXTENDED ABSTRACT tations. Collectively, the papers in this session provide distinctive If we were to listen to the depiction of influencers in the main- insights into these issues. stream press and some academic articles, we might conclude that The first paper revisits the megaphone effect and provides a re- persistent efforts from meritocratic consumers will lead to them be- vision to the theory by demonstrating the struggles of ordinary con- coming influencers. Duffy and Wissinger (2017) refer to this mythol- sumers in their quest for becoming influencers. By acknowledging ogized portrait of the path to become an influencer as ‘glamorized.’ their struggles, this paper provides a needed corrective to the myth In reality, becoming an influencer is a difficult project that is unlikely of meritocratic and serendipitous journey of becoming and being an to yield results (Duffy 2016). Despite this acknowledgment of the influencer. Findings are expanded to explain why consumers leave struggles inherent to becoming and being an influencer, we know fields more generally. surprisingly little of what exactly these struggles are and how they The second paper provides a mixed method analysis of food affect ordinary consumers’ quest for influence. These are the two re- images on Instagram and explores how different visual framing strat- search questions we answer in this project. egies contribute to the engagement strategies of top-ten food influ- These research questions are important for many reasons. encers. Combining quantitative and qualitative analysis, the authors Centrally, there were more than 3.7 million ads by influencers on show distinct patterns of representation of food using strategic com- Instagram in 2018, and 90% of Instagram campaigns in 2018 used position and visual framing. These patterns also reflect long-standing micro-influencers—influencers that have somewhere between 1000 distinctions between the raw and the cooked. and 100 000 followers (HubSpot, July 15th, 2019; Wired, Apr 22nd, The third paper investigates the phenomenon of pet influencers. 2019). Micro-influencers represent about 25% of the Instagram user Through studying account managers who generate platform content base, or about 250 million people (Mention.com 2018). These de- around the performances of their pets, the authors uncover visual and velopments are giving ordinary consumers—consumers who do not rhetorical strategies of their engagement. The participants implement possess a privileged position in social fields (McQuarrie, Miller, anthropomorphism through intertextual manipulation of the visual and Phillips 2013)—increasing opportunities to capitalize on their and textual content and orchestration of realistic or fantasy perfor- (micro) influence. This has translated into an increasing number mances. of ordinary consumers wanting to become influencers: More than 75% of people aged between 5 and 38 state wanting to become one

(DailyMail, May 22nd, 2017; People, Nov 5th, 2019). The glamorized ney to propose an alternative to the dominant theorization of how picture of influencers is thus problematic, as it represents a highly consumers become influencers. Instead of consumers taking risks biased understanding of this role and the path to become an influ- when sharing their passion in a consumption, and how they accumu- encer. Theoretically, this is a glaring omission in a rapidly rising and late followers by doing so (see Erz and Christensen 2018; McQuar- important role of the digital economy. Practically, this mythologized rie, Miller, and Phillips 2013), we suggest conceptualizing at least a understanding of what is required to become an influencer can lead segment of influencers as digital entrepreneurs whose journey to be- consumers to pursue an identity project that is bound to fail. come influencers is active and intentional; For example, they readily We answer our research questions by analyzing repeat inter- create a brand from the get-go, strategically think about their market views with 23 micro-influencers with different followings: (1) less positioning, and use numerous market resources (e.g., buying fol- than 10k followers, (2) between 10k and 50k followers, and (3) lowers) to achieve their objective. Lastly, we generalize the struggles between 50k and 100k followers. We complement these interviews we identify to any consumer highly involved in some consumption with all Instagram posts and a month of Instagram Stories from our activity (e.g., see Stebbins 1982 on serious leisure) to provide three participants, as well as archival data in the form of articles for and on main reasons why they leave consumption fields. influencers (166 pages). We leverage McQuarrie, Miller, and Phillips (2013) work and Aesthetics of Food: The Role of Visual Framing concentrate on three main stages that explains the movement of Strategies for Influence Building on Instagram an ordinary consumer to online influencer: acquiring an audience, maintaining and expanding an audience, and deriving benefits from EXTENDED ABSTRACT an audience. Social media produces a megaphone effect that offers a mass When entering the field and starting to acquire an audience, audience to ordinary consumers (McQuarrie and Phillips 2013). Im- ordinary consumers face struggles associated with the breadth of ages of food dominate social media, which has revolutionized the expertise required to perform their role and the work demanded to world of gastronomy and food (Allué 2013). Food is a central part better their expertise to do so. For example, consumers realize that, of defining identity, as well as shaping collective, cultural identities in addition to being proficient in successfully making risky taste (Lupton 1996). The transformative, aesthetic aspects of food often displays (McQuarrie, Miller, and Phillips 2013), they also need to reflect social status and prestige. Food and eating, beyond survival know how to use the Instagram algorithm to their advantage, create and sustenance, can be understood as an aestheticized expression of thematically related images (vs. only one risky display of taste), and identity, status, and taste. Such taste expressions can be regarded as do proper make-up and poses. a taste regime that helps to explain how individuals gain cultural When maintaining and expanding their audience, consumers capital in the marketplace – often via social media (Arsel and Bean need to address changes in their audience expectations, which now 2013). By posting food images and seeking feedback from followers demand continuous content creation. We show how consumers take through the “likes”, the influencers provide repeated stimuli to con- two strategies to address this: they broaden their expertise to mul- trol followers’ behavior. Owing to the value creation effect, influenc- tiple fields (e.g., from fashion to food, travel, and design), or deepen ers are becoming a taste maker, and building an informal social norm their expertise in their focal field of interest (e.g., from fashion stylist (Hackman 1992) which will influence how followers perceive food. to critic, historian, and tailor). This increases demands in terms of Food serves as a powerful system of communication. Food and consumer expertise, which exponentially heightens the involvement eating connect consumers to their biological and cultural heritage required to continue pursuing their quest for influence. We also find (Allen 2012). Food is the basis of many consumer habits and rituals that consumers become objectified by their audience, which brings (Marshall 2005). The aesthetics of food consumption can be traced its own set of struggles associated with addressing negative and un- back to Epicureanism, an individualistic and communal philosophy wanted comments and requests. that emphasizes the central role of pleasure, laying the foundation Lastly, when converting their audience into benefits, consum- for aesthetic appreciation, including appreciation of food and taste. ers need to understand how to position themselves effectively in the French writer Brillat-Savarin introduced the idea of you are what market for influencers (i.e., have an effective positioning from the you eat, and articulated a philosophical approach to cooking (Brillat- perspective of brands who might hire them), how to develop their Savarin 2009). During the twentieth century, food marketing has unique brand, and how to deal with commercial clients. Here again, shifted from nutrition appeals to more hedonic taste appeals. Intrin- we find that consumers are often ill-equipped to address the scope of sic motivation such as seeking leisure becomes the main purpose tasks asked from them, which have become increasingly removed of participating in online interaction. Perceived enjoyment derived from what led them to start wanting to be an influencer in the first from the theory of reasoned action (TRA) and the technology accep- place. tance model (TAM) significantly influence online community’s loy- Each of these stages come with its own set of difficulties, which alty (Hsu and Lu 2007), this study, was inspired by these theories to leads influencers to abandon their project. At least in our sample, explain how influencers apply visual framing strategies as intrinsic these difficulties led to more than half our participants (13 outof motivation on technology use. To better understand influence mar- 23) to discontinue their influencer activities. And the number of fol- keting in the context of how technology interacts with gastronomy, it lowers did not play a role in convincing influencers to continue their is essential to analyze online images that connect taste consumption quest: We have many instances of Instagram influencers with more with visual consumption (Schroeder 2002). than 50 000 followers who abandoned their quest, for example, be- Among different social media platforms, Instagram special- cause converting their audience was deemed too difficult. izes in efficiently distributing visual rhetoric on a personal or global Our work contributes to the literature on influencers and con- scale to its over 500 million daily active users (Statista 2018). Food sumer involvement in a consumption field. First, the struggles we influencers create an informal social norm via the repeat use of a identify lead us to provide an important revision as to how ordinary circle loop starting from food preparation decision to feedback col- consumers become influencers (cf. Erz and Christensen 2018; Mc- lection and post recreation. This value creation process prompts the Quarrie, Miller, and Phillips 2013). Second, we use this revised jour- influencer to become a taste maker considering the preferences of 876 / Monetizing the Megaphone: How Consumers and Firms Use Platforms to Conquer the Attention Economy

“visual consumers”. In the historical view, this online food images Our analytical categories of raw and cooked food draw inspira- sharing is different from classical business model using their unique tion from anthropologist Claude Lévi-Strauss’s influential work on virtual “words”. It is a complex practice that has relational, social, the raw and the cooked as basic categories for understanding human and cultural significance. However, the visual aspect of these prac- culture (1983). Perhaps inadvertently, it appears that influencers’ tices, especially as it related to influencer marketing often remains social media posts reflect fundamental, long-standing distinctions understudied. Therefore, the purpose of this study is to identify fram- of the raw and the cooked. Of course, raw food, when appearing ing strategies that influence consumer engagement on social media. on Instagram, is generally highly aestheticized. However, we find The behavior of highlighting certain facets and concealing others is this distinction meaningful, and useful for understanding some ba- a central tenet of framing theory (Entman 1993). Framing affects sic elements of food posting. Moreover, these factors including so- how audiences interpret and react to scenarios, images, and text. By cial norm, perceived enjoyment, and cultural intermediaries have a taking advantages of various rhetoric tools like symbols, metaphors, significant effect shaping followers’ behavior via the value creation and depictions, visual imagery operates as a framing device (Rodri- process. Our framework emphasizes the importance of how food can guez and Dimitrova 2011). be accommodated within the framework of a meal, and how taste The research questions this paper aims to answer are: could be used as an approach to describe aesthetic norms. The inter- • RQ1:What do the top ten most-followed food channel Ins- action between real-life meal gatherings and social media creates a tagram influencers post on their accounts? combined type of “commensal” experience, one that fosters a sense • RQ2:What framing strategies do the top ten most-followed of community, despite occurring online (Bouvier 2018). Food rep- food channel Instagram influencers use to visually repre- resents an ideal way to achieve such a sense. This study represents sent food posts to facilitate consumer engagement? a step toward a more cogent understanding of food influencers on Instagram, informed by historical discussion of aesthetics and taste. The Instagram accounts for this study were selected based on In addition to implications such as the visual framing effects of ad- a formal ranking by Statista, which is an online statistic, market re- vertising for influencer marketing, future studies may investigate and search, and business intelligence portal that provides access to data the role of food influencers as ‘cultural intermediaries’ as described from market and opinion research institutions. Statista presented the Bourdieu (1984) to help us explain the set of activities and profes- most-followed food influencers on Instagram in the United States sions in the creation of markets and consumption economies like as of January 2018 (Statista 2018). Based on the reliable external how music bloggers mediate and orient consumer taste (Arriagada assessment data, this study used content analysis to investigate how and Cruz 2014). different categories of visual framing mutually or independently contribute to the popularity of these top ten food influencers. Orchestrating Pet Influencers: Rhetorical and Visual Using the food-related images posted by these food influencers from January 1 to December 31, 2017, and selected images that Strategies in Creating Mediated Platform Content received more than 10,000 “likes” on the influencer`s account. The EXTENDED ABSTRACT ten influencers included two men and eight women; one of the ac- As we are writing this, one of the most popular influencers in counts did not indicate gender. Images were randomly selected in the world is not a person but a cat with 4.3 million followers on each influencers’ Instagram, one image per month, obtaining a total Instagram. Nala is an 8-year-old cat who became Instagram famous of N = 120 posts. after her human companion created the account to share images with Content analysis views data such as images, text, and expres- her family who resides overseas. Nala’s audience has expanded to sions to be seen and interpreted for their meanings (Krippendorff other platforms, such as Facebook, YouTube, and Twitter, where she 2012). We also combined qualitative content analysis with frequency has 2.6M followers, 27K subscribers, and 24.9K followers, respec- and Chi Square-tests. We developed a codebook of attributes after tively, creating a multi-platform megaphone. observing most of the images and divided them into two parts: in- Despite the prevalence of pet influencers, there is a paucity of fluencers and food. The coding categories are partially based on the studies on the subject matter. While influencers or microcelebrities Duncan (1990) study, which offers a visual analysis using contextual who manage their own accounts have been explored by scholars readings of visual photographic imagery. Images were analyzed in (Abidin, 2015, 2016, 2018; Khamis, Ang, & Welling, 2017; Mar- 13 nominal categories. Acceptable levels of inter-coder reliability wick, 2013, 2015; McQuarrie et al., 2012; Senft, 2008, 2013)—also via Krippendorff’s alpha were achieved and ranged from 0.86 to 1.0 see papers in this session—, the mediated nature of pet influencers, for each variable tested. particularly the ways humans orchestrate their animal companions The findings show distinct patterns in how food is represented performatively, manage their online presence, and capitalize on the in Instagram posts. Food influencers prefer to post images about attention for their pets require further attention. This is important be- cooked food, without little decoration, using high contrast colors and cause what sets apart the phenomenon from others is its two bound- close-up shots. Most posts do not include “background” elements ary conditions: 1) The duality and mutuality of presence of human such as clothing, facial expression, or proximity to influencers. account managers alongside their pets in social media performances, Random placement and casual layout are more of a reflection of the and 2) The second-degree performance and sociality that is mediated influencer’s own leisure lifestyle; strategies also reflect influencers’ through the pets and with other pet companions. These boundary social network identity construction. Raw food images were found conditions allow contextualizing a new theorization of the triadic to be associated with cluttered composition and far away shoots, and mediated relationship between humans, their animal compan- whereas cooked food images were associated with high contrast and ions, and their audience in platforms. close-up, especially the top-down camera angle. Cooked food im- Our method is a combination of interviews and Netnographic ages serve as a justification agent for consumers, “thereby reducing observations. We conducted semi-structured interviews with 15 pet the conflict associated with the subsequent indulgent consumption account managers with at least 500 followers. We also analyzed the experience and increasing taste perceptions” (Poor, Duhachek, and content of pet influencer posts for visual and textual content. Lastly, Krishnan 2013, 124). through a public pet account, we participated in introspective en- Advances in Consumer Research (Volume 48) / 877 gagement with the practice. We performed iterative and inductive content are internet memes—defined as “(a) a group of digital items analytical procedures in analyzing our data. sharing common characteristics of content, form, and/ or stance, Our findings first demonstrate that Instagram is home toa which (b) were created with awareness of each other, and (c) were loosely networked bundle of micro-communities of cat accounts, circulated, imitated, and/or transformed via the Internet by many us- centered around feline appreciation and fandom. These connected ers” (Shifman, 2014, p. 41). Internet memes are a form of cultural micro-communities are built around hashtags singling out breeds, rhetoric (Grundlingh, 2018; Milner, 2013) and are unique given “the colors, and micro-interests with frequent overlaps and blurring of speed of their transmission and the fidelity of their form” (Davidson, community boundaries. Through a shared language (now coined 2012, p. 122). Internet memes can build report with online consum- as Meowlogisms by Podhovnik, 2018), collective celebration of ers and foster consumer engagement (Gelb, 1997). Previous research milestones, and collective action to support each other such as fun- highlights the convivial nature of internet memes, which act as a draising, pet account managers (PAMs) exemplify the three tenets type of digital leisure shared by internet insiders (Bauckhage, 2011). of community (Muniz and O’Guinn 2001): consciousness of kind, However, this form of leisure can also represent a lucrative opportu- rituals and traditions, and moral responsibility. nity for influencers and brands alike through monetizing the attention Second, we show that within this nexus of micro-communities, economy (Drenten, Gurrieri, and Tyler, 2020; Soha and McDowell, PAMs are motivated to create and maintain engagement for two pri- 2016). Most research to date explores how consumers create memes, mary reasons: connections or fame. These findings mirror and revise separate from marketing intervention. In contrast, our study explores existing typologies on social media communities (Kozinets 1999; brand-generated internet memes. We ask, how do brands create and Martineau and Arsel 2017). capitalize upon internet memes for commercial means? Third, we show that regardless of their goals, PAMs create an- To explore our research question, we turn to the context of Tik- thropomorphic performances for their pets through intertextual sto- Tok, a social media platform (formerly Musical.ly) which allows us- rytelling, switching back and forth between engaging with the pets ers to create and share short-form videos, using a library of “sounds” through proxy conversations and engaging with their human com- (e.g., music, user-generated sounds, television/movie clips). With panions more directly. over 1.5 billion app downloads and nearly 105 million in January Our data shows that storytelling can be either realistic (staying 2020 alone, TikTok is rapidly growing (Sensor Tower, 2019; 2020). within the boundaries of the sociomaterial capacities of animals), For this study, we specifically focus on sponsored hashtags, a paid or fantastical (expressing imaginary or impossible situations). When advertising placement option in which brands pay for custom spon- choosing to make posts on their accounts, PAMs can simply display sored hashtags to drive engagement in marketing campaigns. Spon- the raw images they capture by camera, such as their cat sitting on sored hashtags are placed on the app’s discovery-oriented ‘for you a couch, or they can decide to superimpose the original image onto page’ to engage consumers immediately upon opening the TikTok another one which showcases different settings or characters, such app. Data collection began in April 2019 and was conducted by a as their cat playing golf. Additionally, the same can occur for the cross-cultural research team. The researchers identified sponsored textual content: it could simply reflect the literal or figurative content hashtags via daily monitoring of the ‘for you page’ on TikTok and of the actual image, such as: “my cat sits on the counter,” or integrate created a database of sponsored hashtags and accompanying visu- intertextual narratives which may expand the literal object: “my cat al- and text-based content. Data collection resulted in a sample of slayed the dragons and conquered the seven kingdoms.” These story- 102 sponsored hashtags from 87 companies, such as Nike, Chipotle, telling techniques frame their pet as not only having human emotion- Burberry, and other multinational brands. Data were managed and ality and physicality but also allows the PAMs leverage on shared analyzed through qualitative social media methods (Sloan and Quan- tropes and narratives such as the case of Game of Thrones to boost Haase, 2017), using both Nvivo 12 Plus software and manual coding. engagement and familiarity. Textual cues are frequently supported Preliminary findings offer a framework of marketing memei- by visual manipulation, such as editing images to make the cat look fication wherein brands initiate the reproduction of content through like a knight. Lastly, PAMs deliberately choose the account’s voice three types of calls to action: impersonation, transformation, and (human versus cat) to supplement this storytelling. Building on the self-expression. Marketing memeification unfolds on a continuum patterning of these strategies, we develop a typology of PAM ac- of individual creativity in which consumers are increasingly encour- count management styles. aged to insert their own creative ideas when developing content in Our work aims to contribute to a growing body of literature on conjunction with a sponsored hashtag. In each case, paid influenc- performances in social media (Abidin, 2017; boyd, 2011; Burgess ers set the exemplary creative standard, while everyday consumers and Green, 2008; García-Rapp, 2017; Marwick, 2013; Strangelove, engage with marketing memeification as a form of unpaid consumer 2010; Senft, 2013). We show how social media enables performa- leisure. Our study suggests TikTok is unique in using sounds, or son- tive mediated interactions and mediated connections between hu- ic anchors, to drive the marketing memeification process. mans through their pet companions, how animals are orchestrated First, brands initiate marketing memeification through imper- for attention economy, and how this shifts humans’ connections with sonation-based calls to action, which require little to no individual their pets and other humans. Furthermore, we develop a typology of creativity. Impersonation involves movement-oriented actions (e.g., performative orchestration of Instagram pets. Our typology also pro- dance, lip-sync) with specific guidelines for the performance. For vides firms tools for identifying and recruiting the most appropriate example, to advertise Season 3 of Netflix’s Chilling Adventures of influencer based on their performance style. Sabrina, the media company created the #StraightToHell hashtag accompanying a short, easy to replicate dance with the song “Straight Marketing Memeification on TikTok: Initiating and to Hell” by Sabrina Spellman as the sonic anchor. Through imper- Leveraging Consumer Creativity for Commercial Means sonation, consumers are mobilized to recreate content, exemplified by influencers, directly as it is shown. While some users may per- EXTENDED ABSTRACT form the dance moves differently, due to ability and style, the brand- In a saturated digital advertising space, brands increasingly de- related hashtag does not directly invoke creativity. sire ‘viral’ content (Berger and Milkman, 2012). One form of viral 878 / Monetizing the Megaphone: How Consumers and Firms Use Platforms to Conquer the Attention Economy

Second, through transformation-based calls to action, which Allué, T. (2013). Food sector communication and online give more narrative and creative leeway to consumers. Transfor- influencers. Catalan Journal of Communication & Cultural mation involves transition-oriented stories, wherein consumers are Studies, 5(2), 311-314. asked to show a change or evolution. For example, in promoting Arsel, Z., & Bean, J. (2013). Taste regimes and market-mediated the re-release of Wendy’s ‘spicy chicken nuggets’, the brand encour- practice. Journal of Consumer Research, 39(5), 899-917. aged consumers to demonstrate transformational stories using the Bauckhage, C. (2011). Insights into internet memes. In Proceedings #SpicySnap hashtag challenge and the song “Snap Yo Fingers” by of the Fifth International AAAI Conference on Weblogs and Lil Jon as the sonic anchor. Through transformation, consumers are Social Media, 42-49. mobilized to interpret the brand’s call to action and tell their own Berger, J., & Milkman, K. L. (2012). What makes online content creative stories loosely related to the branded product. For this chal- viral?. Journal of Marketing Research, 49(2), 192-205. lenge, influencers prominently feature Wendy’s food products in Bernazzani, S. (2019). Micro-influencer marketing: a their content; however, everyday consumers use the suggested sound comprehensive guide, HubSpot, October 16th, available and hashtag to show their own creative transformations. at: https://blog.hubspot.com/marketing/micro-influencer- Third, through self-expression-based calls to action, which marketing. inspire consumers to reflect their personal points of view. Self-ex- Boyd, d., 2011. Social network sites as networked publics:. In Z. pression involves identity-oriented displays of uniqueness in which Papacharissi, A networked self. New York: Routledge, 39-58. consumers are asked to interpret the hashtag, with open-ended param- Bourdieu, P. (1984). Distinction: a social critique of the judgment eters. For example, beauty retailer Ulta Beauty created the #BeautyIs of taste, Cambridge, MA: Harvard University Press. hashtag campaign, with the song “Here I Am” by Fleur East as the Bouvier, E. (2018). Breaking bread online: Social media, sonic anchor, inviting consumers to broadly define beauty in their photography, and the virtual experience of food. At The own terms. Through self-expression, influencers and everyday con- Interface/Probing The Boundaries, 97, 157-172. sumers promote their own unique viewpoints and perspectives, us- Brillat-Savarin, J. A. (2009). The physiology of taste, or, ing the brand as a launching point. User videos range in diversity meditations on transcendental gastronomy. New York: in messaging, format, and content creators themselves (e.g., gender, Everyman’s Library. ethnicity, age). Thus, on the marketing memeification continuum of Burgess, J., & Green, J. (2018). YouTube: online video and creativity, branded hashtags and accompanying sounds increasingly participatory culture. New York: John Wiley & Sons. become a pathway for consumers to share their own authentic self- Ian P. Bigley & James M. Leonhardt (2018) Extremity Bias in expression, while indirectly promoting the brand itself. User-Generated Content Creation and Consumption in Social While previous work focuses on the organic evolution of Media, Journal of Interactive Advertising, 18(2), 125-135. memes as user-generated content (Bigley and Leonhardt, 2017), our Daily Mail (2017). Forget being a nurse or doctor, three quarters findings extend existing research to better understand how marketing of today’s children would rather be Youtubers and vloggers,” intervention spurs memeification for commercial means and to dem- Daily Mail, May 22nd, available at: https://www.dailymail. onstrate how brands leverage the status and visibility of social media co.uk/news/article-4532266/75-cent-children-want- influencers to set creative standards. Partitioning devices including YouTubers-vloggers.html. sonic anchors and branded hashtags are initiated by the brand; how- Davidson, P. (2012). The language of internet memes. The Social ever, marketing memeification decentralizes control of the brand in Media Reader, 120-134. the process. In line the theoretical perspective of memes (Shifman, Drenten, J., Gurrieri, L., & Tyler, M. (2020). Sexualized labour 2014), our study suggests meaning is created on TikTok during the in digital culture: Instagram influencers, porn chic and the transmission of cultural content. At a meso-level, this transmission monetization of attention. Gender, Work & Organization, process is propelled by sonic anchors (e.g., songs) which work to 27(1), 41-66. coalesce consumption communities, reflecting the importance of Duffy, B. E. (2016). The romance of work: gender and aspirational sounds as objects in consumption research (Patterson and Larsen, labour in the digital culture industries. International Journal of 2019). This study represents a first step in understanding how brands Cultural Studies, 19(4), 441–57. operate on TikTok, wherein consumers’ creative performances are Duffy, B. E. and E. Wissinger (2017). Mythologies of creative actively initiated, surveilled, and leveraged for commercial means. work in the social media age: fun, free, and ‘just being me.’ International Journal of Communication, 11, 4652-71. REFERENCES Duncan, M. C. (1990). Sports photographs and sexual difference: Abidin, C. (2015). Communicative intimacies: Influencers and Images of women and men in the 1984 and 1988 Olympic perceived interconnectedness. Ada: A Journal of Gender, New Games. Sociology of Sport Journal, 7(1), 22–43. Media, and Technology, (8). Available: https://adanewmedia. Entman, R. M. (1993). Framing: toward clarification of a fractured org/2015/11/issue8-abidin/. paradigm. Journal of Communication, 43, 51-58. ______(2016). “Aren’t these just young, rich women doing vain Erz, A. and A-B. H. Christensen (2018). Transforming consumers things online?”: Influencer selfies as subversive frivolity. into brands: tracing transformation processes of the practice of Social Media+ Society, 2(2), 2056305116641342. blogging. Journal of Interactive Marketing, 43, 69–82. ______(2017). # familygoals: family influencers, calibrated García-Rapp, F. (2017). Popularity markers on YouTube’s attention amateurism, and justifying young digital labor. Social Media+ economy: The case of Bubzbeauty. Celebrity Studies, 8(2), Society, 3(2), 2056305117707191. 228-245. ______(2018). Internet celebrity: understanding fame online. Gelb, B. D. (1997). Creating “memes” while creating advertising. Emerald Publishing. Journal of Advertising Research, 37(6), 57-60. Allen, J. S. (2012). The omnivorous mind: our evolving relationship Grundlingh, L. (2018). Memes as speech acts. Social Semiotics, with food. Cambridge, MA: Harvard University Press. 28(2), 147-168. Advances in Consumer Research (Volume 48) / 879

Khamis, S., Ang, L., & Welling, R. (2017). Self-branding,‘micro- Patterson, M., & Larsen, G. (2019) Listening to consumption: celebrity’ and the rise of Social Media Influencers.Celebrity Towards a sonic turn in consumer research. Marketing Theory, Studies, 8(2), 191-208. 19(2): 105-127. Kozinets, R. V. (1999). E-tribalized marketing?: The strategic Podhovnik, E. (2018). The Purrification of English: Meowlogisms implications of virtual communities of consumption. European in Online Communities: A qualitative description of a Management Journal, 17(3), 252-264. selected# blackcat community on Instagram. English Today, Krippendorff, K. (2012). Content analysis: An introduction to its 34(3), 2-16. methodology. London: Sage. Poor, M., Duhachek, A., & Krishnan, H. S. (2013). How images Lévi-Strauss, C. (1983). The raw and the cooked, trans J. of other consumers influence subsequent taste perceptions. Weightman and D. Weightman. Chicago: University of Journal of Marketing, 77(6), 124-139. Chicago Press. Rodriguez, L., & Dimitrova, D. V. (2011). The levels of visual Lupton, D. (1996). Food, the body and the self. London: Sage framing. Journal of Visual Literacy, 30(1), 48-65. Publications. Schroeder, J. (2002). Visual consumption. New York: Routledge. Martineau, E., & Arsel, Z. (2017). Managing communities of co- Senft, T. M. (2008). Camgirls: Celebrity and community in the age creation around consumer engagement styles. Journal of the of social networks (Vol. 4). Peter Lang. Association for Consumer Research, 2(2), 179-195. ______(2013). Microcelebrity and the branded self. A companion Martineau, P. (2019). Inside the weird, and booming, industry of to new media dynamics, 346- 354. online influence,.Wired , April 22nd, available at: https://www. Sensor Tower (2019, November 14). TikTok Clocks 1.5 Billion wired.com/story/inside-the-industry-social-media-influence/. Downloads on The App Store and Google Play. Retrieved Marwick, A. E. (2013). Status update: Celebrity, publicity, and January 2, 2020, from https://sensortower.com/blog/tiktok- branding in the social media age. New Haven: CT: Yale downloads-1-5-billion. University Press. Shifman, L. (2014). The cultural logic of photo-based meme genres. ______(2015). You May Know Me from YouTube:(Micro‐) Journal of Visual Culture, 13(3), 340-358. Celebrity in Social Media. In A companion to celebrity, P. Sloan, L., & Quan-Haase, A. (2017). The SAGE Handbook of David Marshall and Sean Redmonds, (Eds.), 333-350, New Social Media Research Methods. Thousand Oaks, CA: SAGE York: Wiley. Publishing. McQuarrie, E. F., J. Miller, and B. J. Phillips (2013). The Soha, M., & McDowell, Z. J. (2016). Monetizing a Meme: megaphone effect: taste and audience in fashion blogging, YouTube, Content ID, and the Harlem Shake. Social Media + Journal of Consumer Research, 40(1), 136–58. Society. https://doi.org/10.1177/2056305115623801 Mention (2018). 2018 Instagram Report, available at: https://info. Statista. (2018). Number of daily active Instagram users from mention.com/hubfs/Instagram%20Engagement%20Report%20 October 2016 to September 2017 (in millions). Retrieved from 2018%20%7C%20Mention.pdf. https://www.statista.com/statistics/657823/number-of-daily- Milner, R. M. (2013). Pop polyvocality: Internet memes, public active-instagram-users/ participation, and the Occupy Wall Street movement. Stebbins, R. A. (1982). Serious leisure: a conceptual statement, The International Journal of Communication, 7, 2357-2390. Pacific Sociological Review, 25(2), 251-72. Muniz, A. M., & O’guinn, T. C. (2001). Brand community. Journal Strangelove, M. (2010). Watching YouTube: Extraordinary videos of Consumer Research, 27(4), 412-432. amples Re-Examining Consumer Identity on Digital Platforms Chairs: Ela Veresiu, York University, Canada Linda Tuncay Zayer, Loyola University Chicago, USA Marie-Agnès Parmentier, HEC Montréal, Canada

Paper #1: How Digital Platforms Shape Family Identities and 4) how do online portals contribute to contemplative consumer Jenna Drenten, Loyola University Chicago, USA identity creation? Linda Tuncay Zayer, Loyola University Chicago, USA Overall, this special session asks: how do digital platforms (re) shape consumers’ individual and collective identities? We anticipate Paper #2: #Instagrannies?: How Mature Women Reassemble that our session will be relevant for scholars engaged with broad Aging, Fashion and Retirement in Digital Times questions that relate to marketing in a digital world, and consumer Ela Veresiu, York University, Canada researchers with a specific focus on participatory media consump- Marie-Agnès Parmentier, HEC Montreal, Canada tion. We also wish to engage in dialogue with practitioners and social Paper #3: An Exploration into How Influencers Shape Gender scientists working in influential social media organizations such as Identity Among Gen Z Consumers Pinterest through a discussion of our respective empirical findings. Kathrynn Pounders, The University of Texas at Austin, USA Paper #4: The Contemplative Consumer Identity: Bridging How Digital Platforms Shape Family Identities Spirituality and the Marketplace Richard Kedzior, Bucknell University, USA EXTENDED ABSTRACT Daiane Scaraboto, University of Melbourne, Australia Past scholarship has documented how consumers in online environments construct their identities (Schau and Gilly 2003; Belk SESSION OVERVIEW 2013), including navigating identity transitions (Drenten and Zayer From the explosive popularity of podcasts to the rapid rise of 2018). Additional research has highlighted the role of digital tech- social media influencers, seismic shifts in technology and new media nologies in shaping family practices (see Epp, Schau, and Price 2014 consumption are taking shape across the globe. As digital platforms study on geographic dispersion of families), reinforcing rituals and – defined as “a set of online digital arrangements whose algorithms intimacy among couples (see Su 2016 study on digital messaging), serve to organize and structure economic and social activity” (Ken- and providing a means to negotiate meanings of home (see Ca- ney and Zysman 2016) – increasingly take on a larger role in individ- balquinto 2018 study on mobile media use of Filipino transnational ual consumers’ lives, the digital world holds widespread implications families). Importantly, Epp et al. (2014) highlight how families can for consumer identity construction (Belk 2013). These consumer expand their material, expressive and imaginative capacities through identities are neither static nor binary (e.g. online versus offline, see technologies such as Skype during times of separation. However, the Kedzior 2015), but rather assembled and at times disassembled with- manner in which technologies, such as social media platforms, im- in and between various digital platforms, such as social networking pact family identities is yet to be explored in great depth. To address sites, image-based or audio based social platforms, and hybrid sites. this gap in the literature, our research asks: how are family identities Indeed, past scholarship has examined the ways in which consumers shaped by digital platforms? and influencers actively engage in digitally mediated spaces (e.g., We draw on Epp and Price’s (2008) sociological conceptualiza- Drenten and Zayer 2018, Drenten, Gurrieri and Tyler 2018, Epp, tion of ‘being a family’ as a “bundle” of identities, which include Price and Schau 2014; Gurrieri and Cherrier 2013, Molesworth and collective, relational (couple, parent-child or sibling identities), and Denegri-Knott 2013, Schau and Gilly 2003). individual identities as they interact with communication forms and However, while there has been much discussion regarding the symbolic marketplace resources. The authors refer to Bennett, Wolin, more negative impact of digital platforms (Baccarella et al. 2018) and McAvity’s (1988, 212) definition of family identity as the “ge- on individuals and families, consumer research scholarship has yet stalt of qualities and attributes,” and “the family’s subjective sense of to fully explore how digital platforms (re)shape and/or disrupt con- its own continuity over time, its present situation, and its character.” sumer identities, and what these processes mean not only for con- The context for our research is the social media platform, Pinterest, sumer welfare outcomes, but also for marketing practitioners. In which allows users to search, save and share content in the form of this special session we examine this timely oversight by bringing “pins” and “boards” related to their interests akin to a bulletin board. together an inclusive team of authors from both academia and the Two-thirds of the pins on the site represent brands or products. The social media industry, exploring digital consumption issues through site boasts that 28% of all US social media users are Pinterest users multiple theoretical lenses, like family identity theory, intersectional with a monthly active user base of 300 million and a median age of theory, and assemblage theory, as well as methodological paradigms 40 (Omnicore Agency Report 2020). including quantitative and qualitative perspectives. The first two pa- To explore family identity in the digital era, we draw upon a pers of this session will focus on group identity construction in the dual methodology of in-depth interviews and visual content analysis. digital age, while the last two papers will highlight the role of digital First, interviews with 14 men and women, ages 21 to 45, who use platforms on individual consumer identity creation. Specifically, the Pinterest were conducted, resulting in over 500 pages of text. Visual empirical papers, which all have data collected and are in mature data was also captured as informants shared their personal Pinterest stages of theorization, will address the following research questions: boards. Second, a series of searches were conducted via Pinterest 1) how are family identities (re)shaped by digital platforms? 2) how Trends (trends.pinterest.com), using “Family” as the initial keyword, are female influencers and consumers’ aging identities shaped and then systematically exploring the first 20 results (e.g., family dinner constrained by social media images? 3) how do Instagram influenc- ideas, family room ideas, family pictures). Themes related to iden- ers impact their consumers’ gender identity and gender expression? tity and family were identified using a constant comparison method

Advances in Consumer Research 880 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 881 and iteratively referring back to the existing literature (Strauss and #Instagrannies?: How Mature Women Reassemble Corbin 1998). Aging, Fashion and Retirement in Digital Times We find Pinterest provides symbolic marketplace resources in the form digital virtual consumption ‘objects’ which are embedded EXTENDED ABSTRACT in communications forms (e.g., narratives, rituals, intergenerational Until recently, the idea of paying close attention to mature transfers, social dramas, everyday interactions) and help consumers women’s sense of style would have been met with puzzlement and construct and maintain collective family identities. Exploratory visu- little interest from consumers and the fashion industry alike. After al analysis of the most popular pins in the Pinterest Trends database all, to be fashionable and feminine has typically been viewed as indicates that it is replete with digital virtual resources to shape and belonging to the young. What then could older female consumers enact family identities (e.g., family time, family tree, family tradi- possibly teach the market about being stylish? Following the unex- tions) as well as to help families overcome potential barriers (e.g., pected popularity of the Advanced Style phenomenon, discussions inspirational family quotes, addressing family problems). about ageism, gender, and fashion have begun to draw increasing at- Emergent from our findings are four unique forms of family tention in the popular press (e.g., Grinberg 2012; La Ferla 2018). The identity resources: relational, functional, transformational, and ex- @advancedstyle Instagram account created in 2008 by American pressive. Due to space limitations of the abstract, abbreviated quotes street style photographer Ari Seth Cohen helped fuel these critical from the data are presented here. First, relational identity resources conversations. Cohen’s celebration of the personal styles of regular reflect connection and bonding with fellow family members, such mature women, which are commonly understood as 50 years of age as pinning family game night ideas or books to read at bedtime. El- or older (Twigg 2013), has launched a flourishing online movement. len, a college student, uses Pinterest to save tips for playing ‘Words A decade after its creation, the social media account has more than with Friends,’ a mobile game that she plays each day with her dad. 278,000 followers and 4,300 posts, boasts a hashtag (#advanced- Ellen’s pins enhance her long-distance gameplay experiences. She style) used more than 169,000 times, is regularly featured in major says, “… my dad wants to play with me; of course, I’m going to fashion magazines (e.g., Vogue) and mainstream media (e.g., The play with him. This [pin] had different words that a scrabble player Guardian), and has expanded into the realm of books (e.g., Cohen should memorize so like highest possible points on the first turn.” 2012, 2016; Cohen and Schraer 2013) and feature length films (e.g., Second, functional identity resources reflect pragmatic family opera- Cohen and Feig 2014). tions, such as recipes for weeknight meals or chore charts for kids. In light of this social media and traditional market success story, For example, Carolyn pins content related to planning an upcoming and in an effort to answer research calls for greater examination of Disney vacation, including a Disney packing list and meal tips for the intersection between the sociocultural categories of age and gen- eating at the theme park. “I used Pinterest a ton before I went to Dis- der in the marketplace (Twigg 2013), we explore how women over ney because I wanted to make sure I was fully prepared because Dis- 50 are embracing the online opportunity to increase their visibility ney was a whole different kind of vacation, you have to be ready.” by consuming the photo-sharing digital platform called Instagram, as Carolyn’s pins reflect desired organization and preparedness for her well as amplify their voices and representations in the fashion market family. Third, transformational identity resources reflect evolutions by becoming Instagram style influencers. As such, our research aims in family structure or values, such as efforts to get healthy together are: 1) to expand our knowledge of influencer marketing dynamics, or moving in together. Hailey is recently engaged and pins home which heavily favors young prosumers, by examining an under re- décor ideas for moving in with her future husband. She says, “…it searched older female segment; and 2) to investigate Advanced Style gives me inspiration … it just makes me feel inspired to actually do prosumers’ impact on broader market and societal representations of stuff rather than to just continue to watch HGTV.” Fourth, expres- age, gender, and retirement. sive family identity resources reflect on shared meanings and be- To do so, we combine the intersectional and assemblage theo- liefs, such as faith and emotional concern for family members. Ellen retical lenses and focus on the case of 10 popular female Advanced discusses her family’s belief to “make sure you hug your loved ones Style Instagram influencers, defined as ordinary Internet users who today and every day” and states “…if I see something that is along have an established credibility and audience, and therefore can mon- lines of that I’m totally pinning it.” In some cases, family identity etize their online following through advertorials, brand collabora- resources intersect serving as multiple family identity resources. For tions, and public appearances (Abidin 2016). While prior consumer example, pins for planning a family vacation may be both functional research has examined mature consumers as retirees (Price, Arnould, and relational. These family identity resources are leveraged through and Curasi 2000; Schau, Gilly, and Wolfinbarger 2009) and as part digital media (e.g., Pinterest) to help families reinforce or reimagine of elderly consumption ensembles (Barnhart and Peñaloza 2013), their collective identities. our study pays particular attention to a new kind of older prosumer As Epp and Price (2008) point out, the marketplace can both en- or producer-consumer hybrid, who, through their savvy digital con- able and constrain the identity construction practices of families, and sumption and glamor labor (Wissinger 2014), reassemble the imag- that there are barriers to enactment, including geographic dispersion, ery of aging, gender, and retirement for their combined millions of time constraints, and lack of monetary resources. We find Pinter- online followers. est diminishes these barriers by offering asynchronous participation Methodologically, we employed a qualitative case study and a seemingly endless array of digital virtual consumption as re- method, as it “has become a favored methodology for researching sources. Importantly, contrary to popular press and prior research macroscopic […] questions concerning markets and cultures from indicating that digital technologies are disrupting and introducing an interpretive perspective” (Holt 2002, 73). To help contextualize new vulnerabilities to family relationships and roles (Abbasi and our market phenomenon, we first collected an overview of popular Alghamdi 2017; Hertlein 2012; Lewis et al. 2015), we find that the written, audio, and visual materials about the Advanced Style move- social media platform Pinterest provides new or reframed opportuni- ment since its inception in 2008. Following this initial phase of data ties for reinforcing and reimagining family identities. collection, we focused on 10 popular Advanced Style influencers, where we collected all online media interviews (291 documents) and YouTube video interviews (93 videos) featuring any of these 10 882 / Re-Examining Consumer Identity on Digital Platforms influencers. To supplement the archival dataset, we conducted a net- rectly impact the purchase decisions of large audiences (De Veirman, nography to understand the social interaction in this contemporary Cauberghe, and Hudders 2017; Djafarova and Rushworth 2017). digital communications context (Kozinets 2019). Here, each author While a body of research in marketing has begun to investigate the individually engaged in online participant observation as regular effectiveness of influencers, and factors that may intercede their ef- fans by following the 10 influencers on Instagram from our personal fectiveness, there is a lack of work investigating how influencers accounts for approximately six months, periodically sharing direct affect consumers’ gender identity and gender-related consumption. messages on the platform with each other about our online experi- As such, this work is an attempt to increase our understanding ences (Kozinets, Scaraboto, and Parmentier 2018). We also down- of how influencers impact consumers’ gender identity and expres- loaded fans comments for 30 randomly selected recent consecutive sion. First, an exploratory study, which was conducted to determine posts per influencer to better understand fan interactions. Lastly, we whether or not influencers affect gender identity and gender expres- collected metrics on the 10 influencers’ Instagram accounts regard- sion, is presented. Based on these findings, the theoretical founda- ing their audience, engagement, and brand mentions from the third- tions of social comparison theory and self-concept clarity are used to party analytics platform HypeAuditor. develop a set of predictions that will be tested using a between-sub- In accordance with the established principles of qualitative jects experimental design to shed light on to how influencers impact research (Glaser and Strauss 1967), data collection continued until consumers’ gender identity and gender expression. theoretical saturation was reached. The complete dataset was ana- Gender is considered to be a cultural category that is reinforced lyzed using the hermeneutic approach (Thompson 1997) typically by socially constructed and contested assumptions and norms. Prior utilized for qualitative data. We find that the Advanced Style inter- work has shown media produce and communicate social signifiers, sectional assemblage was created through a specific historical pro- including gender, specifically depictions of femininity and masculin- cess (DeLanda 2016) that moved from invisible to visible and then ity (Hirschman and Stern 1994; Stevens 2012; Zayer et al. 2012). progressed from influence to precarious power. In the first phase of According to Schroeder and Zwick (2004, 24), “advertising dis- the process, our influencers created the intersectional assemblage by course both reflects and creates social norms” and media and adver- first recognizing their invisibility in the fashion market as women tising representations influence both cultural and individual percep- over 50, then deliberately deciding to become style activists, next tions of gender identity, femininity, and physical attractiveness. strategically using hashtags, and finally making deliberate and con- An exploratory study was conducted among 81 undergraduate sistent choices about their Instagram style. In the second phase, our students (74% female, Mage =20.2) to assess whether influencers on prosumers stabilized the intersectional assemblage by first recogniz- the social media platform Instagram has impact on perceptions of ing their responsibility as influencers, then engaging with their on- gender identity and gender expression. Participants were provided line fans, and finally accepting strategic brand collaborations. with a definition of gender identity and a definition of gender ex- Overall, we aim to contribute to the nascent, yet important lit- pression from the Human Rights Campaign website. They were then erature stream investigating intersectionality in contemporary mar- asked to answer four open-ended questions to determine whether keting images (Gopaldas and Siebert 2018). Our focus is the inter- influencers impact their gender identity, their gender expression, as sectionality of age and gender in consumer-driven images, which is a well as the gender identity and gender expressions of their friends. timely endeavor since the Advanced Style phenomenon spotlighting The responses to these questions were read and coded by the fashionable older women has received minimal scholarly attention researcher and graduate assistant. Each identified main themes and to date (Jermyn 2016; Tiidenberg 2018). Specifically, our work of- sub-themes for each question, and then combined findings from the fers three important insights. First, in contrast to previous research questions to identify common themes and sub-themes. Findings sug- on influencer marketing, which has focused on young females, we gest that participants largely did not believe that influencers held provide a process to understand how mature women directly impact sway over their gender identity, although this varied based on one’s the ageist fashion and beauty industries by becoming Instagram in- confidence in their own gender identity. Participants overwhelm- fluencers. At the same time, we offer a direct extension to the retire- ingly believe influencers play a large role in their gender expression. ment consumption literature that finds “retirement explicitly privi- A sub-theme of social comparison with the influencer also emerged leges consumption (living) over production (work)” (Schau et al. when analyzing the question about gender expression. Specifically, 2009, 256). In sharp contrast, we find that retirement is no longer participants often discussed comparing oneself with an influencer in purely a consumption affair but rather a new form of prosumption, terms of femininity and masculinity. Finally, in assessing responses whereby retirees or those approaching legal retirement age willingly to the question about whether influencers impact perceptions of gen- find creative ways to monetize their consumption activities and hob- der identity among their friends, responses were overwhelmingly bies. Finally, our research extends gender studies by conceptualizing positive with most participants stating that influencers play a large gender as an assemblage of market-mediated human and non-human role in their friends’ gender identity and gender expression. Analysis technological elements in the digital age. also revealed that when discussing the effect that influencers have on their peers’ gender identity, it was often in reference to friends or An Exploration into How Influencers Shape Gender peers that may not be as self-assured in their gender identity. Identity Among Gen Z Consumers A dearth of literature has examined the effects that traditional advertising has on female consumers’ self-perceptions, largely fo- EXTENDED ABSTRACT cusing how exposure to idealized imagery negatively impacts wom- Influencers attract millions of followers by sharing content en’s self-perceptions (D’Alessandro and Chitty 2011; Hogg and Fra- curated from their daily lives on platforms such as Instagram and gou 2003; Pounders, Rice, and Mabry-Flynn 2016; Richins 1991; YouTube and typically evolve around one particular domain. For Smeesters and Mandel 2006). The bulk of this work uses social com- example, popular categories of influencers include interests such as parison theory (Festinger 1954), which states individuals engage in fashion, beauty, and fitness. Marketers and advertisers have enthusi- spontaneous comparisons to others in order to evaluate the self on a astically embraced influencers as they have are perceived as reliable particular attribute to explain these relationships.Similarly we pre- sources of information among consumers and have the power to di- dict that consumers will engage in spontaneous comparisons with Advances in Consumer Research (Volume 48) / 883 influencers and that these comparisons will impact consumers’ self- Suddaby 2019). Yet, as Redden (2016, 5) notes, “there often appears perceptions in the form of self-esteem, femininity, and satisfaction to be a reticence to examine the market dynamics more closely” in with physical appearance. We also anticipate that these comparisons spiritual contexts, “even where it would provide explanatory power.” are related to gender expression. Further, we expect that these rela- Hence, more research is needed on “the role of market actors; in par- tionships will be moderated by type of influencer. ticular to the role of spiritual practitioners, entrepreneurs and leaders, Self-concept clarity refers to “the extent to which self-beliefs in shaping the spiritual market offer” (Huseman and Eckhardt 2019, are clearly and confidently defined, internally consistent, and stable 395). (Campbell et al. 1996). Accordingly, when consumers have low self- Our study addresses these gaps by examining the market- concept clarity, their self-concept is less coherent and more unstable. making dynamics in the spiritual field of Tarot reading. We focus on Prior research has found that low self-concept clarity has two main how online spiritual portals contribute to spiritual consumer identity effects on consumer behavior: it increases the degree to which con- creation. Tarot is a deck of 78 cards whose origins date back to Re- sumers are impressionable and influenced by external self-relevant naissance Italy. Although much of its history is shrouded in mys- information (Campbell 1990) and reduces the psychological resourc- tery, since late XVIII century, Tarot evolved into a fortune-telling es consumers have available to process identity-related information and divinatory device, with its cards acquiring symbolic and esoteric (Campbell 1990). For example, Campbell (1990) found that consum- meanings. Recent shifts in the spiritual marketplace have made Tarot ers with low self-concept clarity take longer to determine if an adjec- more accessible (Gregory 2016), as materials are available online tive represents their identity. We believe that self-concept clarity will and resources (e.g. Tarot decks and books) have proliferated in the moderate the relationship influencers have on self-perceptions. Spe- market. cifically, we predict influencers are more likely to impact consumers’ In collecting and analyzing data for this project, we follow the (a) gender identity and (b) gender expression when consumers have methodological recommendations for studying spiritual consump- low self-concept clarity about their gender identity. tion (Ozanne and Appau 2019). We are sensitized to the socio-his- These predictions will be examined in a 2 (influencer type) X 2 torical context in which consumers’ interest and engagement with (self-concept clarity) between-subjects experiment using influencers Tarot as a spiritual path happens, and explored indigenous ways of identified in the exploratory study. The sample will consist of female interpreting and theorizing the phenomenon. As part of a larger ef- consumers ages 18-25. Data will be collected in April and analyzed fort to collect ethnographic and netnographic data on the market of using ANOVA and Process Model 8. Findings will be ready for pre- Tarot, we collected data from Lilly Tarot (pseudonym), one of the sentation by the 2020 Annual ACR conference. It is expected that the largest online portals for Tarot readers. Founded in 1999 by Belinda findings from this work will contribute to the literature on gender and Maffei (pseudonym), Lilly Tarot is one of the earliest online spiritual identity by shedding light into how influencers and online platforms portals in the emergent Tarot market. Belinda produced a podcast impact gender identity among Gen Z women. series of 146 episodes (average length 40 minutes) in which she introduces Tarot to those interested in becoming readers. Her podcast The Contemplative Consumer Identity: Bridging series reached 2 million daily visits. Each episode was accompanied Spirituality and the Marketplace by a full transcription, which we downloaded and archived, total- ing 479 pages of single-spaced text. We also read all the additional EXTENDED ABSTRACT resources included in each episode (e.g. card spread templates and A growing stream of research examines the conflation of mar- links to websites). A common feature in the Lilly Tarot podcast series kets and spirituality, focusing on the roles consumers assume at this is interviews with other actors in the Tarot market (e.g. book authors, intersection (Huseman and Eckhardt 2019). Two approaches can card deck designers, numerologists). Interviewees often discussed be identified. The first assumes a critical stance on the interface of their trajectories as Tarot experts, and their perceptions of how Tarot spirituality and markets, noting appropriation, distortion, and com- has evolved through the decades. moditization of religions or other belief systems (Rindfleisch 2005). The podcast series offers us a window into the development It characterizes the (re)emergence of spiritual systems as consumer- of the Tarot market, as it allows us to “account for changes from a ist markets or “spiritual supermarkets” (Redden 2016, 2). For ex- specific point of view” (Hagberg 2016, 114). We identify two mar- ample, Rindfleisch (2005, 346) argues that entrepreneurs in the New ket-making strategies through which online portals and spiritual en- Age field treat complex traditional systems of beliefs in “partial and trepreneurs act to bridge spirituality and marketization: First, is “The superficial” ways, which “reduces and decontextualizes their mean- High Priestess” – designing the contemplative consumer. Belinda ings,” thereby making them palatable for consumer markets. The and her interviewees leave no room for listeners to assume roles second approach is reconciliatory and examines the hybridization of other than reader and client – seller and buyer – in the Tarot read- commercial and spiritual worlds as a phenomenon of interest without ing market. Learners are considered entrepreneurs and encouraged assuming that one value regime is morally or ideologically superior to practice by offering free readings for friends and family to accu- to the other (Scaraboto and Figueiredo 2017; Gibson-Graham 2008). mulate “confidence enough to charge.” Yet, in any of these roles, Be- Here, consumer identities seamlessly blend with other non-commer- linda argues, individuals should focus on unlocking “the true power cially defined roles of community members, neighbors, pilgrims, or of Tarot [which] is helping you get in touch with yourself and your faithful church members (Santana and Botelho 2019; Kedzior 2013). intuition.” Before reading to oneself or a client, Tarot readers should Bridging these approaches, recent research has examined the “center yourself, ground yourself, and also connect yourself in with experiences of consumers who engage with spiritual market of- the Universal energy that’s around you.” Tarot cards become a de- ferings. Huseman and Eckhardt (2019, 393) introduce the concept vice to “slow down, focus, and reconnect to oneself and to others.” of consumer spirituality, defined as “the interrelated practices and We find that this is the key characteristic of the contemplative con- processes that people engage in when consuming market offerings sumer, who uses Tarot (and other spiritual practices) as a pathfinder (products, services, places) that yield ‘spiritual utility’.” This emer- on the journey to the Higher Self.The second market-making strat- gent stream of research highlights that markets need to be shaped to egy identified is the “The Magician” – empowering consumers for accommodate spiritual consumption (Huseman and Eckhardt 2019; market expansion. The online portal allows consumers to do Tarot 884 / Re-Examining Consumer Identity on Digital Platforms

“their own way.” Belinda reminds listeners that “anyone can do this De Veirman, M., Cauberghe, V., and Hudders, L. (2017), if it’s their calling, and it feels right for them.” Interviewees often “Marketing through Instagram Influencers: The Impact of echo this discourse: “it is your reading practice, and you’re not work- Number of Followers and Product Divergence on Brand ing for an employer, so you don’t have to do it anyone else’s way! Attitude,” International Journal of Advertising, 36(5), 798- […] go your own way, and charge what you’re worth.” By empower- 828. ing Tarot readers, spiritual entrepreneurs work to expand the market, Drenten, J. and L.T. Zayer (2018), “The Role of Digital Virtual offsetting the barriers to entry into the esoteric category (Rinallo and Consumption in Navigating Risk-Laden Life Events,” the Maclaran 2016). Journal of Association for Consumer Research, 3(1), 46-62. Overall, our analysis suggests that the online portal provides Drenten, J., L. Gurreri, and M. Tyler (2018),“Sexualized labour an abundance of resources that materialize these two market-mak- in digital culture: Instagram influencers, porn chic and the ing strategies, as well as provide a template for Tarot consumers to monetization of attention,” Gender, Work, and Organization, reconcile tensions between spirituality and marketization as they 27(1), 41-46. transition toward becoming spiritual entrepreneurs. How consumers Epp, A. M., and L. L. Price (2008), “Family Identity: A Framework respond to spiritual entrepreneurs’ efforts to shape the market should of Identity Interplay in Consumption Practices,” Journal of be the subject of future research. Consumer Research, 35(June), 50-70. Epp, A. M., L. L. Price, and H. J. Schau (2014), “The Role of REFERENCES Brands and Mediating Technologies in Assembling Long- Abbasi, I.S. and N. Alghamdi (2017), “When flirting turns into Distance Family Practices,” Journal of Marketing, 78 (May), infidelity: The Facebook Dilemma,”The American Journal of 81-101. Family Therapy, 45(1), 1-14. Festinger, L. (1954), “A Theory of Social Comparison Processes,” Abidin, C. (2016), “Aren’t These Just Young, Rich Women Doing Human Relations, 7(2), 117-140. Vain Things Online?”: Influencer Selfies as Subversive Gibson-Graham, J. K. (2008), “Diverse economies: performative Frivolity,” Social Media + Society, 2(2), 1-17. practices for other worlds,” Progress in Human Geography, Baccarella, C., T. Wagner, J. Kietzmann, and I. McCarthy (2018), 32(5), 613-632. “Social media? It’s serious! Understanding the dark side of Glaser, B. G. and A.L. Strauss (1967), The Discovery of Grounded social media,” European Management Journal, 36, 431-438. Theory Strategies for Qualitative Research, Chicago: Aldine. Barnhart, M. and L. Peñaloza (2013), “Who Are You Calling Old? Gregory, K. (2016), “In the Cards: From hearing ‘things’ to human Negotiating Old Age Identity in the Elderly Consumption capital” In Katherine Behar (Ed.), Object-Oriented Feminism, Ensemble,” Journal of Consumer Research, 39(6), 1133-53. Minneapolis: University of Minnesota Press. Belk, R. W. (2013), “Extended Self in a Digital World,” Journal of Grinberg, E. (2012), “Aging stylishly, online and in the streets,” Consumer Research, 40 (October), 477-500. CNN, April 28. Bennett, L., S. Wolin and K. McAvity (1988), “Family identity, Gopaldas, A. and A. Siebert (2018), “Women over 40, foreigners ritual, and myth: A cultural perspective on life cycle of color, and other missing persons in globalizing transitions,” In C. Falicov (Ed.), Family Transitions. New mediascapes: understanding marketing images as mirrors of York: Guilford Press. intersectionality,” Consumption Markets & Culture, 21(4), Cabalquinto, E. C. (2018), “Home on the move: negotiating 323-46. differential domesticity in family life at a distance,”Media, Gurrieri, L. and H. Cherrier (2013), “Queering beauty: Culture & Society, 40(6), 795-816. Fatshionistas in the fatosphere,” Qualitative Market Research: Campbell, J. D. (1990), “Self-Esteem and Clarity of the Self- An International Journal, 16(3), 276-295. Concept,” Journal of Personality and Social Psychology, Hagberg, J. (2016), “Agencing practices: a historical exploration 59(3), 538. of shopping bags,” Consumption Markets & Culture, 19(1), Cohen, A.S. (2012), Advanced Style, Brooklyn, NY: Powerhouse 111-132. Books. Hertlein, K. M. (2012), “Digital Dwelling: Technology in Couple Cohen, A.S. and I. Schraer (2013), Advanced Style The Coloring and Family Relationships,” Family Relations, 61(3), 374-387. Book, Brooklyn: Powerhouse Books. Hirschman, E. C., and Stern, B. B. (1994), “Women as Cohen, A.S. and E.J. Feig (2014), Advanced Style, United States: Commodities: Prostitution as Depicted in The Blue Angel, Bond/360. Pretty Baby, and Pretty Woman,” ACR North American Cohen, A. S. (2016), Advanced Style: Older and Wiser, Brooklyn: Advances. Powerhouse Books. Hogg, M. K. and A. Fragou. (2003), “Social Comparison Goals and D’Alessandro, S. and B. Chitty (2011), “Real or relevant beauty? the Consumption of Advertising: Towards a More Contingent Body Shape and Endorser Effects on Brand Attitude and Body View of Young Women’s Consumption of Advertising,” Image,” Psychology and Marketing, 28(8), 843-878. Journal of Marketing Management, 19(7-8), 749-780. Djafarova, E., and Rushworth, C. (2017), “Exploring the credibility Holt, D.B. (2002), “Why Do Brands Cause Trouble? A Dialectical of online celebrities’ Instagram profiles in influencing the Theory of Consumer Culture and Branding,” Journal of purchase decisions of young female users,” Computers in Consumer Research, 29(1), 70-90. Human Behavior, 68, 1-7. Husemann, K. C. and G. Eckhardt (2019), “Consumer spirituality,” DeLanda, M. (2016), Assemblage Theory. Edinburgh: Edinburgh Journal of Marketing Management, 35(5-6), 391-406. University Press. Jermyn, D. (2016), “Pretty past it? Interrogating the post-feminist makeover of ageing, style, and fashion,” Feminist Media Studies, 16(4), 573-89. Advances in Consumer Research (Volume 48) / 885

Kedzior, R. (2013), “Materializing the spiritual. Investigating Santana, J. and D. Botelho (2019), “‘If it comes from Juazeiro, it’s the role of marketplace in creating opportunities for the blessed’! Liquid and solid attachment in systems of object consumption of spiritual experiences,” In D. Rinallo, L. Scott, itineraries of pilgrimages,” Journal of Marketing Management, and P. Maclaran (Eds.), Spirituality and consumption. New 35(5-6), 451-66. York: Routledge, 178-194. Scaraboto, D. and B. Figueiredo (2017), “Holy Mary goes’ round: Kedzior, R. (2015), “Digital Materiality—A Phenomenological Using object circulation to promote hybrid value regimes in Explo- ration,” In K. Diehl and C. Yoon (Eds.), NA—Advances alternative economies,” Journal of Macromarketing,37(2), in Consumer Research, Vol. 43, Duluth, MN: Association for 180-92. Consumer Research, 275-281. Schau, H.J., and M. C. Gilly (2003), “We Are What We Post? Self- Kenney, M. and J. Zysman (2016), “The Rise of the Platform Presentation in Personal Web Space,” Journal of Consumer Economy,” Issues inScience and Technology, 32(3), http:// Research, 30(3), 385-404. issues.org/32-3/the-rise-of-the-platform-economy/ Schau, H.J., M.C. Gilly and M.Wolfinbarger (2009), “Consumer Kozinets, R.V. (2019), Netnography: The Essential Guide to Identity Renaissance: The Resurgence of Identity-Inspired Qualitative Social Media Research, London: Sage. Consumption in Retirement,” Journal of Consumer Research, Kozinets, R.V., D. Scaraboto and M.-A. Parmentier (2018), 36(2), 255-76. “Evolving netnography: how brand auto-netnography, a Schroeder, J. and Zwick, D. (2004), “Mirrors of Masculinity: netnographic sensibility, and more-than-human netnography Representation and Identity in Advertising Images,” can transform your research,” Journal of Marketing Consumption Markets and Culture, 7(1), 21-52. Management, 34(3-4), 231-42. Smeesters, D., and Mandel, N. (2006), “Positive and Negative La Ferla, R. (2018), “The Glamorous Grandmas of Instagram,” The Media Image Effects on the Self,”Journal of Consumer New York Times, June 20. Research, 32(4), 576-82. Lewis, A., T. Knight, G. Germanov, M.L. Benstead, C.I. Joseph and Stevens, L. (2012), “Feeling Bodies,” in Marketing: Aesthetics, L. Poole (2015), “The Impact of Family Functioning of Social Emotions and Gender. The Marketing Review, 12(2), 141-59. Media Use by Depressed Adolescents: A Qualitative Analysis Strauss, A. L., and Corbin, J. M. (1998), Basics of qualitative of the Family Options Study,” Frontiers in Psychiatry, 6 research: Techniques and procedures for developing grounded (September) https://www.ncbi.nlm.nih.gov/pmc/articles/ theory, Thousand Oaks: Sage. PMC4585291/ Su, H. (2016), “Constant connection as the media condition of love: Molesworth, M. and J. Denegri-Knott (2013), “Digital Virtual where bonds become bondage,” Media, Culture & Society, Consumption as Transformative Space,” in The Routledge 38(2), 232-47. Companion to Digital Consumption, ed. Russell W. Belk and Suddaby, R. (2019), “Spiritual myths of consumption: puritanism, Rosa Llamas, New York: Routledge, 223-34. transcendentalism and the consubstantiation of the American Omnicore Agency Report (2020, January 13), “Pinterest by the consumer,” Journal of Marketing Management, 35 (5-6), 410- Numbers: Stats, Demographics & Fun Facts,” available at: 26. https://www.omnicoreagency.com/pinterest-statistics/. Thompson, C.J. (1997), “Interpreting Consumers: A Hermeneutical Ozanne, J. L., and S. Appau (2019), “Spirits in the marketplace,” Framework for DerivingMarketing Insights from the Texts Journal of Marketing Management, 35(5-6), 451-66. of Consumers’ Consumption Stories,” Journal of Marketing Pounders, K., D. H. Rice and A. Mabry-Flynn. (2017), Research, 34(4), 438-55. “Understanding How Goal-Striving, Goal Orientation, and Tiidenberg, K. (2018), “Visibly ageing femininities: women’s Shame Influence Self-Perceptions After Exposure to Models in visual discourses of being over-40 and over-50 on Instagram,” Advertising,” Psychology and Marketing, 34(5), 538-55. Feminist Media Studies, 18(1), 61-76. Price, L. L., E.J. Arnould and C.F. Curasi (2000), “Older Twigg, J. (2013), Fashion and Age: Dress, the Body and Later Life, Consumers’ Disposition of Special Possessions,” Journal of New York: Bloomsbury Academic. “What makes Pinterest Consumer Research, 27(2), 179-201. different?” https://business.pinterest.com/en, accessed Redden. G. (2016), “Revisiting the spiritual supermarket: does February 10, 2020. the commodification of spirituality necessarily devalue it?,” Wissinger, E. (2015), “#NoFilter: Models, Glamour Labor, and the Culture and Religion, 1-19. Age of the Blink,” Interface, 1(1), 1-20. Richins, M. L. (1991), “Social Comparison and the Idealized Zayer, L. T., K. Sredl, M.-A. Parmentier, and C. Coleman (2012), Images of Advertising,” Journal of Consumer Research, 18(1), “Consumption and Gender Identity in Popular Media: 71-83. Discourses of Domesticity, Authenticity, and Sexuality,” Rinallo, D., L. Scott and P. Maclaran (2013), Consumption and Consumption Markets & Culture, 15(4), 333-57. Spirituality. New York: Routledge. Rindfleish, J. (2005), “Consuming the Self: New Age Spirituality as ‘Social Product’ in Consumer Society,” Consumption, Markets and Culture, 8(4), 343-360. Consumers’ Preferences and Their Inferences about Preferences Under Uncertainty Chair: Randy Yang Gao, New York University, USA

Paper #1: Why (and When) Are Uncertain Price Promotions than its worst payoff, not because they find the prospect inherently More Effective Than Equivalent Sure Discounts? aversive, but because they want to exploit the deal for greater profit. Celia Gaertig, University of Chicago, USA Whereas the certain offer is construed as a regular transaction, and Joseph P. Simmons, University of Pennsylvania, USA market norms prohibit exploitation in such cases, the game-like features of the uncertain offer enable profit-seeking, which gives the Paper #2: A Curious Case of Curiosity: An Integrative Review appearance of risk aversion. The Uncertainty Effect disappears when of Recent and Seemingly Contradictory Findings it is made difficult or impossible to exploit the uncertain offer. Christopher K. Hsee, University of Chicago, USA Together, these four papers enrich our understanding of con- Bowen Ruan, University of Iowa, USA sumer decision-making under uncertainty, through well-powered Paper #3: Consumers Confuse Consensus with Strength of empirical investigations, careful synthesis of existing research, and Preferences novel theoretical insights. We believe this session will prove not only Graham Overton, Bocconi University, Italy intriguing to researchers in consumer judgment and decision making, Joachim Vosgerau, Bocconi University, Italy but also useful to the broader audience of marketing practitioners and Ioannis Evangelidis, ESADE Business School, Spain their customers. Paper #4: Risk Aversion or Profit Seeking? Explaining the Uncertainty Effect Why (and When) Are Uncertain Price Promotions More Randy Y. Gao, New York University, USA Effective Than Equivalent Sure Discounts? Minah H. Jung, New York University, USA EXTENDED ABSTRACT SESSION OVERVIEW What do consumers prefer - a sure discount on a purchase or an Uncertainty is prevalent in real life. Understanding uncertainty X% chance to get the product at an even greater discount? Although better will help improve a wide variety of decisions, for example, the ubiquity of sure discounts might lead one to infer that they are how to pique interests and incentivize customers (for marketers), more effective, recent research on this question suggests the oppo- when to seize on an attractive offer (for consumers), and whether site: Consumers seem to prefer uncertain promotions to equivalent to attend a special session (for conference attendees). Uncertainty is promotions that offer a discount or reward with certainty (Goldsmith also pervasive in the academic literature, with previous researchers and Amir 2010; Mazar, Shampanier, and Ariely 2017; also see Shen, identifying both its costs (Gneezy, List, & Wu, 2006) and benefits Fishbach, and Hsee 2014). In an important paper, Mazar, Shampani- (Goldsmith & Amir, 2010; Shen, Fishbach, & Hsee, 2015). How can er, and Ariely (2017) demonstrated that people are more likely to we reconcile these seemingly paradoxical findings? Moreover, much purchase a product when it comes with a chance to get it for free than of the information consumers encounter daily is inherently uncer- when it comes with a sure discount of equal expected value. This tain and often expressed in probabilities or proportions. How well suggests that marketers should make more use of uncertain promo- do consumers understand these representations? The current session tions than they currently do. brings together four papers that shed new light on how people make In the current project, we offer a theory that more fully explains judgments and decisions under uncertainty, from both empirical and these results, and that predicts that the effect will only emerge when theoretical perspectives. the sure discount feels small. In four studies (N = 6,713), we find that The first paper by Gaertig and Simmons finds that uncertain uncertain price promotions are more effective than equivalent sure price promotions are not always more effective than sure discounts discounts only when those sure discounts feel small. Specifically, we of equal expected value. Specifically, the authors propose that con- find that uncertain promotions are relatively more effective when the sumers are more likely to prefer an uncertain price promotion to a sure discounts are actually smaller (Studies 1 and 2) or when the sure sure discount only when the sure discount feels small. discounts are made to feel smaller by presenting them alongside a The second paper by Hsee and Ruan presents an integrative larger discount (Study 3) or by framing them as a percentage-dis- framework that helps reconcile the apparent discrepancies in the count rather than a dollar amount (Study 4). existing literature regarding the effect of curiosity and uncertainty. In Study 1 (N = 2,302), we asked participants to choose among The authors propose that the resolution of uncertainty and curiosity three hotels, one of which was promoted. We manipulated whether always bring positive utility, regardless of the actual outcome. When the promotion was certain (10% off) or uncertain (10% chance to people are curious and such expected positive resolution utility is get it for free), and whether the price of the promoted hotel was low heavily weighted, people might engage in uncertainty-resolving be- ($48) or high ($480). Participants were more likely to choose the haviors even when the expected outcome utility is negative. promoted hotel when it came with a chance-for-free promotion than The third paper by Overton, Vosgerau, & Evangelidis finds that when it came with a sure discount, b = .057, SE = .021, p = .006, consumers confuse consensus information (e.g., 70% prefer product and when the price was low ($48) than when it was high ($480), b A over product B) with strength of preferences for the products (e.g. = .041, SE = .021, p = .045. However, a significant interaction, b = how much better A is rated than B). Specifically, consumers don’t .084, SE = .041, p = .041, revealed that the difference between the anticipate that an option preferred by fewer people often have ratings two promotion types was only significant for the low-price condi- similar to – and sometimes even higher than – ratings of an option tion (p = .001), but not for the high-price condition (p = .609). Thus, preferred by more people. the chance-for-free promotion only outperformed the sure discount The final paper by Gao and Jung proposes a novel explanation when the sure discount was small ($4.80). Because the percentage for the Uncertainty Effect (Gneezy, List, & Wu, 2006). The authors associated with the promotion was held constant at 10%, this effect argue that people are willing to pay less for an uncertain prospect cannot be explained by consumers’ overweighting of small probabil-

Advances in Consumer Research 886 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 887 ities. In Study 2 (N = 852), we replicated the findings from Study 1 example, some studies suggest that curiosity is “bad”—leading peo- in a within-subjects design, where participants were asked to directly ple to expose themselves to miseries (Hsee & Ruan, 2016; Lau et al., choose between the two types of promotions. 2018; Oosterwijk, 2017), while other studies suggest that curiosity is In Study 3 (N = 1,188), we manipulated how large or small “good”—leading to happiness and joy (Kang et al., 2009; Noordew- a sure discount felt by manipulating the context. Participants were ier & van Dijk, 2017; Ruan et al., 2018; Golman, Loewenstein, Mol- asked to imagine that they were buying a $50 training session, and nar, & Saccardo, 2019). that they could choose between receiving a $5 sure discount or a Closely related to curiosity is the notion of uncertainty, because 10% chance to get it for free. We presented the training session curiosity arises from uncertainty in information. Like recent find- alongside a gym membership which was either priced at $480/year ings on curiosity, recent findings on uncertainty also yield seemingly (with a $48 discount) or $40/month (with a $4 discount). Participants contradictory findings. For example, Gneezy, List and Wu (2006) were more likely to prefer the 10% chance-for-free promotion over find that people dread uncertainty, so much so that they would rather a sure discount of $5 when the context was a $48 discount of the have a dominated certain reward (e.g., $10 for sure) than a dominat- yearly membership price than when the context was a $4 discount of ing uncertain reward (e.g., an even chance at $10 and $20; see also the monthly membership price, b = .229, SE = .108, p = .034. Thus, Newman and Mochon 2012; Simonsohn 2009), while Shen and co- asking consumers to consider the same discount in the context of a authors (Shen, Fishbach & Hsee 2015; Shen, Hsee & Talleon 2018) larger discount increased people’s relative preference for an uncer- find that people embrace uncertainty, so much so that they are more tain promotion. motivated to work for a dominated uncertain reward (e.g., an even In Study 4 (N = 2,371), we tested an additional way to manipu- chance of $5 or $10) than a dominating certain reward (e.g., $10 for late how large a sure discount feels, namely by manipulating whether sure). Other authors also report positive effects of uncertainty (Gold- the sure discount is framed as a percentage or a dollar amount. We smith and Amir 2010; Wilson, Centerbar, Kermer, & Gilbert, 2005). first conducted a pretest which revealed that a sure discount is rat- We propose resolution utility (the experienced utility of resolved as smaller when it is framed as a percentage off than when it is ing uncertainty and the accompanying curiosity) as a unifying theo- framed as a dollar amount (p < .001). In Study 4, we then asked par- retical construct that elucidates and integrates the seemingly dispa- ticipants to imagine that they were buying a product that came with rate findings in the literature. Our theoretical framework builds on one of two types of promotions, either a 10% sure discount or a 10% three central propositions. chance to get it for free, and that they could choose which they would The first proposition is that resolution utility is positive. Put like to receive. We manipulated the price of the product to be low differently, regardless of the outcome that resolves uncertainty, the ($11) or high ($311), and, importantly, we also manipulated whether mere resolution of uncertainty and the accompanying curiosity al- the sure discount was presented as a percentage (i.e., 10% off) or ways yields positive utility. This is consistent with prior research as the equivalent dollar amount (i.e., $1.10/$31.10 off). Participants suggesting that curiosity is “good” and can lead to happiness and joy. were more likely to choose the chance-for-free promotion over the For example, learning the answer of a trivia and thus resolving one’s sure discount when the price was low than when it was high, b = curiosity about it is sufficient to produce positive feelings (Kang et .241, SE = .075, p = .001. More importantly and in line with our pre- al., 2009; Ruan et al., 2018). Learning or experiencing a negative test, participants were also more likely to choose the chance-for-free outcome may feel bad, such as one learning she has lost in a gamble. promotion over the sure discount when the sure discount was framed However, that is because the outcome utility—the overall utility of as a percentage than when it was framed as a dollar amount, b = .523, the actual outcome(s) of an uncertainty—is negative. Above and be- SE = .075, p < .001. The interaction was significant as well,b = .352, yond the negative utility of an outcome, the mere resolution of un- SE = .151, p = .019, such that the framing of the sure discount had certainty may still produce positive resolution utility. a larger effect in the low-price condition. Thus, the results from this Building on positive resolution utility, the second proposition study suggest that we can alter people’s preferences for uncertain concerns uncertainty-resolving behavior. Specifically, we propose promotions simply by changing whether the sure discount is framed that the tendency to engage in (and repeat) a behavior that resolves as a percentage or as a dollar amount. uncertainty and the accompanying curiosity depends on expected Our findings are inconsistent with two leading explanations of resolution utility and expected outcome utility (the expected utility consumers’ preferences for uncertain over certain promotions – di- of the possible outcomes of an uncertainty because it is ex ante). For minishing sensitivity and the overweighting of small probabilities example, when people are presented with a box and need to decide – and suggest that people’s preferences for uncertainty are more whether or not to open it, the expected resolution utility is the ex- strongly tethered to their perceptions of the size of the sure outcome pected pleasure of finding out what is in the box, and the expected than they are to their perceptions of the probability of getting the outcome utility is the expected utility of the possible outcomes of the uncertain reward. opened box. We predict that people would open the box if expected overall utility is positive and would not open it if expected overall A Curious Case of Curiosity: An Integrative Review of utility is negative. Recent and Seemingly Contradictory Findings Our last proposition considers the relative weights of these two utilities and holds that when people are curious, they put more EXTENDED ABSTRACT weight on the expected (positive) resolution utility than the expected Many studies have been published recently on curiosity (e.g., outcome utility (which can be negative, neutral, or positive). There- Golman & Loewenstein, 2018; Hill, Fombelle, & Sirianni, 2016; fore, it is possible that the former may override the latter, leading Hsee & Ruan, 2016; Jepma, Verdonschot, Van Steenbergen, Rom- people to resolve uncertainty even if they expect the outcome utility bouts, & Nieuwenhuis, 2012; Kang et al., 2009; Kobayashi, Rav- to be negative. Indeed, Hsee and Ruan (2016) find that because par- aioli, Baranès, Woodford, & Gottlieb, 2019; Lau, Ozono, Kurato- ticipants were curious about whether a prank pen would give electric mi, Komiya, & Murayama, 2018; Noordewier & van Dijk, 2017; shocks if clicked, they clicked the pen just to resolve their curiosity, Oosterwijk, 2017; Ruan, Hsee, & Lu, 2018). However, the findings even though doing so also put them at risk of being shocked. In this of these studies are disparate or even apparently contradictory. For 888 / Consumers’ Preferences and Their Inferences about Preferences Under Uncertainty case, the expected pleasure of uncertainty resolution overrode the Participants were shown two wines, A and B, and told that both expected pain of receiving an electric shock. wines had been rated by consumers on a scale from 1 to 10. Par- Our framework can also explain why an uncertain incentive ticipants were further told that either 90% preferred A to B, 65% (e.g., 20¢ or a 50¢) can be more motivating than a certain incentive preferred A to B, or were given no additional information (control). (50¢) that dominate the uncertain incentive (Shen et al. 2015; 2018). We then asked participants which was more likely, that A’s average A certain incentive produces only one type of positive utility, name- rating is 4 points higher than B’s or that it is 1 point higher than B’s ly, the utility of receiving the discount (i.e., outcome utility). How- (the latter is the correct answer in all conditions). Correct answers ever, an uncertain incentive yields two types of positive utilities: the were incentivized with a bonus. The incorrect answer of 4 points utility of resolving the uncertainty (i.e., resolution utility) and the higher was chosen 31.5% of the time in the control condition, was utility of receiving the discount (i.e., outcome utility). Although the chosen 51.3% in the 65% consensus condition, and 69.2% of the outcome utility of an uncertain incentive may be lower than that of time in the 90% consensus condition (all chi-squares against control a certain incentive, the resolution utility of the uncertain incentive p < .001), suggesting that consumers incorrectly infer strength of can be greater than their difference in outcome utility, resulting in preferences from consensus. greater overall utility of the uncertain incentive, especially if people Study 2 (N = 400, MTurk, https://aspredicted.org/blind. are curious about it. php?x=zx2y4f) used the same paradigm as Study 1 (wine A was pre- This paper is important both theoretically and practically. Theo- ferred by 60[90]% to wine B). Participants were given three answer retically, it builds on Loewenstein’s (1994) seminal work on curios- options: ratings of wine A and B differ by 0 to 1, by 2 to 3, or by 4 to ity and provides an integrative framework that reconciles seemingly 5 points. Like in Study 1, participants were more likely to incorrectly contradictory recent findings on the topic. Practically, this review is choose one of the larger preference differences (compared to 0 to 1) timely, because the amount of information people seek and receive in both the 60% (74%) and the 90% consensus condition (89%; all nowadays is unprecedented in human history; curiosity—the desire binomial tests against equal choice shares p < .001). for information—is a main driving force in this information age and Study 3 (N = 798, MTurk, https://aspredicted.org/blind. deserves better understanding. php?x=w6xq3q) tested whether consumers are aware that a less preferred product B can still have a higher average rating than the more Consumers Confuse Consensus with Strength of preferred product A. Participants learned that two hotels, A and B, Preferences had been rated on 5-point scales and that 65% of consumers preferred Hotel A to Hotel B. Participants were then asked to choose EXTENDED ABSTRACT which of the below answers were possible and incentivized with a Imagine you are shopping for a new car. You are deciding be- bonus to select the most complete answer (option 3). tween a Honda Accord and a Toyota Camry. Looking online for 1. Average rating of Hotel A = 3.8, average rating of Hotel comparisons of these cars, you find a poll according to which 72% B = 3.2 of consumers prefer the Accord to the Camry. You may infer from 2. Average rating of Hotel A = 3.2, average rating of Hotel this information that consumers like the Accord much better than the B = 3.8 Camry. In fact, such inferences are natural and occur spontaneously 3. Both options 1) and 2) above are possible (Kardes et al. 2004, JCP). Surprisingly, they are not necessarily cor- 4. Neither option 1) or 2) are possible rect. The Accord may have received a higher average rating than the Option 1 was chosen more frequently than the, most complete Camry, but it is in fact more likely that the two cars received very answer, option 3, (43.7% vs 30.3%, binomial test p = .002). In an- similar ratings. It is even possible that the Camry was rated better other condition, we tested whether participants would make the same than the Accord! mistake when inferring consensus information from average ratings. In general, it is more likely for two (average) ratings to be close Contrary to our expectations, the majority of participants failed to each other than far apart. When ratings are made on scales from 1 again to choose the complete answer, similarly to the former condi- to 5, for example, it is much more likely to observe a difference of 1 tion (34.8%, χ2(1) =1.24, p = .266). than of 4 points, because there are many more possible combinations Finally, Study 4 (N = 604, MTurk, https://aspredicted.org/blind. of ratings resulting in a difference of 1 (e.g., 5 & 4, 4 & 3, 3 & 2, etc.) php?x=rp84jg) tested whether strategically displaying consensus than there are for a difference of 4 (5 & 1). This combinatorial reason or strength of preferences information influences consumer choice. holds no matter what proportion prefers one car to the other. Hence, Additionally, we tested if such an effect is driven by anchoring or even if 72% prefer the Accord to the Camry, it is more likely that the confusing consensus with strength of preferences. Participants were two cars received similar than vastly different ratings. told that Hotel A costs $135 and Hotel B costs $120. In the consensus But how could the Camry receive a better average rating conditions, participants were asked “Given that 60[90]% gave Hotel than the Accord? That is possible when the rating distributions are A a higher rating than Hotel B, which hotel would you choose?” In skewed (as is typically the case with online ratings). If the Camry’s the strength of preferences conditions, participants were either told rating distribution is right-skewed and the Accord’s distribution is that Hotel A received a 0.4 [1.4] higher rating than Hotel B on a left-skewed, the Camry can have a higher average rating than the 5-point scale, or that Hotel A received a 400 [1,400] higher rating Accord while a greater proportion of consumers prefer the Accord than Hotel B on a 5000-point scale. The 5000-point scale was includ- to the Camry. ed to test the influence of anchoring. We verified analytically that the We show that consumers are not aware of these regularities, strength of preferences (0.4 or 1.4 on a 5-point scale) represent the which leads them to confusing consensus information (a measure of mean of the consensus information (60% or 90% preferring A to B). preference heterogeneity or preference uncertainty) with strength of For ratings on the 5-point scale, participants were more likely to preferences. choose the more expensive hotel when given consensus rather than In Study 1 (N = 700, MTurk, https://aspredicted.org/blind. strength of preferences information (z = 5.13, p < .001). Though the php?x=c8rj88), we tested whether consumers interpret consensus data are not consistent with anchoring, the effect did not occur on the information beyond what it conveys about strength of preferences. 5000-point scale. Advances in Consumer Research (Volume 48) / 889 Risk Aversion or Profit Seeking? Explaining the Nelson, 2012; Kim, Natter, & Spann, 2009). In Study 4 (N = 1211), Uncertainty Effect participants were asked how much they would be willing to pay to watch a preferred movie, a less preferred movie, or for a coin flip that EXTENDED ABSTRACT gives them a ticket to watch either one of the two movies (randomly Uncertainty abounds in the marketplace. Traditional sweep- chosen). Half of participants in each condition were further told that stakes and lotteries aside, recent years have also witnessed rising the theater uses PWYW pricing, and they could pay any amount they popularity of “mystery boxes”, whose content is unknown to the want. Participants’ WTP for the less preferred movie was lower un- buyer at the time of purchase. How much do consumers value under PWYW (M = $5.50 vs. M = $7.32), p < 0.001. Interestingly, certain offers? The answer appears to be “not much”, according to participants’ WTP for the uncertain movie actually increased under research on the Uncertainty Effect (UE). Gneezy, List, & Wu (2006) PWYW (M = $5.26 vs. M = $4.22), p = 0.001. In fact, while we found that people were willing to pay even less for an uncertain pros- replicated the UE without PWYW ($7.32 vs. $4.22, p < 0.001), we pect than for its worst payoff. On the other hand, uncertain offers did not observe the UE under PWYW ($5.50 vs. $5.26, p = 0.644). have been found to be just as enticing (Goldsmith & Amir, 2010) In Studies 5-6 (N = 1611), we replicated all the results above: under or even more enticing (Shen, Fishbach, & Hsee, 2015) than its best PWYW, WTP for the less preferred movie decreases while WTP for payoff. the uncertain movie increases, ps < 0.001. In neither study did we In the current project, we propose a novel explanation for the find the UE under PWYW (p = 0.567 and p = 0.189). These results UE that may help reconcile these seemingly contradictory findings: are consistent with our theorizing that lower valuation in the classic people are willing to pay less for an uncertain prospect, not because uncertainty paradigm is driven by the perceived market norms. Un- they find the prospect inherently aversive, but because they want to der PWYW, however, such norms shift such that people do not maxi- exploit it by paying little. While a certain offer is construed as a regu- mally exploit under uncertainty even when they can by paying zero. lar purchase, and market norms prohibit excessive exploitation, the In summary, results from six studies challenge the “risk aver- novel and game-like features of an uncertain offer allows or even sion” account of the UE, and support our proposed “profit seeking” encourages people to exploit the offer. In six studies (N = 4430, all account, which is also consistent with the finding that the UE is preregistered on AsPredicted), we found evidence consistent with unique to pricing measures (Moon & Nelson, in press). To the best our proposed explanation. of our knowledge, we are the first to propose such an explanation for In Study 1 (N = 401), we ask a series of straightforward ques- the UE, instead of an aversion to risk (Simonsohn, 2009; Newman & tions to examine people’s perception of the uncertain (vs certain) of- Mochon, 2012), bad deals (Yang, Vosgerau, & Loewenstein, 2013), fers, as well as their considerations when deciding how much to pay. or unexplained transaction features (Mislavsky & Simonsohn, 2018). Participants were randomly assigned to evaluate either a certain offer Our results have both important practical implications and theoreti- (a $50 gift certificate), or an uncertain offer (a coin flip that gives cal contributions, by calling attention to consumers’ construal of a either a $50 gift certificate or a $100 gift certificate). Participants felt transaction during preference elicitation (Fischhoff, Welch, & Fred- the uncertain offer was more like a game and less like a regular pur- erick, 1999). chase than the certain offer,p s < 0.001. More importantly, as we predicted, participants in the uncertain condition also stated that when REFERENCES they considered how much they would be willing to pay, it was less Fischhoff, B., Welch, N., & Frederick, S. (1999). Construal important to be fair to the seller but more important to maximize the processes in preference assessment. Journal of Risk and profit, compared to participants in the certain condition,p s < 0.001. Uncertainty, 19, 139–164. If our “profit seeking” account holds, then we would not ob- Gneezy, A., Gneezy, U., Riener, G., & Nelson, L. D. (2012). serve the UE when the uncertain prospect offers neither room nor Pay-what-you-want, identity, and self-signaling in markets. need to seek profit. We believe that in the classic paradigm, because Proceedings of the National Academy of Sciences, 109(19), the value of the gift cards is fixed, it offers participants an oppor- 7236-7240. tunity to exploit the offer by paying little. We remove opportunity Gneezy, U., List, J. A., & Wu, G. (2006). The uncertainty effect: for exploitation by removing such “guaranteed amount” setup in the When a risky prospect is valued less than its worst possible next two studies. Study 2 (N = 600) asked participants (between- outcome. The Quarterly Journal of Economics, 1283-1309. subjects) how much they would be willing to donate to either a pre- Goldsmith, K., & Amir, O. (2010). Can uncertainty improve ferred charity, a less preferred charity, or an uncertain charity (ran- promotions? Journal of Marketing Research, 47(6), 1070- domly chosen between the two). We found that people were willing 1077. to donate significantly more to an uncertain charity (M = $35.57) Golman, R., & Loewenstein, G. (2018). Information gaps: A than to a less preferred charity (M = $27.30), p = 0.010. We also con- theory of preferences regarding the presence and absence of ceptually replicated the findings in a commercial, non-charity setting information. Decision, 5(3), 143-164. in Study 3 (N = 607): participants were willing to add significantly Golman, R., Loewenstein, G., Molnar, A., & Saccardo, S. (2019). more money to a gift card, if the gift card was from an uncertain The demand for, and avoidance of, information. Working retailer (M = $55.60) than when it was from a less preferred retailer Paper. (M = $42.18), p < 0.001. Taken together, results from Studies 2-3 Hill, K. M., Fombelle, P. W., & Sirianni, N. J. (2016). Shopping suggest that people are not simply averse to uncertainty per se. under the influence of curiosity: How retailers use mystery Finally, in Studies 4-6, we test the moderating role of the salient to drive purchase motivation. Journal of Business Research, exchange norms on consumers’ valuation under uncertainty. If peo- 69(3), 1028-1034. ple pay little for the uncertain offer because they perceive it as salient Hsee, C. K., & Ruan, B. (2016). The Pandora effect: The power and market norms, then they should be less likely to exploit the uncertain peril of curiosity. Psychological science, 27(5), 659-666. offer when it becomes less appropriate to be excessively exploitative. It has been shown that the salient norms under PWYW deviate from the traditional market transaction norms (Gneezy, Gneezy, Riener, & 890 / Consumers’ Preferences and Their Inferences about Preferences Under Uncertainty

Jepma, M., Verdonschot, R. G., Van Steenbergen, H., Rombouts, S. Newman, G. E., & Mochon, D. (2012). Why are lotteries valued A., & Nieuwenhuis, S. (2012). Neural mechanisms underlying less? Multiple tests of a direct risk-aversion mechanism. the induction and relief of perceptual curiosity. Frontiers in Judgment and Decision Making, 7(1), 19-24. behavioral neuroscience, 6, 5. Noordewier, M. K., & van Dijk, E. (2017). Curiosity and time: Kang, M. J., Hsu, M., Krajbich, I. M., Loewenstein, G., McClure, from not knowing to almost knowing. Cognition and Emotion, S. M., Wang, J. T.-y., et al. (2009). The wick in the candle of 31(3), 411-421. learning epistemic curiosity activates reward circuitry and Oosterwijk, S. (2017). Choosing the negative: A behavioral enhances memory. Psychological Science, 20(8), 963-973. demonstration of morbid curiosity. PloS one, 12(7), e0178399. Kardes, F. R., Posavac, S. S., & Cronley, M. L. (2004). Consumer Ruan, B., Hsee, C. K., & Lu, Z. Y. (2018). The Teasing Effect: Inference: A Review of Processes, Bases, and Judgment An Underappreciated Benefit of Creating and Resolving an Contexts. Journal of Consumer Psychology, 14(3), 230-256. Uncertainty. Journal of Marketing Research, 55(4), 556-570. Kim, J.-Y., Natter, M., & Spann, M. (2009). Pay what you Want: A Shen, L., Fishbach, A., & Hsee, C. K. (2015). The Motivating- New Participative Pricing Mechanism. Journal of Marketing, Uncertainty Effect: Uncertainty Increases Resource Investment 73(1), 44-58. in the Process of Reward Pursuit. Journal of Consumer Kobayashi, K., Ravaioli, S., Baranès, A., Woodford, M., & Gottlieb, Research, 41(5), 1301-1315. J. (2019). Diverse motives for human curiosity. Nature Human Shen, L., Hsee, C. K., & Talloen, J. H. (2018). The Fun and Behaviour, 3(6), 581-595. Function of Uncertainty: Uncertain Incentives Reinforce Lau, J. K. L., Ozono, H., Kuratomi, K., Komiya, A., & Murayama, Repetition Decisions. Journal of Consumer Research, 46(1), K. (2018). Hunger for Knowledge: How the Irresistible Lure 69-81. of Curiosity is Generated in the Brain. bioRxiv, 473975. Simonsohn, U. (2009). Direct risk aversion: Evidence from risky Loewenstein, G. (1994). The psychology of curiosity. prospects valued below their worst outcome. Psychological Psychological Bulletin, 116, 75-98. Science, 20(6), 686-692. Mazar, Nina, Kristina Shampanier, and Dan Ariely (2017), “When Wilson, T. D., Centerbar, D. B., Kermer, D. A., & Gilbert, D. T. Retailing and Las Vegas Meet: Probabilistic Free Price (2005). The pleasures of uncertainty: prolonging positive Promotions,” Management Science, 63 (1), 250–66. moods in ways people do not anticipate. Journal of personality Mislavsky, R., & Simonsohn, U. (2018). When risk is weird: and social psychology, 88(1), 5-21. unexplained transaction features lower valuations. Yang, Y., Vosgerau, J., & Loewenstein, G. (2013). Framing Management Science, 64(11), 5395-5404. influences willingness to pay but not willingness to accept. Moon, A., & Nelson, L. D. (In press). The uncertain value of Journal of Marketing Research, 50(6), 725-738. uncertainty: When consumers are unwilling to pay for what they like. Management Science. Advances in Neuroforecasting: Forecasting Consumer and Firm Choice using Neural Data Chairs: Steven D. Shaw, University of Michigan, USA Carolyn Yoon, University of Michigan, USA

Paper #1: Using Neural Data to Improve Forecasts of Market- on whole-brain representations of reward value elicited in response Level Behavior to health-related news articles. This paper highlights the advantages Alexander Genevsky, Erasmus University, The Netherlands of using whole-brain patterns associated with reward valuation in ad- Lester Tong, Stanford University, USA dition to previously used region-specific and self-report predictors of Brian Knutson, Stanford University, USA health information sharing. The third paper in this session proposes a novel method to predict pre-launch product sales by combining Paper #2: Population News Sharing is Reflected in Distributed retailer data about competitors with information from traditional sur- Reward-Related Brain Activity veys and fMRI data, and gauges the added value of each information Bruce P. Doré , University of Pennsylvania, USA source (market data, survey, fMRI). Findings illustrate what manag- Christin Scholz, University of Amsterdam, The Netherlands ers can learn and the key benefits from running fMRI studies and/ Elisa C. Baek, University of California, Los Angeles, USA or traditional surveys before new product launch. The fourth paper Emily B. Falk, University of Pennsylvania, USA in this session aims to test neuroforecasting in a novel market, using Paper #3: Top or Flop: Quantifying the Value Of fMRI Data in online dating profiles, and attempts to differentiate between how af- the Prediction of Success of New Products fective and informational components of choice can forecast market Marton Varga, INSEAD, France level outcomes. Anita Tusche, Queen’s University, Canada In sum, the papers in this special session address the following Paulo Albuquerque, INSEAD, France research questions: 1) how do behavioral and neural measures per- Nadine Gier, Heinrich Heine University, Germany form when the sample of participants is more or less representative Bernd Weber, University of Bonn, Germany of the market?, 2) can neuroforecasting prediction be improved upon Hilke Plassmann, INSEAD, France by expanding from previous region-based neural approaches to the Paper #4: Neuroforecasting Aggregate Choice in Online Dating: use of multivariate brain patterns?, 3) what is the value of neurofore- Predicting Aggregate Choices From Small Samples Using Neural casting to managers attempting to forecast sales?, and 4) do affective and Behavioral Measures and informational components of stimuli contribute differently to Steven D. Shaw, University of Michigan, USA neuroforecasts and can we forecast online dating decisions? Vinod Venkatraman, Temple University, USA Carolyn Yoon, University of Michigan, USA Using Neural Data to Improve Forecasts of Market-Level Behavior SESSION OVERVIEW The nascent field of consumer neuroscience holds great promise EXTENDED ABSTRACT for generating new insights and theory in consumer behavior. Re- cently, evidence has emerged indicating that neural activity in the Objective brain’s valuation system may be used to forecast market-level be- Consequential decisions in business and public policy are often havior of companies, products, and advertising performance. This based on forecasts of population-level behavior. Any improvement in phenomenon, called ‘neuroforecasting’, is the use of brain activity our ability to accurately predict aggregate-level responses will maxi- from a small sample of individuals to forecast choices or prefer- mize the likelihood of product success and minimize waste of valu- ences of a separate, large population group of individuals (Knutson able time and resources. Motivated by theoretical challenges from & Genevsky, 2018; Falk & Scholz, 2018). At present, a number of economics (Bernheim 2008) and the promise of practical applica- neuroforecasting studies have illustrated this phenomenon, including tions (Smidts et al. 2014), researchers have begun to explore whether for music album sales (Berns & Moore, 2012), calls to a smoking brain activity in laboratory samples can forecast aggregate choice cessation helpline (Falk, Berkman, and Lieberman, 2012), various ad and how neural measures compare with more traditional measures metrics (Venkatraman et al., 2015; Falk et al., 2016), news sharing (e.g., ratings, choices; Venkatraman et al. 2015). Although a grow- (Scholz et al., 2017), online media consumption (Tong et al., 2020), ing literature demonstrating the capacity of neural data to forecast and microlending (Genevsky & Knutson, 2015). However, many aggregate level behavior (see Knutson and Genevsky 2018)brain ac- open research questions surrounding neuroforecasting remain. This tivations were recorded while smokers viewed three different televi- special session brings together work by leading researchers in this sion campaigns promoting the National Cancer Institute’s telephone field to expand on our understanding of neuroforecasting. hotline to help smokers quit (1-800-QUIT-NOW, no research has The papers in this session aim to expand on our understanding explored the mechanisms that account for these out-of-sample pre- of neuroforecasting. In the first paper, authors examine the role of dictions. generalizable and idiosyncratic choice processes by exploring how Despite evidence to support neuroforecasting, important ques- forecasts based on behavioral and neural measures are differentially tions remain regarding how individual choice processes scale to in- impacted by the representativeness of the study sample to the mar- form forecasts of aggregate choice. Collected evidence from the ex- ket population, and show that when compared to behavioral mea- isting neuroforecasting literature suggests a “partial scaling” account, sures, neural data is less impacted by the representativeness of the in which some choice components may forecast aggregate behavior sample. This suggests that neural activity may be a more generaliz- more accurately than others (and by extension, more accurately than able index of preference across individuals than self-report measures individual choice itself, which constitutes the integration of all sup- or observed behavior. In the second paper, authors seek to build a porting components). Specifically, with respect to fMRI markers, ac- model of neural predictors of health information diffusion, focusing tivity associated with positively aroused affect in the ventral striatum

(including the NAcc), loss anticipation in the anterior Insula (AIns), market, only neural activity was significantly associated with the and value integration in the medial prefrontal cortex (MPFC). Thus, market preferences (coef. = .656, SE = .217, p < .01). we predicted that primary neural responses indexed by striatal and insular activity may generalize more across individuals and thus Discussion forecast aggregate behavior more accurately than neural regions as- We find that when compared to behavioral measures, neural sociated with idiosyncratic value integration (e.g., MPFC). Further, data is less impacted by the representativeness of the sample. These these basic neural responses may also forecast behavior more accu- data suggest that neural activity may be a more generalizable index rately than individual choice itself, which by definition incorporates of preference across individuals than self-report measures or ob- these generalizable and idiosyncratic neural signals. In this paper, we served behavior. In fact, for non-representative samples, only neural examine the role of generalizable and idiosyncratic choice processes activity, and not behavioral measures, were significant predictors of by exploring how forecasts based on behavioral and neural measures market preference. In a representative sample, both behavioral mea- are differentially impacted by the representativeness of the study sures and neural activity accounted for significant and independent sample to the market population. variance in market preference. These data provide clear and practical implications regarding Methods the use cases in which neural measures can provide value to market- In phase one, participants in the scanner were presented with ers. First, in situations in which representative samples are difficult projects from a popular online crowdfunding platform and were to obtain, due to lack of information about the target audience or an asked to make incentive compatible funding decisions. Neural activ- inability to find suitable participants, neural measures may provide ity was collected while participants viewed and then made decisions the best option for accurately forecasting market preference. Second, regarding these projects. In phase two, a large internet sample (n= even when a representative sample can be established, neural mea- 3000) was asked to make similar funding choices regarding the same sures can add significant predictive value to models of market pref- projects presented during the imaging phase. In addition to the fund- erence. Further, this work is the first attempt to explore the neural ing choices, demographic characteristics were also collected about and psychological mechanisms that underlie and account for neural the internet participants. The laboratory sample’s behavior and neu- prediction of market behavior and represents a significant conceptual ral activity were then used to forecast the larger internet market’s advance in our understanding how choice scales from individual to preferences. Using traditional demographic variables, we construct aggregate levels. two independent marketplaces; one more and one less well represented by our laboratory sample. Forecasts of the project preferences Population News Sharing is Reflected in Distributed in the two marketplaces were calculated independently and then con- Reward-Related Brain Activity trasted. Based on previous work, neural predictions focus on activity in primary affective (i.e., nucleus accumbens, insula) and value EXTENDED ABSTRACT integration (i.e. MPFC) regions. We hypothesized that in the well Information that diffuses widely in the media environment can represented market both behavioral and neural measures would dem- influence the behavior of individuals and shape broader directions onstrate significant prediction accuracies. However, in the less well of societal change. Previous studies have shown that information in represented market we will observe a much steeper decline in accu- targeted brain regions can be used to predict this diffusion of in- racy for behavioral prediction and a relatively stable prediction rate formation. However, by focusing on these targeted regions, the ap- for the neural measures. This finding would support the hypothesis proach in previous work has discarded information from the majority that primary neural components of choice generalize more than oth- of the brain. Here, we sought to build a model of neural predictors ers, and thus forecast out-of-sample aggregate behavior better than of health information diffusion, focusing on whole-brain represen- individual choice itself. tations of reward value elicited in response to health-related news articles. Beyond enriching our scientific understanding of how and Results why health information spreads throughout a population, models of We first explored the association between the laboratory mea- this kind could ultimately be used to forecast and enhance the impact sures and market preferences independently for the representative of communications at population scale. and non-representative marketplaces. Using the laboratory sample’s Current neurobiological models posit that the functional neu- behavior, we find a significant correlation with the preferences in roanatomy of reward value extends beyond the core striatal and the representative marketplace (r = .375, p < .05) but not in the non- ventromedial prefrontal (vmPFC) regions that have received empiri- representative marketplace (r = .249, n.s.). This finding indicates that cal attention in previous population prediction studies. These mod- the ability of a sample to forecast market-level behavior is dependent els propose that diverse brain systems interact to rapidly propagate on the representativeness of the sample. However, when we used reward-related information throughout the brain, generating a dis- neural activity from the lab sample we find significant, and nearly tributed value representation that directs cognition and behavior in identical, correlations with preferences in the representative (r = a multi-faceted manner. However, it is unclear whether distributed .470, p < .01) and non-representative marketplaces (r = .447, p < brain representations of reward value hold information that can be .01). This suggests that neural prediction of aggregate behavior is used to predict the impact of messaging, beyond responses in core less influenced by the representativeness of the sample. striatal and vmPFC regions. Regression analyses further explored these relationships. Lin- We sought to address this gap in knowledge with two neuro- ear regression models including both behavioral and neural predic- imaging studies that quantified functional brain responses to New tors of preferences in the representative market indicated that both York Times health news articles and used these responses to predict lab behavior (coef. = .319, SE = .151, p < .05) and neural activity sharing of these articles in the broader population of readers. In par- (coef. = .613, SE = .211, p < .01) accounted for significant and inde- ticular, we addressed two specific questions. First, does expression pendent variance in the market. However, in the non-representative of a meta-analytically defined whole-brain reward valuation-related pattern predict population-level information sharing? Second, does Advances in Consumer Research (Volume 48) / 893 expression of this pattern improve prediction of population-level in- ficacy of the pattern was not reducible to patterns of activity within formation sharing beyond what can be predicted from activity within core brain reward regions but rather depended on larger-scale pat- reward-related brain regions and self-reports? terns of activity distributed widely across cortical, subcortical, and brainstem systems. These findings highlight the advantages of using Results whole-brain patterns associated with reward valuation in addition to Our initial question was whether expression of the meta-an- previously used region-specific and self-report predictors of health alytically defined reward-value related pattern was predictive of information sharing. population-level article sharing (collected via the New York Times If the brain holds information that can be used to forecast large- website). From the perspective of practitioners seeking to explain scale behavior, how do we best characterize and model this activity? variance in the population-level success of messaging, it is important Here we suggest that, in addition to summed activity within core to characterize the predictive efficacy of our models in terms of how value-related brain regions, it is useful for neural prediction models well they explain article-to-article variance in population sharing. to incorporate information about distributed brain representations of Therefore, we also estimated the relationship between the average value. Overall, this work contributes to our growing ability to use expression of the meta-analytically defined reward value pattern for neural data to forecast out-of-sample outcomes and augments our each article (averaged across all the perceivers in a study) and the understanding of the brain mechanisms that underlie how informa- population sharing of those articles. This indicated that the standard- tion diffuses (or fails to diffuse) across a population of individuals. ized article-level relationship between pattern expression and population article sharing was β = .33, 95%CI[.12, 54] in Study 1 and β = Top or Flop: Quantifying the Value of fMRI Data in the .41, 95%CI[.21, .62] in Study 2. Prediction of Success of New Products In a next step, we sought to understand the extent to which a model incorporating expression of the reward-related pattern showed EXTENDED ABSTRACT increased predictive accuracy relative to reduced models including Managers are keen to obtain the best possible forecast about only activity in a reward-related region of interest (ROI) and subjec- future sales, so that they can plan with operations ahead and adjust tive ratings of the articles. First, we fit models that controlled for marketing policies if necessary. This is especially the case for new activity within the reward-related region (spanning vmPFC and ven- products for which, by definition, no previous sales information is tral striatum), and self-reports of article value (reading intentions in available. Some innovations might turn out to be best-sellers, ex- Study 1, sharing intentions in Study 2), finding that expression of the ceeding previous expectations. If managers anticipate such future reward value pattern predicted population article sharing above and success, they can previously plan to avoid stock-out, for instance. beyond these other predictors in both Study 1, β = .07, 95%CI[.01, Other innovations might become clear failures, which can even en- .14], and in Study 2, β = .12, 95%CI[.04, .21]. danger the reputation of the company. Having a good forecast about As above, in order to understand the practical value of each the fiasco can support managers in their decision whether to launch predictor in forecasting out-of-sample sharing, we estimated the the product at all. variance explained by our brain and self-report predictors when ag- We propose a novel method to predict pre-launch product sales gregating data from all perceivers up to the article level. Combining by combining retailer data about competitors with information from the data from Studies 1 and 2, the aggregate article-level R2 was .18, traditional, representative surveys in which participants were asked 95%CI[.05, .30], for a model 1 including only subjective ratings of to give their views about existing and new products as well. Impor- the articles, .24, 95%CI[.10, .36], for a model 2 including subjective tantly, we further add functional magnetic resonance imaging (fMRI) ratings and reward value-related ROI activity, and .34, 95%CI[.19, data to the model, in which we measure the brain activity of partici- .46], for a model 3 including subjective ratings, reward value ROI pants during survey questions. activity, and reward value pattern expression.To estimate the out-of- Since the seminal work of Bass (1969), various authors have sample predictive accuracy of these linear models, we used Bayesian studied the adoption of new products (Lee et al., 2003; Neelamegham leave-one-out (LOO) cross validation to derive LOO-adjusted devi- and Chintagunta, 1999; Lenk and Rao, 1990). Traditionally, such ance values (LOOIC) reflecting expected out-of-sample predictive studies relied on either revealed or stated preference data, but more error. The model including ratings, reward value ROI, and reward recently some utilized a combination of these two (Helveston et al., value pattern expression showed improved predictive fit relative to 2018; Hardt et al., 2017; Phaneuf et al., 2013; Morikawa et al., 2002; model 1 including only ratings, ∆LOOICm3-m1 = -12.7, SE=7.3, and Harris et al., 1999). As the standard goal of this field is to make in- relative to model 2 including ratings and reward value ROI activ- ference about the real-world market, choice experiments have been ity, ∆LOOICm2-m1 = -8.9, SE=5.6. Overall, these results indicate that improved to better mirror the outcomes when consumers face the the expression of the reward value-related pattern substantially im- purchase decision in the store as opposed to the lab (Gilbride et al., proved accuracy in predicting population-level article sharing, with 2008; Allenby et al., 2005; Wertenbroch and Skiera, 2002). a full model reaching on the order of one-third of the variance ex- Recently, a growing body of literature established the link plained. between the activity in certain areas of the brain and real-world monetary outcomes. Kühn et al. (2016) found that their fMRI data Discussion correlated with supermarket sales of existing products. fMRI also We used neuroimaging to ask whether whole-brain response explained a large part of the variance in real-world advertising elas- to health-relevant news articles could predict large-scale, out of ticities (Venkatraman et al., 2015) beyond stated preference survey, sample, sharing of those articles, beyond previously identified brain of market funding outcomes (Genevsky et al., 2017), and of the suc- and self-report predictors. Our results indicated that expression of cess of microloan requests (Genevsky and Knutson, 2017). Given a distributed pattern of brain activity meta-analytically associated the explanatory power of the recorded brain images, we believe that with reward valuation substantially improved accuracy in predicting incorporating fMRI into an economic model that builds on pooling population sharing of the health news articles, beyond previously revealed and stated preference data will lead to more accurate sales identified brain and self-report predictors. Further, the predictive ef- forecasts. 894 / Advances in Neuroforecasting: Forecasting Consumer and Firm Choice using Neural Data

We obtained data from a large German supermarket chain about a small group of participants (n=~30-50) are shown unfamiliar, nov- the weekly sales of 19 innovative grocery products launched within el, real-world stimuli, and their neural activity data is used to predict the last two years. We have information on the weekly quantity sold the real-world outcomes of the stimuli once the market has matured from each new product (at least in its first year), and from two or (several months or years later). Such neuroforecasts have been re- more competitors (at least in their last year before the new prod- ported for song downloads (Berns & Moore, 2012), call-back rates uct launch). In each store of the retailer the management can decide (Falk, Berkman, and Lieberman, 2012), various ad metrics (Ven- whether to offer a certain product at a given week and has some flex- katraman et al., 2015; Kuhn, Strelow, and Gallinat, 2016; Falk et ibility in price setting as well. al., 2016), and loan/funding appeals (Genevsky and Knutson, 2015; Survey participants representative of those who typically shop Genevsky, Yoon, and Knutson, 2017). In nearly all of these studies, at our retailer indicated their preference for some of the innovations the NAcc and MPFC (separately or together) were the key contribu- and their respective competitors. Participants first saw pictures of the tors to forecasting aggregate level choices. However, exactly how products without price information and were asked to indicate how these two brain areas contribute to aggregate choice prediction, their much they like the product and its packaging on a 7-point Likert- prediction weights or under which conditions each/either is effective, scale. Thereafter, they were asked to indicate their purchase intention is not yet clear. at the price recommended by the retailer on a 4-point Likert-scale. Knutson & Genevsky (2018) propose that the AIM model (Sa- Furthermore, we recruited 44 participants in order to measure manez-Larkin & Knutson, 2015) could provide a useful framework their brain activity while answering the questions of the same survey. for understanding neuroforecasting. Specifically, affective compo- In addition, these participants were asked to indicate their purchase nents of individual choice may be most useful for aggregate fore- intention on a 4-point Likert-scale in an incentive compatible ex- casting, whereas neural regions associated with integrating affective periment in which they had monetary incentives to reveal their true information into one’s goals and context may be more predictive of preferences. The brain activity of the participants was measured us- individual choices (captured by brain activity in the NAcc and vMP- ing fMRI. Consequently, we have variables reflecting the activity of FC, respectively). To test the AIM model’s efficacy for explaining three pre-defined regions of interest (i.e., the vmPFC, vStr and an- neuroforecasting aggregate choice, we will attempt to forecast the terior Insula) brain areas at the time when the product is viewed and success of online dating profiles. when the purchase intention is stated. Online dating profiles provide a suitable context for this re- We develop a structural economic model, in which consumers search question because: 1) online dating profiles have both affective form preferences about existing and innovative products and make (e.g., facial image) and informational (e.g., occupation or personali- purchase decisions accordingly. Consumer choice is modeled via the ty) components, and 2) online dating is a novel market for neurofore- utility function, which incorporates covariates from all three data casting. To this end, we created 36 standardized dating profiles using sources. We derive the likelihood of the stated preference data using faces from the Chicago Face Database (Ma, Correll, & Wittenbrink, a Multivariate-Ordered-Probit framework, while for the aggregated 2015). Profiles were orthogonalized using a 3 x 2 x 2 x 3 factorial de- market data using a market-share based approach. These models are sign: attractiveness (high, medium, low), age (19-23, 24-28), facial connected via a set of common parameters that measure the latent expression (neutral, smiling), and profile description (hobbies/likes, attractiveness of a particular product. We combine the three likeli- SES/occupation, and personality traits). During the task, participants hoods to form a joint objective function to maximize. were randomly shown all 36 profiles (within-subjects), made a bi- Using the estimated utility-based coefficients we forecast the nary ‘like’ or ‘pass’ choice for each profile, and rated each profile on sales and the number of adopting retailers of the innovative products attractiveness, career prospects, likability of personality, and likeli- before there are launched. We repeat these sales forecasts with using hood that the individual in the profile will ‘like’ them back. only a subset of the datasets (market data, survey, fMRI), so that we Market level descriptive statistics revealed that participants rate can gauge the added value of each information source. With our ap- profiles with a smiling face higher on positive personality charac- proach, manager can learn the benefits of running fMRI studies and/ teristics and believe the individuals in these profiles will be more or traditional surveys before new product launch. willing to ‘like’ them back. Additionally, SES/occupation indicators in a profile description led to participants rating the profiles as Neuroforecasting Aggregate Choice in Online Dating: having better career prospects. Factorial design attractiveness bins Predicting Aggregate Choices From Small Samples Using correlated highly with participants’ attractiveness ratings of profiles. Neural and Behavioral Measures Regression results indicated that aggregate in-lab sample (n = 45) reported levels of attractiveness, career prospects, personality, and EXTENDED ABSTRACT likelihood of ‘like’ back of profiles explained a significant amount The ability to predict aggregate, market level choices from a of variance in profile choice likelihood within sample (F(4, 31) = 2 small sample of individuals can provide tremendous value in a va- 45.21, p < .01, R Adj = 0.83) and out of sample/at the market-level 2 riety of domains (e.g., product success, political elections). In this (n = 273; F(4, 31) = 15.82, p < .01, R Adj = 0.63). Training a model paper, we advance our understanding of decision-making and choice with these same variables to predict the aggregate choice likelihood by assessing behavioral and neural approaches to forecasting aggre- of randomly selected holdout profiles (75/25 train/test; 50 iterations) gate choices in an online dating context both within a small sample resulted in correct choice likelihood prediction (+/- 5% of actual (n < 50) and in an independent simulated market sample (n > 250). In choice likelihood) 43.5% of the time within sample, and 33.0% of accordance with the affective-integration-motivation (AIM) frame- the time for the market population. work, we propose that affective neural components of individual In accordance with the AIM model, we hypothesize: H1: Af- choice are most useful for aggregate forecasting, whereas neural fective neural components, captured by activity in the NAcc, will components associated with the integration of information are most be most highly correlated with perceived attractiveness and facial useful for individual-level prediction. expression, and more likely to forecast aggregate choice. H2: Neu- At present, a small number of studies have provided a proof of ral activity associated with integration, captured by activity in the concept for neuroforecasting aggregate choice. In these experiments, medial prefrontal cortex, will be most highly correlated with per- Advances in Consumer Research (Volume 48) / 895 ceptions of career prospects and personality and will be most useful Knutson, B., & Genevsky, A. (2018). Neuroforecasting Aggregate for predicting choices at the individual-level. Market-level data have Choice. Current Directions in Psychological Science, 4, been collected; fMRI scanning has been postponed due to the CO- 096372141773787–6. VID-19 pandemic. Kühn, S., Strelow, E., & Gallinat, J. (2016). Multiple “buy buttons” in the brain: Forecasting chocolate sales at point-of-sale based REFERENCES on functional brain activation using fMRI. NeuroImage, 136, Allenby, G., Fennell, G., Huber, J., Eagle, T., Gilbride, T., Horsky, 122-128. D., et al. (2005). Adjusting choice models to better predict Lee, H., Smith, K. G., & Grimm, C. M. (2003). The effect of market behavior. Marketing Letters, 16(3-4), 197-208. new product radicality and scope on the extent and speed of Bass, F. M. (1969). A new product growth for model consumer innovation diffusion.Journal of Management, 29(5), 753-768. durables. Management science, 15(5), 215-227. Lenk, P. J., & Rao, A. G. (1990). New models from old: Forecasting Baek, E. C., Scholz, C., O’Donnell, M. B., & Falk, E. B. (2017). product adoption by hierarchical Bayes procedures. Marketing The Value of Sharing Information: A Neural Account of Science, 9(1), 42-53. Information Transmission. Psychological Science, 28(7), Ma, D. S., Correll, J., & Wittenbrink, B. (2015). The Chicago 851–861. face database: A free stimulus set of faces and norming data. Bartra, O., McGuire, J. T., & Kable, J. W. (2013). The valuation Behavior Research Methods, 47(4), 1122–1135. system: a coordinate-based meta-analysis of BOLD fMRI Morikawa, T., Ben-Akiva, M., & McFadden, D. (2002). Discrete experiments examining neural correlates of subjective value. choice models incorporating revealed preferences and NeuroImage, 76, 412–427. psychometric data. Advances in Econometrics, 16, 29-56. Berns, G. S., & Moore, S. E. (2012). A neural predictor of cultural Neelamegham, R., & Chintagunta, P. (1999). A Bayesian model popularity. Journal of Consumer Psychology, 22, 154–160. to forecast new product performance in domestic and Falk, E., & Scholz, C. (2018). Persuasion, influence, and value: international markets. Marketing Science, 18(2), 115-136. Perspectives from communication and social neuroscience. Phaneuf, D. J., Taylor, L. O., & Braden, J. B. (2013). Combining Annual Review of Psychology, 69, 329–356. revealed and stated preference data to estimate preferences Falk, E. B., Berkman, E. T., & Lieberman, M. D. (2012). From for residential amenities: A GMM approach. Land Economics, neural responses to population behavior: Neural focus group 89(1), 30-52. predicts population-level media effects.Psychological Science, Samanez-Larkin, G. R., & Knutson, B. (2015). Decision-making 23, 439–445. in the ageing brain: Changes in affective and motivational Falk, E. B., O’Donnell, M. B., Tompson, S., Gonzalez, R., circuits. Nature Reviews Neuroscience, 16, 278–289. et al. (2016). Functional brain imaging predicts public Scholz, C., Baek, E. C., O’Donnell, M. B., Kim, H. S., Cappella, health campaign success, Social Cognitive and Affective J. N., & Falk, E. B. (2017). A neural model of valuation and Neuroscience, 11(2), 204–214. information virality. Proceedings of the National Academy of Genevsky, A., & Knutson, B. (2015). Neural affective mechanisms Sciences of the United States of America, 114(11), 2881–2886. predict market-level microlending. Psychological Science, 26, Schultz, W. (2015). Neuronal Reward and Decision Signals: From 1411–1422. Theories to Data. Physiological Reviews, 95(3), 853–951. Genevsky, A., Yoon, C., & Knutson, B. (2017). When brain beats Tong, L. C., Yavuz Acikalin, M., Genevsky, A., Shiv, B., & behavior: Neuroforecasting crowdfunding outcomes. Journal Knutson, B. (2020). Brain activity forecasts video engagement of Neuroscience, 37(36), 8625-8634. in an internet attention market. Proceedings of the National Gilbride, T. J., Lenk, P. J., & Brazell, J. D. (2008). Market share Academy of Sciences of the United States of America, 117(12), constraints and the loss function in choice-based conjoint 6936–6941. analysis. Marketing Science, 27(6), 995-1011. Venkatraman, V., Dimoka, A., Pavlou, P. A., Vo, K., Hampton, W., Hardt, N., Kim, Y., Joo, M., Kim, J., & Allenby, G. M. (2017). Bollinger, B., . . . Winer, R. S. (2015). Predicting advertising Reconciling Stated and Revealed Preferences. Working paper success beyond traditional measures: New insights from Harris, K. M., & Keane, M. P. (1998). A model of health plan neurophysiological methods and market response modeling. choice: Inferring preferences and perceptions from a Journal of Marketing Research, 52, 436–452. combination of revealed preference and attitudinal data. Vickery, T. J., Chun, M. M., & Lee, D. (2011). Ubiquity and Journal of Econometrics, 89(1-2), 131-157. specificity of reinforcement signals throughout the human Helveston, J. P., Feit, E. M., & Michalek, J. J. (2018). Pooling brain. Neuron, 72(1), 166–177. stated and revealed preference data in the presence Wertenbroch, K., & Skiera, B. (2002). Measuring consumers’ of RP endogeneity. Transportation Research Part B: willingness to pay at the point of purchase. Journal of Methodological, 109, 70-89. Marketing Research, 39(2), 228-241. Sharing and Consumer Experience: How Does Different Forms of Sharing Affect Consumer Evaluations of Experience Chairs: Der-Wei Huang, Indiana University, USA Shanker Krishnan, Indiana University, USA

Paper #1: The Paradox of Social Interaction in the Sharing on social media. The authors’ findings show that when consumers see Economy photos taken from the actor’s perspective (vs. observer’s perspec- Dafna Goor, Harvard University, USA tive), consumers generate more favorable attitude toward the experi- Amir Grinstein, Northeastern University, USA ence of an event, and they are more likely to share this experience Nailya Ordabayeva, Boston College, USA on social media. Meike H. Morren, Vrije University, The Netherlands The fourth paper demonstrates that the anticipated reaction from the audience when sharing online has an important impact on Paper #2: The Meme Economy: How Internet Memes Impact experience enjoyment. Moving away from the belief that known oth- Consumption ers don’t judge as harshly as unknown others, the paper shows that Nicole Kim, University of Maryland, USA when consumers share pictures of indulgence foods online, they an- Jared Watson, New York University, USA ticipate less positive judgements from people they know (vs. mostly Paper #3: Let Me See How I Look: How Visual Perspective don’t know) and enjoy less the food after sharing the picture. Affects Consumer Experience and Sharing Behaviors Together these four papers examine how consumers behave in Der-Wei Huang, Indiana University, USA a sharing context, whether it being offline (in a sharing economy) or Shanker Krishnan, Indiana University, USA online (on social media), and investigate the underlying mechanisms. Paper #4: Indulging Online: A Moral Dilemma The fours papers contribute to the literature in sharing economy, Marie Ozanne, Cornell University, USA emotions, and consumer decisions and judgement. Anna Mattila, Pennsylvania State University, USA The Paradox of Social Interaction in the Sharing SESSION OVERVIEW Economy Sharing is an important part of consumers’ consumption experiences and decision-making. Interestingly, there are different ways EXTENDED ABSTRACT in which consumers can share. Consumers may share resources that The sharing economy has been rising in the past decade. Its key are intangible, such as information and experiences, or consumers sectors are expected to grow in revenue from $15 billion in 2014 to may share products and physical goods, such as taxi rides and hos- $335 billion in 2025 (PwC 2015). However, little is known about us- tels. Existing research has looked at how sharing experiences could ers’ experience on peer-to-peer platforms in the sharing economy or affect consumer memories, enjoyment, and perceptions of morality the relationship between the sharing economy and wellbeing (Eck- (e.g., Barasch, Zauberman, and Diehl, 2018; Campbell and Wint- hardt et al. 2019). Rather, prior research focused on the economic erich, 2018; Tamir et al., 2018). Advancing the existing findings, in value of the sharing economy (Eckhardt and Bardhi 2015; Neoh et the current session, we bring together four papers that examine how al. 2017). Yet, our pilot study (N=148, Mturk) shows that consum- consumer experiences and decision-making are influenced by differ- ers’ social, in addition to financial, motivations to enter the sharing ent modes of sharing. economy are greater than other motivations (e.g., environmental, The first paper examines how sharing products and services af- cultural; p’s<.001; see also Milanova and Maas 2017). Furthermore, fects consumer experience in a shared economy. In particular, the our analysis of ads of ten leading sharing economy brands (N=79, authors show that despite the rising demand for the sharing economy broadcasted between 2016-2018) reveals that companies capitalize that is also driven by social motivation, consumers rate their expe- on social drivers (68.4% of the ads used social interaction narratives rience in the sharing economy lower, compared to the traditional and 89.9% community narratives). economy (e.g., hostel vs. Airbnb, taxi vs. uber, restaurant vs. shared Building on these preliminary findings on the substantial role of dining). This paper also demonstrates that such sharing is caused by social interaction in the sharing economy, we suggest that consump- social anxiety and increase in awareness to social interactions. tion experiences in the sharing economy may elicit greater social The second paper examines a novel mode, internet ‘memes,’ awareness in consumers than parallel experiences in the traditional that consumers use to share information on social media. Specifi- economy (e.g., Uber vs. taxi, Airbnb vs. hostel, Eatwith vs. restau- cally, the authors examine how consumers make inferences about rant). Consequently, focus on social interaction may paradoxically a target content referenced in a meme, and that it can impact subse- decrease consumer satisfaction and even translate into lower ratings. quent consumption decisions. Specifically, memes, compared to oth- Five studies use mixed methodologies, combining lab experi- er social media posts in non-meme format, can increase consumption ments and archival data from a leading sharing economy company, attitudes toward the target content. This is because the use of memes to test our predictions. increase the perception that the target content has gone viral, invok- Study 1 tested the effect of the sharing (vs. traditional) economy ing consumers’ FOMO (i.e., fear of missing out), thereby increasing on consumer satisfaction. Participants (N=167, Mage=33.6, 38% fe- consumption intentions. male, MTurk) imagined they went on a one-week trip. In the shar- The third and fourth papers investigate how sharing contexts ing (vs. traditional) economy condition, they imagined staying at an or sharing behavior affect the consumption experience and some in- Airbnb (vs. YMCA) apartment, where they got a private room but vestigate when we are more likely to share. The third paper dem- shared house amenities with other people. Participants rated the shar- onstrates how visual perspective of photos of an experience affect ing economy experience as less satisfying (M=3.35) than the tradi- consumer attitude toward an experience, and their sharing behavior tional economy experience (M=3.90, p=.026). Next, we investigate the role of social interaction. Advances in Consumer Research 896 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 897

First, an exploratory study (N=211, Mage=37.38, 32.5% female, portantly, reduced social interaction increased satisfaction and well- Mturk) revealed that, across product categories (transportation, ac- being compared to the shared and increased greater social interaction commodation, dinning, master class, and wine tasting), consumers conditions (p’s < .043) to the level of the control condition (p=.176). perceive the sharing (vs. traditional) economy as involving greater Our work answers an urgent call for research on the nature and social interaction (t(210)=5.76, p<.001) and stress (t(210)=5.04, role of the social environment in the sharing economy and its impact p<.001). However, the P2P (sharing) vs. traditional providers’ levels on consumer wellbeing (Eckhardt et al. 2019), and adds useful in- of knowledge and expertise were not perceived to be significantly sights to marketing practice. It also contributes to sociology theories different t( (210)=1.60, p=.11). on group dynamics by documenting a unique setting of initiated so- Study 2 examined the effect in a setting with social appeal – cial interactions in an economic context and the unexplored impact shared dining, and investigated the underlying psychological proof group size. cess. Undergraduate students (N=160, Mage=19.9, 49% female) imagined going to dinner with a friend in a shared dining event (vs. The Meme Economy: How Internet Memes Impact a restaurant) called EatOut in a nearby city. At the venue there was Consumption a chef who served traditional Italian food and ten diners engaged in conversations. The results revealed that the sharing economy EXTENDED ABSTRACT decreased satisfaction (M=4.34) and overall well-being (comfort, In the current society, a major portion of word-of-mouth and in- enjoyment, and pleasant experience; M=3.93) compared to the tra- formation sharing takes place online (Chen and Xie 2008; Godes and ditional economy (M=4.74, p=.025; M=4.23, p=.009; respectively). Mayzlin 2004; Kozinets et al. 2010). Among a flood of information Importantly, social awareness mediated the effect (“I would feel exchanged online, certain content receives more attention and be- more socially aware,” “The presence of others would bother me,” comes viral (Godes and Silva 2012). As virality indicates greater im- and “The interactions with the chef and the other guests might feel pressions and sales for firms, prior work have investigated if certain awkward”; α=.58; a×b=-.0870, 95% CI=[-.1627,-.0287]). types of content are more likely to achieve viral status online, such We also tested the role of privacy concerns which may impact as ones that are more humorous and emotional (Berger and Milkman consumer experience (“I am concerned that because I went to EatOut, 2012; Warren et al. 2018). However, less is understood about how some people may have my personal information” and “Other people consumers might infer virality of certain online contents and whether at EatOut are likely to invade my privacy”; r=.57, p<.001). Privacy the perception of virality can drive consumption behaviors. concerns also mediated the effect (a×b=-.0630, 95% CI=[-.1533,- In this paper, we propose that internet ‘memes’ evoke the in- .0064]), but importantly, a competitive mediation analysis (Model 4) ference that a content has become viral, and that this perception in- showed that only social awareness mediated the effect, ruling-out a creases consumption intention. Memes are often created by consum- relevant alternative explanation. ers, usually in a picture and word caption format. Originally coined Study 3 sought to document the role of social interaction in the as a term to describe the mutation and evolution of ideas by Richard sharing economy in a real-world setting. We used a supervised ma- Dawkins, a key characteristic of a meme is that it mutates an original chine learning approach (Support Vector Machine; SVM) to analyze content to add a novel idea. For instance, an image of a grumpy cat reviews (N=25,251 English reviews out of 31,041 reviews in all lan- combined with a written caption, ‘Love is in the air? Get out the gas guages) from a world-leading shared dining platform. First, 3,000 mask’ is a meme that mutates the original content, a grumpy cat. sentences were randomly sampled from the reviews and labeled as 1 Even though millions of consumers encounter and share memes on if referenced social interaction or 0 if not (four independent coders; social media and firms realize its importance (McCrae 2017), con- α>.8). The best-fitting SVM model was used to classify the rest of sumer research has not yet investigated this phenomenon. We theo- the sentences. Then, for each review, we created a social interac- rize that when encountering a meme, consumers infer that the focal tion score which represents the percentage of sentences in a review content (e.g., grumpy cat) has become viral. This perception of viral- that mentioned social interaction. Due to skewness of the data, we ity invokes a sense of FOMO for consumers who do not understand treated ratings as five stars (=1) or fewer stars (=0). As expected, a the focal content (i.e., fear of missing out; Hayran et al. 2016; Rifkin logit regression of social interaction on rating revealed a negative et al. 2015), which leads to an increased intention to consume. effect (β=-.155, p=.0135), indicating that social awareness decreased To provide preliminary support for our predictions, a 2-cell be- consumer satisfaction. The effect holds when controlling for location tween-subjects design (N=77) was used in study 1. The meme condi- and price as well as with full variation in star ratings. Interestingly, tion participants imagined that while browsing on social media, they a greater number of guests at dinner decreased ratings (p<.001; also saw that their friend had “posted a meme about a new TV show,” accounting for the individual’s party). The negative effect of party while the non-meme condition imagined that their friend had “made size on ratings was mediated by greater emphasis on social interac- a post about a new TV show.” As predicted, participants who saw a tion in the reviews. social media post in the form of a meme (vs. non-meme) reported Finally, Study 4 sought to manipulate social interaction and greater consumption intention toward the new TV show (3 items examine the effect in another context – ride sharing, where social measuring their interest, likelihood of watching trailer, and descrip- motivation is considered less meaningful. Participants (N=407, tion of reading about the TV show; M=5.13 vs. 4.49, p=.018). In a subsequent consequential choice, memes also significantly increased Mage=34, 38% female, Mturk) imagined that they took an Uber to get to a meeting in the city. The ride arrived quickly and they got to the the likelihood that participants chose to read the TV show mentioned meeting on time. Participants read that they were the only passenger in the social media post (63% vs. 38.5%; χ2(1) = 4.70, p = .041). (control) vs. sat next to another passenger (shared) vs. sat next to Study 2 had two objectives. First, we aimed to replicate the ef- another passenger who tried to start a conversation (increased social fects from study 1 using real memes and non-meme posts (i.e., an interaction) vs. another passenger who was wearing headphones and image and caption format), rather than relying on the participant to concentrated on their phone (decreased interaction). The shared and imagine a post. Second, we tested the mediation process of perceived high social interaction conditions decreased consumer satisfaction virality. In a 2-cell between-subjects design (N=147), participants and well-being compared to the control condition (p’s < .001). Im- imagined browsing on Twitter. Next, participants in the meme condi- 898 / Sharing and Consumer Experience: How Does Different Forms of SharingAffect Consumer Evaluations of Experience tion were presented with a post by CBS Sports, with a meme about more intense and readily direct and accessible feelings of the events, Kawhi Leonard, a basketball player who had won the NBA cham- whereas the observer’s perspective generates less emotional cues, pionship title. Those in the non-meme condition also saw a similar and people tend to scrutinize themselves in the scenario, as if they post, but the image in the post was simply a photo of Kawhi Leonard. were watching a movie of themselves (Zhang and Yang, 2015).We Consistent with the results of study 1, seeing a tweet that employed a examine how visual perspective affects consumers’ attitude toward meme increased consumption intention, compared to a tweet by the the experience and their sharing intention (Study 1), through the lev- same source that employed a non-meme image (M=4.54 vs. 3.61, el of present focus the consumers have on experiences. p=.007). Also, as predicted, three items measuring perceived virality Temporal focus is the extent to which individuals characteristi- of Kawhi Leonard’s title revealed that a tweet employing a meme cally direct their attention to the past, present, and/or future (Shipp, (vs. non-meme) was perceived to have become more viral (M=5.30 Edwards, and Lambert, 2009; Zimbardo and Boyd, 1999). Individu- vs. 4.78, p=.031). Perception of virality significantly mediated con- als focus on the past, present or the future to varying degrees. Such sumption intentions (b=.3930, SE=.1728, 95% CI: [.0555, .7253]). psychological differences in orienting their thoughts at different time Study 3 was aimed to test directionality of the effects, such that frames, is capable of eliciting affective reactions and attitude at dif- memes indeed increase consumption intention, rather than other ferent degrees. forms of social media posts decreasing consumption intention. In a When engaging in an experience, consumers are oriented to fo- 3-cell between-subjects design (control vs. traditional ad vs. meme; cus on the current experience. Being immersive in the experience N=145), participants saw a social media post about a new TV series enhances consumers’ orientation to focus on the present experience. by AmazonPrime. Participants saw one social media post by an ac- Building on this notion, we propose that visual perspective of photos count called FilmAndTVGuru, either with no image, a poster of the will affect the level of present focus, since the actor’s perspective show, or a meme. Similar to the previous study results, the meme encourages consumers to focus on the present. This increases their condition significantly increased consumption intention compared to attitude toward an experience and sharing intention. On the other the control condition (M=3.52 vs. 2.76, p=.042). While the tradition- hand, the observer’s perspective orients consumers to focus less on al ad (M=3.09) did not significantly differ from the control (p=.837) the current experience, leading to less favorable outcomes. nor the meme condition (p=.493). Importantly, the results also dem- Consumers could engage experiences in a solo setting, which onstrate that the meme condition increased the perception of virality are experiences that only involve the participation of the consumer compared to both the control condition (M=3.29 vs. 2.37, p=.006) themselves, or in a group setting, which are events that involve the and the traditional ad condition (M=2.40, p=.008). There was no dif- interdependent participation of two or more consumers who concur- ference between the control and ad condition (p>.95). rently participate and are able to interact with each other (Aronson, Finally, our theory is that when a consumer encounters a meme, Wilson, and Akert, 2015). Solo experiences are shown to induce con- it is perceived to have gone viral, which evokes FOMO (i.e., fear of sumers to avoid public judgments and evaluations. Concentrating on missing out), thus increasing one’s own intention to consume the the experience itself intensifies the effect of the experience, whereas content of the meme. Thus, even if a social media post involves a group experiences could be impacted by the nonverbal reactions of meme, when FOMO is not evoked, such as when a meme is posted companions (Ramanathan and McGill 2007), which serves as dis- by a company, rather than one’s friend, the effects would be attenu- tractions for consumers from focusing on the present event itself. ated. Study 4 tested this using a 2(meme vs. non-meme) X 2(source: Based on these findings, we demonstrate that a solo experience will friend vs. firm) between-subjects design (N=372). In a similar set-up amplify consumers’ attitude toward an experience and sharing inten- as previous studies, participants saw either a post involving a meme tion when the photo is taken from the actor’s perspective, whereas image or a non-meme image. To manipulate the source, participants a group experience will attenuate this effect, serving as a boundary in the friend conditions saw a post by a friend, while those in the condition. firm conditions saw a post by a firm, Hulu. As intended, FOMO was In Study 1, we show that consumers generate more positive at- triggered when participants saw a meme (vs. non-meme) posted by titude toward an experience when the photo is taken from the actor’s a friend (M=3.82 vs. 3.49, p=.067), but the image format did not (vs. observer’s) perspective across three studies with different ma- impact FOMO when the source was a firm (M=2.68 vs. 2.86, NS). nipulations. In Study 1a, 239 participants on Mturk to participate in Further, confirming our predictions, memes increased consumption an imagination task. Visual perspectives of the photos were manipu- intention when the source was a friend (M=3.57 vs. 3.16, p=.083), lated by the images presented to the participants. The photos used in but it did not have impact when the source was Hulu (M=3.53 vs. this study were pretested in a pilot study. Consistent with our pre- 3.62, NS). dictions, participants who were assigned to the actor’s perspective reported that they enjoyed making the flower more (Mactor = 8.32, Let Me See How I Look: How Visual Perspective Affects SD = .10) than those assigned to observer’s perspective condition Consumer Experience and Sharing Behaviors (Mobserver = 8.06, SD = .10) (p = .06). In Study 1b, 169 students were recruited to the lab where they made a paper flower. We held EXTENDED ABSTRACT the photo taker constant by having the lab proctor take the picture for How individuals visualize an event is known as “visual per- participants in both the actor’s and the observer’s perspective condi- spective.” Actor’s perspective (first-person perspective) occurs when tions to alleviate the concern of visual perspective being confounded people visualize an event as if they see it through their own eyes, with the photo takers in different conditions. Consistent with our pre- whereas observer’s perspective occurs when individuals visualize an dictions, participants who were assigned to the actor’s perspective event from an outsider’s perspective (third-person perspective). Past reported that they generated more favorable attitude toward the flow- research has focused on the effects of visual perspective on memo- er-making activity (Mactor = 5.00, SD = .23) than those assigned to ries, emotions, and behaviors (Jiang, Adaval, and Steinhart, 2014; the no-photo (control) condition (Mcontrol = 4.40, SD = .22) (p = Jiang and Wyer, 2008; Libby, Shaeffer, and Eibach, 2009; Libby .04). The difference in attitude between the actor’s and the observer’s and Eibach, 2011; Shaeffer, Eibach, and Slemmer, 2007; Vasquez perspectives is marginal, but those assigned to the actor’s perspec- and Buehler, 2007). For instance, the actor’s perspective triggers tive condition generate marginally more favorable attitude than those Advances in Consumer Research (Volume 48) / 899 in the observer’s perspective condition (Mobserver = 4.28, SD = .24) they know (Argo et al., 2006; Packard and Wooten, 2013). If such (p = .09). There is no significant difference between the control and feedback is not positive, it can have a more devasting effect than the observer’s condition (p = .62). We found the same pattern with feedback from someone more distant (Rubin 1974; Small 2017). enjoyment and sharing behavior as well. In Study 1c, we recruited Consequently, both bragging and inner moral dilemmas should en- 310 students to the lab to participate in a card-making activity. In hance sensitivity to a known (vs. unknown) audience’s judgement. this study, we held the photo-taker constant by having students take Precisely, taking pictures of indulgence foods with the intention to the photos in both the actor’s and the observer’s conditions. Partici- share them online should lower levels of anticipated positive judge- pants who were assigned to the actor’s perspective reported that they ments when those pictures are shared with a mostly known (vs. un- generated more favorable attitude toward the card-making activity known) audience. In turn, lower levels of anticipated positive judge- (Mactor = 4.95, SD = .16) than those assigned to the observer’s per- ments should reduce consumption enjoyment (Robinson and Higgs, spective condition (Mobserver = 4.49, SD = .15) (p = .04). There 2012; Stok et al., 2014). Three studies test our predictions. is no significant difference between the control and the observer’s In the first experiment, participants (N=154) imagined dining condition (p = .39), or the actor’s condition (p = .21). We found the alone at a restaurant and ordering the restaurant’s famous plate. De- same patter for enjoyment and sharing behavior as well. pending on the condition, participants saw a picture a homemade Study 2 shows that whether an experience is in a solo or group apple pie or a fresh fruit salad. Participants were told that before eat- setting moderates the effect that visual perspective has on sharing ing, they decided to take a picture of the dish to share it on Instagram. behavior. We randomly assigned 252 participants on Mturk to one To operationalize audience type, we manipulated the privacy setting of the conditions of a 2 (perspective: actor, observer) x2 (type: solo, of the Instagram account. In the mostly unknown condition, partici- group) between-subjects design. We obtained a significant interac- pants imagined that their account was public and that both their In- tion (F(1, 217) = 3.98, p = .05).. Visual perspectives of the photos stagram followers and other users could see the post. In the mostly were manipulated by the images presented to the participants. The known condition, the account setting was private and access to the photos used in this study were pretested in a pilot study. Planned picture was limited to Instagram followers. To assess whether par- comparison tests revealed, as predicted, that in the solo experience ticipants felt a difference under both conditions, we measured per- condition, participants generate more positive attitude toward the ex- ceived anonymity when sharing with two items adapted from Hite, perience when the photo was taken from the actor’s perspective ver- Voelker, & Robertson (2014) (e.g. “When thinking about who can sus the observer’s perspective (Mactor = 8.33, Mobserver = 7.87, p = see my picture, I am confident that they do not know who I am” (1: .06). However, in the group experience condition, the effect of visual strongly disagree to 7: strongly agree)). As expected, participants felt perspective is attenuated (Mactor = 7.94, Mobserver = 8.15, p = .37). a greater sense of anonymity in a public setting (M=4.06) than in a Across four experiments, we find converging evidence for the private one (M=3.35) (p<.000). We measured anticipated enjoyment effect of visual perspective on consumers’ experiences: the actor’s from consumption with 2 items adapted from Van Boven and Ash- perspective generates more positive outcomes than the observer’s worth (2007) (e.g. “when you think about eating this dessert, how perspective condition. We control for who took the photo across happy does it make you?” (1: very unhappy to 7: very happy)) and studies. Further, we show that this effect is attenuated in a group anticipated positive judgement with three items adapted from Chen setting. (2017) (e.g. “After sharing the food picture, do you think the people with whom you shared this picture will view you more or less favor- Indulging Online: A Moral Dilemma ably?”(1: less favorably to 7: more favorably)). The results revealed a significant interaction effect (F(1,150) = 4.14, p<.05). Specifically, EXTENDED ABSTRACT anticipated positive judgement was lower when sharing indulgence Food pictures are everywhere on social network sites. Yet, de- food picture under a private setting (M = 4.39) than under public spite their predominance online, there has been little research ex- setting (M = 4.82, p = .029). No difference existed with the less in- plaining how sharing pictures of food consumed affects consumption dulgence option (p = .52). Furthermore, participants expected higher enjoyment. levels of anticipated consumption enjoyment when the picture was Previous research found that taking pictures of experiences with shared under a public (vs. private) setting (β: .15, 95% CI: .01 to the intention to share them online (vs. keeping them for oneself) re- .31). duces enjoyment of experiences (Barasch et al., 2017). Furthermore, In a second study, we replicated the experiment of the first study such an effect is accentuated when the picture is shared online with except that food stimuli were changed. We replaced the apple pie acquaintances rather than with friends. Yet, we propose that when it by a homemade “lasagna” and the fruit salad by a “grilled lemon comes to indulgence foods, people derive lower levels of consump- chicken” (pictures pretested for level of indulgence). Results showed tion enjoyment when food pictures are shared with a mostly known significant interaction effect (F(1,143) = 5.73, p = <.05). For the lasa- (vs. unknown) audience. Indulgence foods refer to caloric options, gna plate, anticipated positive judgement was lower under a private such as brownies or lasagna (Locher et al., 2005). This type of food setting (M = 4.56) than under a public one (M = 5.08, p = <.05). No is mainly associated with pleasure (Hirschman and Holbrook, 1982) difference existed with the picture of the grilled chicken (p = .19). and such pleasure is mainly egocentric (Sober and Wilson, 1999). Furthermore, consumers expected lower levels of anticipated con- Yet, sharing egocentric pleasure with known others can backfire. sumption enjoyment when the picture of indulgence food was shared First, the presence of known (vs. unknown) others mitigates under a private (vs. public) setting (β: .20, 95% CI: .02 to.42). egocentrism as each individual is more salient (Alicke and Govorun, In a third study, the experiment was conducted in a lab setting. 2005). Consequently, sharing egocentric pleasure with known others The cover story stated that a restaurant wanted to see how their new might be perceived as bragging (De Angelis et al., 2012). Second, dessert would perform on Instagram. Participants (N=207) had to indulgence foods are often considered as off-limits (Locher et al., try the dessert and post a picture of it on their Instagram account. 2005) and consuming such foods can activate moral dilemmas of It was mentioned that they had to leave the picture on their account “vice” versus “virtue”. In situations where the self-image is ques- for few hours. Participants randomly tasted a brownie (indulgence tioned, individuals are particularly sensitive to feedback from people food) or a fruit salad. Account privacy setting was measured. We also 900 / Sharing and Consumer Experience: How Does Different Forms of SharingAffect Consumer Evaluations of Experience controlled for the effect of audience size by measuring the number of Frank, Mark G. and Thomas Gilovich (1989), “Effect of Memory followers (1: less than 100 to 6: 700 +). The results revealed a sig- Perspective on Retrospective Causal Attributions.,” Journal of nificant main effect of number of followers (p <.01) and a significant Personality and Social Psychology, 57(3), 399–403. interaction effect (F(1,202) = 3.84, p = .05). Specifically, anticipated Fredrickson, Barbara L., Michele M. Tugade, Christian E. Waugh, positive judgement was lower when sharing the picture of indul- and Gregory R. Larkin (2003), “What Good Are Positive gence food in private setting (M = 3.27) than in public (M = 3.86, Emotions in Crisis? A Prospective Study of Resilience and p = .02). No difference existed with the fruit salad option p( = .45). Emotions Following the Terrorist Attacks on the United States Sharing pictures of indulgence foods online lower both levels on September 11th, 2001.,” Journal of Personality and Social of anticipated positive judgement and consumption enjoyment when Psychology, 84(2), 365–76. the audience is mostly known (vs. unknown). Greenwald, Anthony G. (1968), “Cognitive Learning, Cognitive Response to Persuasion and Attitude Change,” REFERENCES in Psychological Foundations 0/ Attitudes, eds. Anthony Argo, Jennifer J., Katherine White, and Darren W. Dahl (2006), Greenwald, Timothy C. Brock, and Timothy C. Ostrom, New “Social Comparison Theory and Deception in the Interpersonal York: Academic Press. Exchange of Consumption Information,” Journal of Consumer Godes, David, and Dina Mayzlin (2004), “Using Online Research 33 (1), 99-108. Conversations to Study Word-of-Mouth Communication, Alicke, Mark D., and Olesya Govorun (2005), “The Better-Than- Marketing Science 23(4), 545-560. Average Effect,”The Self in Social Judgment 1, 85-106. Godes, David, and José C. Silva (2012), “Sequential and Temporal Aronson, Elliot, Timothy D. Wilson, and Robin Akert (2015), Dynamics of Online Opinion,” Marketing Science 31(3), 448- Social psychology, 8th Edition. Pearson, New York, NY. 473. Ayduk, Özlem and Ethan Kross (2010), “From a Distance: Hayran, Ceren, Lalin Anik, and Zeynep Gurhan-Canli (2016), Implications of Spontaneous Self-Distancing for Adaptive “Exploring the Antecedents and Consumer Behavioral Self-Reflection.,”Journal of Personality and Social Consequences of feeling of missing out (FOMO),” ACR North Psychology, 98(5), 809–29. American Advances. Barasch, Alixandra, Kristin Diehl, Jackie Silverman, and Gal Hirschman, Elizabeth C., and Morris B. Holbrook (1982), Zauberman (2017), “Photographic Memory: The Effects of “Hedonic Consumption: Emerging Concepts, Methods and Volitional Photo Taking on Memory for Visual and Auditory Propositions,” The Journal of Marketing, 92–101. Aspects of an Experience,” Psychological Science, 28(8), Hite, Dwight M., Troy Voelker, and Adrian Robertson (2014), 1056–66. “Measuring Perceived Anonymity: The Development of a Barasch, Alixandra, Gal Zauberman, and Kristin Diehl (2018), Context Independent Instrument,” Journal of Methods and “How the Intention to Share Can Undermine Enjoyment: Measurement in the Social Sciences, 5(1), 22–39. Photo-Taking Goals and Evaluation of Experiences,” Journal Hung, Iris W. and Anirban Mukhopadhyay (2012), “Lenses of the of Consumer Research 44 (6), 1220-1237. Heart: How Actors’ and Observers’ Perspectives Influence Berntsen, Dorthe and David C. Rubin (2006), “The Centrality of Emotional Experiences,” Journal of Consumer Research, Event Scale: A Measure of Integrating a Trauma into Ones 38(6), 1103–15. Identity and Its Relation to Post-Traumatic Stress Disorder Jiang, Yuwei, Rashmi Adaval, Yael Steinhart, and Robert S. Wyer Symptoms,” Behaviour Research and Therapy, 44(2), 219–31. (2014), “Imagining Yourself in the Scene: The Interactive Campbell, Margaret C., and Karen Page Winterich (2018), “A Effects of Goal-Driven Self-Imagery and Visual Perspectives Framework for the Consumer Psychology of Morality in the on Consumer Behavior,” Journal of Consumer Research, 41(2, Marketplace,” Journal of Consumer Psychology, 28 (2), 167- )418–35. 79. Jiang, Yuwei, and Robert S. Wyer Jr. (2009), “The Role of Carter, Travis J. and Thomas Gilovich (2012), “I Am What I Do, Visual Perspective in Information Processing,” Journal of Not What I Have: The Centrality of Experiential Purchases Experimental Social Psychology, 45 (3), 486–95 to the Self-Concept,” Journal of Personality and Social Kross, E., O. Ayduk, and W. Mischel (2005), “When Asking ‘Why’ Psychology, 102(6), 1304–17. Does Not Hurt Distinguishing Rumination From Reflective Chen, Yubo, and Jinhong Xie (2008), “Online Consumer Processing of Negative Emotions,” Psychological Science, Review: Word-of-Mouth As A New Element of Marketing 16(9), 709–15. Communication Mix,” Management Science 54(3), 477-491. Koutstaal, Wilma, Daniel L. Schacter, Marcia K. Johnson, Kathryn De Angelis, Matteo, Andrea Bonezzi, Alessandro M. Peluso, Derek E. Angell, and Mara S. Gross (1998), “Post-Event Review in D. Rucker, and Michele Costabile (2012), “On Braggarts Older and Younger Adults: Improving Memory Accessibility and Gossips: A Self-Enhancement Account of Word-of- of Complex Everyday Events.,” Psychology and Aging, 13(2), Mouth Generation and Transmission,” Journal of Marketing 277–96. Research, 49(4), 551–563. Kozinets, Robert V., Kristine De Valck, Andrea C. Wojnicki, Eckhardt, Giana M., and Fleura Bardhi (2015), “The Sharing and Sarah JS Wilner (2010), “Networked Narratives: Economy isn’t about Sharing at all,” Harvard Business Understanding Word-of-Mouth Marketing in Online Review, 28(1). Communities,” Journal of Marketing, 74(2), 71-89. Eckhardt, Giana M., Mark B. Houston, Baojun Jiang, Cait Lambert, Nathaniel M., A. Marlea Gwinn, Roy F. Baumeister, Amy Lamberton, Aric Rindfleisch, and Georgios Zervas (2019), Strachman, Isaac J.Washburn, Shelly L. Gable, and Frank D. “Marketing in the Sharing Economy,” Journal of Marketing, Fincham (2013),“A Boost of Positive Affect: The Perks of 83(5), 5-27. Sharing Positive Experiences,” Journal of Social and Personal Relationships, 30 (1), 24–43. Advances in Consumer Research (Volume 48) / 901

Libby, Lisa K. and Richard P. Eibach (2002), “Looking Back in Robinson, John A. and Karen L. Swanson (1993), “Field and Time: Self-Concept Change Affects Visual Perspective in Observer Modes of Remembering,” Memory, 1(3), 169–84. Autobiographical Memory.,” Journal of Personality and Rubin, Zick (1974), “Lovers and Other Strangers: The Social Psychology, 82(2), 167–79. Development of Intimacy in Encounters and Relationships: Libby, Lisa K., Richard P. Eibach, and Thomas Gilovich (2005), Experimental Studies of Self-Disclosure between Strangers “Here’s Looking at Me: The Effect of Memory Perspective on at Bus Stops and in Airport Departure Lounges can Provide Assessments of Personal Change.,” Journal of Personality and Clues about the Development of Intimate Relationships,” Social Psychology, 88(1, )50–62. American Scientist, 62(2), 182–190. Libby, Lisa K., Eric M. Shaeffer, Richard P. Eibach, and Jonathan Seligman, martin EP, and Mihaly Csikszentmihalyi (2014), A. Slemmer (2007), “Picture Yourself at the Polls: Visual “Positive Psychology: An introduction,” Flow and the Perspective in Mental Imagery Affects Self-Perception and Foundations of Positive Psychology, 279-298. Behavior,” Psychological Science, 18(3), 199–203. Shaeffer, Eric M, Lisa K. Libby, and Richard P. Eibach (2015), Libby, Lisa K., Eric M. Shaeffer, and Richard P. Eibach (2009), “Changing Visual Perspective Changes Processing Style: “Seeing Meaning in Action: A Bidirectional Link between A Distinct Pathway By Which Imagery Guides Cognition,” Visual Perspective and Action Identification Level.,”Journal Journal of Experimental Psychology: General, 144(3), 534. of Experimental Psychology: General, 138(4), 503–16. Shipp, Abbie J., Jeffrey R.Edwards, and Lisa Schurer Lambert Libby, Lisa K., and Richard P. Eibach (2011), “Visual Perspective (2009), “Conceptualization and Measurement of Temporal in Mental Imagery: A Representational Tool That Functions Focus: The Subjective Experience of the Past, Present, and in Judgment, Emotion, and Self-Insight,” Advances in Future,” Organizational Behavior and Human Decision Experimental Social Psychology, 44, 185–245. Processes. Locher, J. L., Yoels, W. C., Maurer, D., & van Ells, J. (2005). Shipp, Abbie J. and Brad Aeon (2019), “Temporal Focus: Thinking Comfort Foods: An Exploratory Journey Into The Social and about the Past, Present, and Future,” Current Opinion in Emotional Significance of Food. Food and Foodways, 13(4), Psychology. 273–297. Sutin, Angelina R., and Richard W. Robins (2008), “When the Lux, Mathias, Marian Kogler, and Manfred Del Fabro (2010), ‘I’ Looks At the ‘Me’: Autobiographical Memory, Visual “Why Did You Take This Photo: A Study on User Intentions Perspective, and the Self,” Consciousness and Cognition, in Digital Photo Productions. ,” Proceedings of the 2010 ACM 17(4), 1386-1397. workshop on Social, adaptive and personalized multimedia Sober, Elliot, and David Sloan Wilson (1999), “Unto Oothers: The interaction and access - SAPMIA 10. Evolution and Psychology of Unselfish Behavior,” Harvard McCrae, James (2017), “Meme Marketing: How Brands Are University Press, 218. Speaking A New Consumer Language”, Forbes https://www. Small, Mario Luis (2017), Someone to Talk to, Oxford University forbes.com/sites/forbescommunicationscouncil/2017/05/08/ Press. meme-marketing-how-brands-are-speaking-a-new-consumer- Stok, F. Marijn, Kirsten T. Verkooijen, Denise TD de Ridder, John language/#153ae3d437f5 BF de Wit, and Emely de Vet (2014), “How Norms Work: Milanova, Veselina, and Peter Maas (2017), “Sharing Intangibles: Self‐Identification, Attitude, and Self‐Efficacy Mediate the Uncovering Individual Motives for Engagement in a Sharing Relation between Descriptive Social Norms and Vegetable Service Setting,” Journal of Business Research, 75, 159-71. Intake,” Applied Psychology: Health and Well-Being, 6(2), Neoh, Jun Guan, Maxwell Chipulu, and Alasdair Marshall (2017), 230–250. “What Encourages People to Carpool? An Evaluation of Warren, Caleb, Adam Barsky, and A. Peter Mcgraw (2018), Factors with Meta-Analysis,” Transportation, 44(2), 423-47. “Humor, Comedy, and Consumer Behavior,” Journal of Packard, Grant, and David B. Wooten (2013), “Compensatory Consumer Research, 45(3), 529-552. Knowledge Signaling in Consumer Word-of-Mouth,” Journal Williams, J. Mark and Keith Broadbent (1986),“Autobiographical of Consumer Psychology, 23(4), 434–450. Memory in Suicide Attempters,” Journal of Abnormal PwC (2015), The sharing economy, Consumer Intelligence Series, Psychology, 95(2), 144–49. https://www.pwc.com/us/en/technology/publications/assets/ Williams, Alishia D. and Michelle L. Moulds (2007), “Cognitive pwc-consumer-intelligence-series-the-sharing-economy.pdf Avoidance of Intrusive Memories: Recall Vantage Perspective Ratner, Rebecca K. and Rebecca W. Hamilton (2015), “Inhibited and Associations with Depression,” Behaviour Research and from Bowling Alone,” Journal of Consumer Research, 42(2), Therapy, 45(6), 1141–53. 266–83. Wyer, Rober S., Jr., I Hung, and Y Jiang (2008), “Visual and Verbal Reis, Harry T., Shannon M. Smith, Cheryl L. Carmichael, Processing Strategies in Comprehension and Judgment,” Peter A. Caprariello, Fen-Fang Tsai, Amy Rodrigues, and Journal of Consumer Psychology, 18(4), 244–57. Michael R. Maniaci (2010),“Are You Happy for Me? how Van Boven, Leaf, and Laurence Ashworth (2007), “Looking Sharing Positive Events with Others Provides personal and Forward, Looking Back: Anticipation is More Evocative than Interpersonal Benefits,”Journal of Personality and Social Retrospection,” Journal of Experimental Psychology: General, Psychology, 99 (2), 311-29. 136(2), 289. Rifkin, Jacqueline, Chan Cindy, and Barbara Kahn (2015), “FoMO: Zhang, Jing and Xiaojing Yang (2015), “Stylistic Properties and How the Fear of Missing out Leads to Missing out,” ACR Regulatory Fit: Examining the Role of Self-Regulatory North American Advances. Focus in the Effectiveness of an Actors vs. Observers Visual Robinson, Eric, and Suzanne Higgs (2012), “Liking Food Less: Perspective,” Journal of Consumer Psychology, 25(3), 449–58. The Impact of Social Influence on Food Liking Evaluations in Female Students,” PloS One, 7(11), e48858. The Pursuit of Fair Market Outcomes in an Unequal World Chairs: Camilla Zallot, Erasmus University, The Netherlands Gabriele Paolacci, Erasmus University, The Netherlands Amit Bhattacharjee, INSEAD, France

Paper #1: Inequality and Inefficiency people fail to consider the underlying values represented by different Serena Hagerty, Harvard Business School, USA sets of units, their preferences for equality can perversely produce Michael Norton, Harvard Business School, USA unequal distributions of rewards and punishments. Duani, Barasch, and Bhattacharjee find that the perceived fairness of unequal market Paper #2: Value Pricing and Distributive Concerns outcomes depends on who is judging them, even when those out- Camilla Zallot, Erasmus University, The Netherlands comes arise from transparently unbiased processes. Men perceive Gabriele Paolacci, Erasmus University, The Netherlands gender wage gaps as equally fair regardless of their direction, while Amit Bhattacharjee, INSEAD, France women’s judgments of institutional fairness depend heavily on who Paper #3: Simple Beliefs about Fairness is disadvantaged. Ellen R. K. Evers, University of California, Berkeley-Haas, Together, these papers offer a broad set of insights on how in- USA equality concerns both shape and constrain the pursuit of fair market Michael O’Donnell, University of California, Berkeley-Haas, outcomes. As this session illustrates, pursuing fairness in an unequal USA world requires addressing difficult societal questions: what consti- Yoel Inbar, University of Toronto, Scarborough, Canada tutes “fair” market outcomes? Who decides? How can we design Paper #4: When Do Wage Disparities Indicate Discrimination? markets to achieve such outcomes? We hope this session will attract Fairness Perceptions Depend on Group Membership a broad audience of researchers studying inequality and fairness, Nofar Duani, New York University, USA economic psychology, and consumer morality, and trigger more con- Alixandra Barasch, New York University, USA tributions on these fundamental questions. Amit Bhattacharjee, INSEAD, France Inequality and Inefficiency SESSION OVERVIEW Inequality is an increasing concern for the public and policy- EXTENDED ABSTRACT makers alike, as much of the world’s wealth continues to reside in the Imagine two people compete in an auction for tickets to a con- hands of a fraction of the population. Beyond those facing resource cert. One bids $100 and the other bids $50. Who do you think wants scarcity, consumers at every level of affluence are constantly con- to go to the concert more? Who values the tickets more? Now imag- fronted with the uncomfortable reality of inequality in the market- ine that the person bidding $100 is Bill Gates and the person bidding place, and must find ways to justify or cope with it. Doing so often $50 is a school teacher. Does the $100 bid still signal a stronger pref- entails navigating conflicts between different standards of fairness. erence for the ticket? Normatively, one of the primary functions of markets is to facilitate Willingness to pay is a commonly used measure in economics the fair allocation of limited resources. In practice, however, con- and public policy to determine an individual’s valuation of a good sumers’ beliefs about what constitutes “fair” outcomes may not align (Sunstein, 2007). Research on willingness to pay assumes that any with the outcomes that markets are designed to achieve. variation in WTP is a result of varying preferences and not attribut- The present session explores the challenges of promoting equi- able to variation in ability to pay. When wealth is not equally distrib- table market outcomes and maintaining fair institutional processes uted, do markets continue to allocate goods efficiently to those who in the face of existing inequalities. The four papers, all in an ad- value them most highly? vanced stage of completion, highlight the inherent arbitrariness and Consideration of wealth distribution is critical because wealth subjectivity that make market fairness so difficult to achieve despite inequality has been steadily rising over the past thirty years (Alva- its widespread appeal. redo et al., 2013). Initial evidence suggests that transparency of in- The first two papers describe how baseline differences in wealth equality (in addition to inequality itself) can reduce cooperation and affect the use of consumer willingness to pay as a signal of preference inhibit market outcomes (Hauser et al., 2016; Nishi et al., 205). The strength. Hagerty and Norton investigate how unequal endowments current research extends these findings by examining the effects of of wealth affect allocative fairness in real-world auctions. They find inequality (and transparency) on the distribution of goods in a com- that because the WTP of wealthier consumers exceeds their true val- petitive market. uations, they are often allocated scarce goods over poorer consumers Study 1 demonstrates the market-level effects of inequality and who actually value them more. Zallot, Paolacci, and Bhattacharjee transparency. Participants (N= 1001) were randomly assigned to one find that the perceived fairness of pricing goods based on customer of five wealth conditions [22/49/72/105/304 points], which represent WTP depends on customer wealth. People find it acceptable for sell- the true income distribution of the United States. They were then ers to maximize profit at the expense of affluent customers but unac- told they could use the points to bid against four other participants to ceptable to charge poorer customers according to their own stated win YouTube videos. If they did not win any videos they would have WTP, suggesting that market exchanges are regarded as opportuni- to complete five more minutes of counting tasks. Any points they ties to redress existing inequalities. did not bid were converted into a monetary bonus. Half of the par- The second pair of papers demonstrates that perceptions of fair ticipants received information depicting how their score compared market design and fair market outcomes are systematically affected to the other participants in their auction group, while the other half by normatively irrelevant factors. Evers, O’Donnell, and Inbar show were given no such information (Transparency vs No Transparency). that the same market allocations are perceived as differentially fair We find that wealthier participants win more videos (F(4, 909)= depending on the units used to describe those allocations. Because 8.80, p=0.00), despite spending a smaller proportion of their budget

(F(4,815)=3.93, p=0.004). As earnings increase, liking becomes a Overall we find that when there is inequality in a competitive weaker predictor of winning a video. The correlation between bid- auction, goods go to those with more money, not to those who value ding and liking is significantly lower for participants in the high- the good the most or will consume it fully. Transparency of inequal- est earning quintile than for those in the lowest earning quintile (22 ity exacerbates these effects. points) (r=.292 vs. r=.432; z=2.96, p=0.02). Even when controlling for liking of a video, participants earning more points were more Value Pricing and Distributive Concerns likely to win any given video (B=0.206, SE=.029, p<0.001). As a result, participants in the highest earning quintile won their least fa- EXTENDED ABSTRACT vorite video at the same rate that participants in the lowest earning Value pricing refers to setting product prices based on custom- quintile won their favorite video (15.1 % vs. 15.3 %; X2 (2, N=361) ers’ willingness-to-pay (WTP). This practice can not only increase =0.00, p=0.98). firm profit, but also enhance consumer welfare: because opportunity There was no main effect of transparency on number of wins, costs differ across consumers due to heterogeneous preferences and but there was a significant interaction: participants in the highest price sensitivities, value pricing can make products available to a quintile won more videos when inequality was transparent while all larger number of customers (Della Vigna & Gentzkow 2019). other participants won fewer videos (F(4,901) = 2.527, p=0.039). Prior work on price fairness has examined situational influences In the transparency condition, the highest quintile took advantage on pricing (cf. Kahneman, Knetsch, & Thaler 1986), but no existing of their relative wealthy by entering the market at higher rates and research has investigated whether relying on customers’ own assess- by bidding significantly more on their non-favorite videos, while all ments of value (i.e., their WTP) is perceived as more acceptable for other quintiles bid less on non-favorites (F(1,801)=3.0, p=0.02). some customers than for others. We present five studies demonstrat- In study 2 we were interested in the effect of inequality trans- ing that value pricing is seen as less fair when selling to poorer cus- parency on subsequent consumption rates. We replicated the proce- tomers, despite the fact that their WTP already reflects their tighter dure from Study 1, with one critical change- participants (N=563) budget constraints. were told they could watch the videos for as long as they would This finding persists across varying levels of profit margin like. We find that -- even controlling for total number of videos won, (Study 1c), and cannot be explained by perceived differences in pref- total number of video-minutes won, rank of video, and percentage of erence strength (Studies 1b-2). Moreover, it holds even when people earnings spent-- participants in the highest quintile watched a small- are explicitly aware that low-income customers have a lower WTP er proportion of their winnings when they were aware of their ad- than high-income customers, thus already reflecting their tighter budget constraints (Study 2). However, this effect is attenuated when vantage (Mtransparent= 0.11 vs. Mnotransp=0.13) while participants in the lowest quintile watched a larger proportion in the transparent con- the seller is also low-income (Study 3), suggesting that it arises because market exchanges are regarded as opportunities to redress ex- dition (Mtransparent= 0.19 vs. Mnotransp=0.15; F(1, 218)=10.88, p=0.01). This interaction is also significant when including all five conditions. isting inequalities. Overall consumption of videos decreases by 7.5% when transpar- Study 1a (N=450) employed a 3-group between-subjects de- ency is introduced in the market. sign. Participants imagined being a business owner selling a prod- In Study 3 we introduce a new paradigm-- an induced valu- uct that cost $25 to produce to customers who would buy it at any ation second-priced auction for gift cards-- modified from Garratt, price up to $75. Customers were described as living in either a dis- Walker and Wooders (2012). Rather than having participants rate the advantaged low-income, average income, or privileged high-income items up for auction as a measure of liking, participants were told neighborhood. Participants indicated what price they thought would explicitly how much they valued each gift card. These valuations be fair to charge, which we report as a proportion of WTP (%WTP) were randomized across participants. In a second-price auction, bid- across studies. Participants believed that low-income customers ding one’s value is always a dominant strategy. Therefore, by provid- should be charged a lower %WTP (76.6%), and high-income cusing participants with explicit values, we are able to identify any bid tomers a higher %WTP (91.1%), compared to average-income cus- greater than the gift card value as an overbid. tomers (85.7%, ps<.001). Again, participants (N=470) were randomly assigned to an en- Using a similar 2-group design, Study 1b (N=302) found that dowment [$.10, $.23, $.33, $.48, $1.38], reflecting the true income this observed difference in fair prices was not due to lower-income distribution of the United States and ensuring that all participants consumers being perceived as less able or willing to pay their stated could afford to overbid on even the most expensive gift card. Partici- WTP. We gauged participants’ comprehension of WTP by asking pants then competed in a second-price auction for five different gift them to indicate the price at which all customers would buy the prod- cards. The value of each gift card was randomized between $0.05, uct. 76% of participants (evenly distributed across conditions, X2<1) $0.06, $0.07, $0.08, and $0.09. correctly indicated the maximum price. Regardless, these partici- We find wealthier participants overbid more participants across pants still selected a lower %WTP as a fair price for disadvantaged all five gift cards (p < 0.001)- there is a main effect of wealth on customers (73.2% vs. 87.6%, X2=26.97, p<.001). Including partici- total amount of overbids (F (4,469) =16.44, p < 0.001) and number pants who failed to interpret the maximum price correctly did not of overbids (F (4,469) =17.66, p < 0.001). Consequently, we find affect these results (73.6% vs. 90.7%, 2X =41.48, p<.001). that the market favors the wealthy, rather than the players who value Study 1c (N=606) then tested whether this finding holds when the cards the most. The wealthier players win more cards overall (F WTP reflects a large versus small profit margin above production (4,469) =29.43, p < 0.001). A player’s endowment condition is a bet- costs ($12 in this study) with a 3 (WTP: $20, $38, $52) X 2 (Cus- ter predictor of winning a card (b= 0.725, SE= 0.05, p <0.001) than tomer Wealth) between-subjects design. Across conditions, we again their private valuation of the card (b= 0.291, SE= 0.04, p <0.001). found a main effect of Customer Wealth, whereby fair prices were Ultimately, 42.8% of the gift cards went to a player in the highest lower for low-income customers (%WTP=73.6% vs. 84.4%, t=4.40, wealth condition, while only 28.1% of gift cards went to a player p<.001). Even at the lowest profit margin (WTP=$20), value pricing who valued it the highest in the auction group. was judged as more immoral when customers were low versus high in income (M=3.16 vs. 2.30, t=3.19, p<.001), supporting our pro- 904 / The Pursuit of Fair Market Outcomes in an Unequal World posed mechanism. Accordingly, a significantly smaller proportion amount of community service, even though the high earner’s time is of participants chose to set prices at customer WTP (24% vs. 48%, more valuable. X2=10.85, p<.001). Moreover, controlling for how much customers In four sets of studies (within each set we vary the domain and were thought to want the product did not affect our results, again whether the outcomes are gains or losses; all N≥200, all relevant p- suggesting that these effects of customer wealth cannot be explained values <.01) we find evidence for the following three points: solely by a desire to reward those perceived to have stronger prefer- 1. When judging the fairness of assigned compensation or pun- ences. ishment, participants base their judgment primarily on the Study 2 (N=700) sought to replicate these findings in a within- dimension the outcome is expressed in. For example, par- subjects design where participants were fully aware of WTP differ- ticipants read about Alan and Bob, who do the same work ences across groups. We also varied whether poorer customers had for the same company. Because Alan has been with the com- a WTP that was lower than, higher than, or equal to that of richer pany longer, he makes 40€ per hour while Bob makes 25€ customers, using a 3 (WTP Difference) X 2 (Customer Wealth) per hour. Both Alan and Bob worked on a Sunday for a high mixed design. Participants imagined that a business owner was sell- priority project and are given a bonus by their boss. In one ing the same product in two different cities where customers had condition participants read that Alan gets a 400€ bonus while different levels of income and interest in the product (as expressed Bob gets a 300€ bonus. Participants judge this outcome to be through their WTP), on average. Across conditions, participants very unfair to Bob (M=68.10 on a 100 point scale with 50 be- set a lower price for customers in the poor city versus the rich city ing perfectly fair, p<.001). In the other condition, participants (%WTP=70.1% vs. 96.1%, F=55.61, p<.001). Most importantly, read that Alan gets 10 hours off while Bob gets 12 hours off even when low-income customers already reported a lower WTP (equivalent to 400€ and 300€in wages respectively). In this than high-income customers, in line with their tighter budget con- condition, participants judge the bonus to be unfair to Alan straints ($30 vs. $40), we still found a significant difference between (M=44.02, p=.01). fair prices (%WTP=74.7% vs. 88.1%, F=5.49, p=0.02). As in Study 2. Participants act as if their reliance on the outcome-dimension 1b, there were no such differences in perceived preference strength, is at least partially a mistake. In Study 2 we use a similar de- indicating that these inferences cannot explain our results. sign to Study 1, but before judging the fairness we “translate” Finally, Study 3 (N= 404) manipulated the relative wealth of the outcome into a different unit. For example, in one study buyers and sellers in a 2 (Buyer Wealth) X 2 (Seller Wealth) be- participants read about Alan and Bob who are both freelanc- tween-subjects design. Replicating our previous results, there was ers in high demand and make $50 and $25 per hour respec- a main effect of buyer wealth (%WTP=71.2% vs. 86.8%, F=44.3, tively. Both got caught running the exact same red light and p<.001). But importantly, this effect was qualified by a significant in- Alan was assigned a $150 [3 hours community service] fine teraction with seller wealth (F=11.42, p<.001). Participants selected while Bob was assigned a $100 [4 hours community service] significantly lower fair prices for poor buyers when the seller was fine. We added 2 more conditions in which we “translate” wealthy (%WTP=64% vs. 87.5%, p<.001), but this effect was at- the fine across dimensions. For example; “A $100 fine means tenuated when the seller was also poor (%WTP=77.8% vs. 85.9%, that Bob will have to work 4 more hours to pay the fine”. p=.069). This result suggests that people regard market exchanges as Participants then again indicated how fair they perceived this opportunities to redress existing inequalities: when the seller is not fine to be. In these conditions, the difference in perceived wealthy enough to enable such redistribution, value pricing to poor fairness attenuated but did not fully disappear. customers becomes more acceptable. 3. People succumb to the same biases when assigning out- Together, these studies show that people believe that sellers are comes. In Study 3 and 4 we have participants assign fines morally obligated to forgo profits and charge poor customers a sys- and bonuses across different dimensions and find that these tematically smaller proportion of what a product is worth to them. judgments follow the same pattern. Interestingly, even when Above and beyond poorer customers’ diminished ability to signal assigning fines in a fully within-participants design, many strong preferences through WTP (cf. Shaddy & Shah 2018), know- participants (~30%) still assign identical hours and identical ingly charging them what they are willing to pay is seen as inherently dollar amounts even when the implications across dimen- less fair and less morally acceptable. Our findings suggest that even sions are calculated right in front of them. In Study 4, we mundane market transactions may be evaluated as opportunities to even observe this behavior in a fully incentive-compatible correct existing inequalities. “Fair” prices based on customers’ own design where failure to realize this bias directly and nega- valuations may not be seen as fair enough, limiting the acceptability tively affects the outcome of the participant. of value pricing strategies. To summarize, we find that people strongly rely on the belief that outcomes should be equitable as the basis for what is fair. How- Simple Beliefs About Fairness ever, when outcomes can be expressed in different dimensions, peo- EXTENDED ABSTRACT ple do not automatically realize this equivalence and fail to take this into account. Even when attention is drawn to the effect of dimension Almost all people value fairness, any many people believe that on perceived equity, participants do not fully correct their judgments. equal treatment exemplifies it. According to this rule, people should receive the same reward for the same work, and pay the same price When Do Wage Disparities Indicate Discrimination? for the same infraction. However, the domain in which these are calculated may lead to inconsistent intuitions. For example, people may Fairness Perceptions Depend on Group Membership feel that it is fair to fine a low earner and a higher earner the same EXTENDED ABSTRACT amount for a speeding ticket, even though the same fine represents a The gender pay gap has been the focus of substantial public different proportion of income for each. At the same time, they may debate and policy intervention. In 2013, American women earned feel it is fair to require both the low and high earner to do the same 82 cents for every dollar earned by men (Bureau of Labor Statistics, 2016), and a recent analysis revealed similar disparities of varying Advances in Consumer Research (Volume 48) / 905 size in all 36 OECD countries (OECD, 2020). However, the extent to versus women. For each task, the difference in earnings was fully which these aggregate differences indicate systemic unfairness and explained by a single reason unrelated to gender: i) spending more discrimination remains in dispute. time on the task, ii) performing better, iii) acting more competitively, A variety of research has thus tried to identify the underlying or iv) acting more cooperatively. Advantage was manipulated be- causes of the gender pay gap and to quantify their relative impact. tween subjects: half the participants learned that men earned more For instance, a recent analysis of the US labor market found that than women on average, while the other half learned that women 62% of the gender pay gap can be explained by observable differ- earned more than men. Participants then responded to the same fair- ences between male and female workers (e.g., work experience, in- ness DVs from Study 1 (αs>0.78). dustry, occupation, education; Blau & Kahn, 2017). Another stream A mixed ANOVA revealed a main effect of who was advan- of research suggests that this aggregate disparity might also reflect taged: overall, participants considered outcome to be fairer when different choices and career preferences between men and women women were paid more than men (F(1,170)=21.90, p<.001). Impor- in the workforce (Cook et al., 2018; Daymont & Andrisani, 1984; tantly, this was qualified by a significant interaction (F(1,170)=13.69, Solberg, 1999). p<.001). Across all four tasks, male participants judged the compen- However, there are conflicting perspectives on the extent to sation structure as equally fair regardless of which gender earned which these underlying causes are actually inconsistent with or more (Fs<2.18, ps>.20). In contrast, female participants responded themselves reflective of historical gender biases. Moreover, it is un- strongly to the direction of the gender gap, finding compensa- clear how these potential explanations affect lay judgments of insti- tion schemes less fair when men versus women were advantaged tutional unfairness, which are critical to understanding public reac- (Fs>14.20, ps<.001). More specifically, female participants rated tion to policies promoting gender equality. the compensation scheme as fairer than male participants when the Accordingly, the current research explores how consumers per- pay gap favored women (M=5.84 vs. 5.36; F=4.86, p=.03), but as ceive unequal market outcomes that arise from explicitly impartial less fair than male participants when men were advantaged (M=4.55 and transparent processes. Specifically, we investigate how fairness vs. 5.21; F=9.14, p=.003). Accordingly, women advocated flat-fee perceptions of the same outcomes might vary across groups. We compensation policies more strongly when men had earned more predict that while members of historically-advantaged groups (e.g., (M=4.03 vs. 5.23; F(1,170)=7.39, p=.007), while men’s support for men) will regard any outcomes resulting from unbiased processes this policy was unaffected by which gender had been advantaged as fair, members of historically-disadvantaged groups (e.g., women) (M=4.06 vs. 3.91; F<1). will judge the same outcomes as less fair, but only when their group Study 3 (N=292) replicated these results on mTurk using a be- is worse off. Members of disadvantaged groups may thus be more tween-subjects design where each participant evaluated the results of likely to interpret systematic group differences in choices, preferenc- one of these hypothetical tasks. Men again rated each of the tasks as es, or abilities as inherently reflective of historic inequalities. Three equally fair regardless of which gender had the advantage (Fs<.07, studies test and support this predicted interaction. ps>.792), while women rated each outcome as less fair when men Study 1 (N=127) was inspired by the results of a recent econo- earned more than when women did (Fs>4.68, ps<.031). metric analysis of Uber drivers’ compensation (Cook et al., 2018). In summary, we find that men and women evaluate gender dif- While the Uber algorithm assigns rides and sets fares through ex- ferences in market outcomes very differently, even when they arise plicitly gender-blind algorithms, results indicate that male drivers from unbiased processes. Hence, advantaged and disadvantaged still make 7% more per hour on the platform. This gender pay gap is consumers may subjectively evaluate the same cues as differentially fully explained by three factors: experience, route preferences, and diagnostic of institutional unfairness. Our findings highlight one rea- driving speed. In this study, participants learned about this observed son policy discussions concerning inequality can be so contentious: wage disparity and these explanatory factors (with order counterbal- perceptions of what constitutes fair treatment versus evidence for anced). For each factor, participants evaluated the fairness of Uber’s discrimination may be in the eye of the beholder. policies on four items (e.g., men have an unfair advantage on Uber, men and women have equal opportunities as Uber drivers; αs>0.8). REFERENCES A 2X3 mixed ANOVA revealed no differences in fairness rat- Alvaredo, F., Atkinson, A. B., Piketty, T., & Saez, E. (2013). The ings across these three factors (F=1.80, p=.167), and no interactions top 1 percent in international and historical perspective. with gender (F=1.06, p=.349). More importantly, we found the pre- Journal of Economic perspectives, 27(3), 3-20. dicted main effect of gender: female participants rated all three con- Blau, F. D., & Kahn, L. M. (2017). The gender wage gap: Extent, tributors to the Uber pay gap as significantly less fair than males trends, and explanations. Journal of Economic Literature, (M=3.97 vs. 4.74; F(1,125)=11.15, p=.001). Moreover, men were 55(3), 789–865. more likely to mention the impartiality of the algorithm in their Bureau of Labor Statistics. (2016). Women in the Labor Force: A open-ended responses (e.g., it’s the same formula; algorithms are not Databook. In US Bureau of Labor Statistics Reports. https:// biased), while women were more likely to provide broader interpre- doi.org/10.2307/144700 tations of each causal explanation (e.g., women might work fewer Cook, C., Diamond, R., Hall, J., List, J. A., & Oyer, P. (2018). The night hours because they would be less safe). gender earnings gap in the gig economy: Evidence from over Because Study 1 was based on real-world observations, we did a million rideshare drivers (No. w24732). National Bureau of not manipulate the direction of the pay disparity (i.e., women always Economic Research. made less than men). Hence, these effects might reflect particular Daymont, T. N., & Andrisani, P. J. (1984). Job preferences, college features of this context or outcomes alone. Studies 2 and 3 were de- major, and the gender gap in earnings. Journal of Human signed to address these limitations. Resources, 408-428. In Study 2 (N=174), subject pool participants were asked to DellaVigna, S., & Gentzkow, M. (2019). Uniform pricing in us evaluate the fairness of four hypothetical lab experiments (order retail chains. The Quarterly Journal of Economics, 134(4), counterbalanced) in which players completed various tasks for pay- 2011-2084. ment, where the average earnings ended up being unequal for men 906 / The Pursuit of Fair Market Outcomes in an Unequal World

Garratt, R. J., Walker, M., & Wooders, J. (2012). Behavior in OECD. (2020). Gender wage gap (indicator). Retrieved February second-price auctions by highly experienced eBay buyers and 26, 2020, from https://data.oecd.org/earnwage/gender-wage- sellers. Experimental Economics, 15(1), 44-57. gap.htm Kahneman, D., Knetsch, J. L., & Thaler, R. (1986). Fairness as a Shaddy, F., & Shah, A. K. (2018). Deciding who gets what, fairly. constraint on profit seeking: Entitlements in the market. The Journal of Consumer Research, 45(4), 833-848. American economic review, 728-741. Solberg, E. J. (1999). Using Occupational Preference in Estimating Nishi, A., Shirado, H., Rand, D. G., & Christakis, N. A. (2015). Market Wage Discrimination: the Case of the Gender Pay Gap: Inequality and visibility of wealth in experimental social Decomposition of a Reduced‐form Wage Equation. American networks. Nature, 526(7573), 426. Journal of Economics and Sociology, 58(1), 85-113. Norton, Michael I., and Dan Ariely. 2011. Building a better Sunstein, C. R. (2007). Willingness to pay vs. welfare. Harv. L. & America—One wealth quintile at a time. Perspectives on Pol’y Rev., 1, 303. psychological science 6(1): 9-12. How You Feelin’? The Role of Internal And External Cues on Consumer Sensory Experiences Chair: Jerry J. Han, University of Technology Sydney, Australia

Paper #1: Sounding Warm: The Role of Audio Pitch on Service sensory experiences. Specifically, Tsai and Zhao find that unpacking Perception (vs. packing) a consumption experience leads to greater sensory en- Jerry J. Han, University of Technology Sydney, Australia joyment even when the objective consumption episode is the same. Michael Lowe, Georgia Institute of Technology, USA Importantly, the authors find that the unpacking frame is more effective only when consumers focus on the remaining subcomponents to Paper #2: Vocal Similarity, Trust and Persuasion in Human-AI be enjoyed (to-go focus), because this increases consumers’ motiva- Agent Interactions tion to complete the sensory experience. Michael Lowe, Georgia Institute of Technology, USA In short, the session sheds light into how internal cues specific Na Kyong Hyun, Georgia Institute of Technology, USA to sensory inputs can influence consumer evaluations and how ex- Paper #3: Sound of Products’ Soundness: The Effect of Product ternal cues from the environment can augment consumers’ senso- Operation Sound on Judgment of Product Quality ry experiences. Importantly, the session will be of interest to both B. Kyu Kim, Yonsei University, South Korea academics and practitioners. For academics, the findings will open Byung Geun Cho, Yonsei University, South Korea new research directions, such as studying other internal cues that can He (Michael) Jia, The University of Hong Kong, China moderate the effect of sensory inputs and other external cues that can Paper #4: The Effect of Unpacking on Consumers’ Sensory affect sensory experiences. For practitioners, the papers provide easy Experience: A Goal-gradient Account to implement recommendations that can improve customer attitudes Claire Tsai, University of Toronto, Canada and satisfaction. Min Zhao, Boston College, USA Sounding Warm: The Role of Audio Pitch on Service SESSION OVERVIEW Perception Sensory experiences play a profound role in consumer decision making (Krishna 2011). For instance, consumers form ex-ante opin- EXTENDED ABSTRACT ions about products and brands while listening to ads on the radio or In 2016, the global radio advertising market reached a value podcasts. Consumers also make ex-post evaluations of a consump- of $32.5 billion. However, little research has looked into how audio tion episode, based on their sensory experience. Despite this impor- characteristics influence consumer decisions (Krishna 2012; Lowe tance, the literature (for a review, see Krishna 2012) surprisingly and Haws 2016). Moreover, the select work in the area has empha- lacks insight into how internal cues of sensory inputs, such as a mes- sized the advantage of low (vs. high) vocal pitches (Lowe and Haws sage’s audio pitch, can affect consumer decision making. Moreover, 2016; Tigue et al. 2012). We add to this literature by looking at how little research has looked into how marketers may enhance consum- higher vocal pitches may positively impact consumer evaluations, by ers’ sensory experiences based on the same objective consumption facilitating perceptions of warmth (vs. competence) for firms. episode. To contribute to the literature on sensory marketing and Perceptions of warmth, which are related to perceptions of so- provide practical guidelines for marketers, the current session poses cial intent (e.g. friendliness, helpfulness, & sincerity), are one of the two research questions: A) How do internal cues of sensory expe- most fundamental forms of social perception (Judd et al. 2005). We riences, such as audio pitch, affect consumer evaluations? and B) argue that such perceptions of warmth may be influenced by audio What external cues can marketers use to enhance consumers’ sensory pitch in advertisements. The literature provides support for this hy- experiences? By studying sensory experiences in regards to both in- pothesis. For example, parents speak in higher vocal pitches when ternal and external cues, the session aims to provide a well-rounded talking to babies (motherese) (Grieser and Kuhl 1988; Shute 1987) perspective to researchers and practitioners. and people perceive robots with higher voices as having better social The first three papers examine how internal cues, such as audio skills and being more pleasant (Niculescu et al. 2013). In contrast, pitch, can influence consumer evaluations, behavior, and judgment. people see low (vs. high) pitch as signaling roughness or anger (Eitan Han and Lowe look into how differences in the audio pitch of ad- and Timmers 2010). vertising messages can influence consumer evaluations of compa- In short, we predict that high (vs. low) pitched advertisements nies. Their paper shows that high (vs. low) pitched audio messages will lead to increased perceptions of warmth and that this will ulti- increase perceptions of warmth (vs. competence) for firms and that mately lead to more positive evaluations for services where warmth this in turn enhances people’s evaluation and preference for services is important. Across five studies we provide evidence for our claims where warmth is important. The second paper by Lowe and Hyun and the underlying process. investigates how vocal similarity between users and AI agents affect trust. Using a novel methodology to quantify vocal similarity, Study 1 the paper finds that higher vocal similarity leads to greater trust for In study 1, we looked at how vocal pitch influences responses AI agents. In the third paper, Kim, Cho, and Jia look at how the toward a senior-care service ad and a financial service ad. The ser- operation sound of products affect consumers’ inferences about the vices were selected based on a pretest, which indicated that consum- product. Specifically, they find that low (vs. high) pitched operation ers prioritize warmth over competence for senior-care services, while sounds lead to higher quality inferences but does not affect infer- they prioritize competence over warmth for financial services. Given ences about user-friendliness. this result, we predicted that a high (vs. low) pitched spokesperson The last paper of the session broadens the view and highlights would be more effective for a senior-care service, but not for a finan- external cues, such as framing effects, that can enhance consumers’ cial service.

Advances in Consumer Research 907 Volume 48, ©2020 908 / How You Feelin’? The Role of Internal And External Cues on Consumer Sensory Experiences The main study had a 2 (Pitch: high vs. low) X 2 (Service: Study 4 senior-care vs. financial) between-subjects design. In the study, par- Study 4 tested our effect in a field setting. In the study, a vol- ticipants first listened to a purported radio ad of a senior-care ser- unteer from the Salvation Army stood next to a donation box, while vice or a financial service. The ad was modulated to be higher or ringing a bell. Alternating each half hour, the volunteer either rang a lower in pitch. After listening to the ad, participants answered three higher or lower pitched handbell. The results showed that shoppers dependent measure items (likelihood to consider firm, likelihood to were more likely to donate when the pitch of the bell being rung was recommend firm to friends, & expected satisfaction). Next, partici- higher (vs. lower) (χ2 (1) = 4.83, p = .028). pants completed a scale, where we measured perceived warmth and perceived competence of the firm. Vocal Similarity, Trust and Persuasion in Human-AI A two-way ANOVA revealed a significant interaction effect of Agent Interactions vocal pitch and industry on the dependent measure index (α = .91) (F(1, 126) = 7.93, p < .01), where high (vs. low) vocal pitch was more EXTENDED ABSTRACT effective for the senior-care ad (F(1, 126) = 4.38, p = .04), while low AI agents (e.g., Apple Siri, Amazon Alexa, Microsoft Corta- (vs. high) vocal pitch was marginally more effective for the financial na, and social robots including Pepper) have become increasingly services ad (F(1, 126) = 3.32, p = .06). Moreover, we ran a moder- prevalent in our daily lives, our homes, and into our workplaces. ated mediation analysis, which showed that the mediation was only Voice recognition technology via machine learning has reached the significant for the senior-care condition (95% CI: [.0440, .4184]), accuracy level of human speech (Arnold, 2018) and algorithms will and not in the financial service condition (95% CI: [-.2224, .1561]). soon recognize all the various aspects in speech including nuances and vocal characteristics (i.e., tonal inflection, mood; Kirby, 2019). Study 2 Also, the use of voice recognition will be advanced to include per- Study 2 aimed to test the generalizability of our effect. First, we sonalization features, similar to face ID or PIN, so that the AI agent tested whether voice pitch could also affect people’s political deci- can respond accordingly to the identified user through the formants sion making. Second, we used a female speaker rather than a male or sound characteristics unique to each person’s vocal tract. This re- speaker. In the study, participants listened to a high or low pitched au- search examines acoustic similarity between an individual consumer dio clip, which was purportedly a speech given by a female politician and the AI agent. We ask this question: how will consumers perceive, in response to a natural disaster. Next, participants rated how much respond to, and be persuaded by an AI agent contingent on similari- they thought the politician would be willing to sacrifice herself, how ties between that agent’s voice and their own voice in both a) pitch, concerned she seemed, and how much they liked the politician. The and b) timbre? three items were averaged into a DV index (α = .84). Replicating our Pitch and timbre are the two most prominent features allowing prior finding, a one-way ANOVA revealed that participants rated the humans or machines to distinguish one voice from another. Pitch female politician higher on the DV index when they heard the higher refers to the “fundamental frequency” of a sound, while timbre is (vs. lower) pitched speech (F (1, 115) = 5.32, p = .02). the unique spectrum of frequencies within a sound, including one’s voice, (here measured in Mel Frequency Cepstral Coefficients [MF- Study 3A & 3B CCs; Logan, 2000]). Pitch of a spokesperson’s voice, for instance, Studies 3A and 3B tested several boundary conditions. In Study can influence message acceptance or automatically convey certain 3A, we tested how individual difference in preference for warmth product attributes (Chattopadhyay et al., 2003; Lowe and Haws, versus competence in a service provider moderates our effect. Here, 2017). Timbre largely accounts for our ability to distinguish between participants listened to either a higher or lower pitched an audio clip voices or instruments, and can also affect consumer perception in that was purportedly from a work out session of a personal fitness various ways (Bruner 1990). In this work, we ask how differences trainer. Next, participants indicated how likely they would be to hire between pitch and timbre in a consumer’s voice and an AI voice the fitness trainer. Finally, participants were asked whether they val- might influence consumer response. ued warmth or competence from a service provider. Consistent with A prevalent body of research in social psychology advocates our theory, a regression analysis revealed a significant interaction for the similarity-attraction effect (Collisson & Howell, 2014; Mon- effect of audio pitch and preference for warmth (F (1, 134) = 4.28, toya, Horton, & Kirchner, 2008), which suggests that we are more p = .04), where participants who preferred warmth showed a greater likely to prefer and be initially attracted to similar others. Although willingness to hire the trainer when they listened to the high (vs. low) research suggests that we prefer to interact with advanced machines pitched recording (F (1, 134) = 3.47 = .06). However, there was no that are more human in nature (Nass et al., 1995; Tapus & Mataric, effect of pitch for those who preferred competence (NS). 2007), the uncanny valley theory (Mori, 1970) suggests that the de- In study 3B, we looked at how chronic differences in relation- gree to which these machines are similar may play a role in that we ship norms moderated the effect. Based on prior research (Li et al. experience an eerie sensation and discomfort when they become too 2018), we predicted that those who have a communal norm will pri- similar to ourselves. oritize warmth over competence and therefore prefer a service pro- We explore how objective vocal similarity between an AI agent vider with a higher pitched voice. Conversely, we predicted a null and a consumer in both pitch and timbre influences user perceptions effect for those with an exchange norm. The experimental stimuli of warmth, competence, and trust in the agent. In our series of stud- for Study 3B was the same as 3A, except that we measured partici- ies, after listening to a single (study 1 and 2) or three different (study pants chronic relationship norms. Consistent with our expectations, 3) AI agents’ recommendation of different products (books, SNL a regression analysis showed that those with a communal norm pre- videos, and movies), we measure participant’s perception of the AI ferred the higher (vs. low) pitched trainer (F (1, 229) = 6.12, p = .01), agent and actual choice of recommended products. We later record whereas there was no effect of pitch for those who had an exchange and analyze each participant’s voice to create an objective measure norm. of acoustic similarity to the AI voice. This objective measure of vocal similarity in pitch and timbre (MFCCs) introduces a new methodological approach measuring the Euclidian distance between the Advances in Consumer Research (Volume 48) / 909

AI agent and each participant’s voice. We demonstrate the effect of sound as operation sound (i.e., sound produced as a consequence of vocal similarity on persuasion and trust (competence, benevolence, operation). Signal sound has been considered as an important design integrity) further showing how listener’s self-esteem plays a role in factor in product design. In contrast, operation sound is often con- moderating sidered merely as “noises” that should be eliminated. In the current Study 1 was conducted with 152 undergraduates. Participants research, we propose that even operation sound can be an important listened to a sample of an AI voice (gender matched) and provided design element that positively influences consumers’ evaluation of their impressions regarding the AI agent. Each participant also had products. Specifically, we argue that when a product produces op- a sample of their voice recorded upon conclusion of the study. Ob- eration sound in a lower pitch, consumers judge the product to be jective similarity between AI and participant voices was calculated “sounder,” compared to when it produces sound in a higher pitch. using a computer algorithm created for this research. Overall, greater We base our prediction on findings from the biology litera- similarity in vocal pitch significantly predicted satisfaction with the ture, which has demonstrated that voice pitch serves as a signal for agent (F = 8.19, p < .01), willingness to further interact with the one’s physical quality (Collins 2000; Jones et al. 2010; Puts et al. agent (F = 4.40, p < .05), perceived sincerity of the agent (F = 4.37, 2012). Both voice pitch and physical quality are influenced by the p < .05), and the overall trust in the agent (F = 7.75, p < .01). level of testosterone, and hence one’s voice pitch tends to be cor- Study 2 (187 undergraduates), used a similar procedure to related with his or her physical quality. Consequently, people tend Study 1, while also including a choice task based on the AI voice’s to judge the physical quality of others based on their voices (Fein- recommendation. In addition, we measure participant’s levels of berg et al. 2008). Those who make lower voices are judged to be self-esteem. We hypothesized that one’s level of self-esteem could healthier, more fertile, and better in physical ability, compared to influence individuals’ responses to the subtle effect of vocal similar- those who make higher voices. Building on the relationship between ity. The more similar the participant’s voice was to the AI agent’s voice pitch and physical quality, we investigate whether the pitch of voice in terms of timbre (MFCCs), the participant was more likely to a product’s operation sound can serve as a signal for the product’s choose the video recommended by the AI agent (MFCC: B = -22.40, physical quality in consumers’ perception. Specifically, we predict SE = 10.96, p < .05). Moreover, the interaction between timbre simi- that products that make lower-pitch operation sound are judged to larity and self-esteem was significant, such that higher self-esteem have better quality than products that make higher-pitch sound. In individuals were actually less influenced by vocal similarity (self- other words, products producing a lower-pitch sound is judged to be esteem x MFCC: B = 3.83, SE = 2.00, p = .055). “sounder” than products producing a higher-pitch sound. In Study 3, we introduce three distinct voices of AI agents recommending three different movies. The results again show that simi- Studies 1A-1C larity in timbre significantly predicts evaluations of warmth (MFCC: Studies 1A (N = 68), 1B (N = 67), and 1C (N = 114) aim to B = -1.82, SE = .625, p < .01) and overall trust in the agent (MFCC: generalize the effect of product operation sound on product qual- B = -1.27, SE = .49, p < .05). Moreover, pitch similarity significantly ity judgement across several product categories, including blenders predicts liking of the agent (B = -.40, SE = .07, p < .001) and two (Study 1A), pens (Study 1B), and hand dryers (Study 1C). All three trust dimensions: competence (B = .14, SE = .05, p < .05) and be- studies used a one-way, two-cell between-subjects, in which partici- nevolence (B = .12, SE = .0, p < .001). We also replicate the moder- pants were randomly assigned to either a high-pitch condition or a ating role of self-esteem on vocal similarity on trust in the AI agent low-pitch condition. To create high- and low-pitch stimuli, we al- (MFCC: B = -5.55, SE = 2.61, p < .05; self-esteem x MFCC: B = .82, tered the pitch of a product’s operation sound using an audio-editing SE = .48, p = .087). software while keeping other aspects of the sound wave constant. Present research aims to contribute to literatures in psycho- In the studies, participants first listened to an audio clip recording a acoustics, similarity-attraction effect, and human-computer interac- product’s operation sound and then judged product’s physical qual- tion. We find that, overall, similar timbre and dissimilar pitch was ity (e.g., durability, sturdiness, and performance). The results of favored. Furthermore, we believe that our research offers useful these three studies consistently show that participants perceived the implications to marketers of new technology devices. Although cer- product to have better quality in the low-pitch condition than in the tain voice-relevant cues (e.g., accent, conversational styles) may be high-pitch condition, and such a difference was not driven by per- more direct and pronounced to individuals when engaging in an in- ceptions of a product’s other attributes, such as luxuriousness and teraction, we explore a very subtle cue that subconsciously influence heaviness. Studies 1A-1C convergently provide initial evidence for consumers’ mindset. Marketers will be able to better understand the our proposed effect of product operation sound on product quality mechanisms and conditions under which we prefer AI agents that are judgement. more personalized to sound similar in pitch and timbre to individual consumers. Study 2 Study 2 (N = 78) aims to further demonstrate that the effect Sound of Products’ Soundness: The Effect of Product of product operation sound on product judgement is specific to ca- Operation Sound on Judgment of Product Quality pability-related judgment (i.e., quality) but not to usability-related judgment (i.e., user-friendliness). For this purpose, Study 2 adopted EXTENDED ABSTRACT a 2 (Pitch: high vs. low) x 2 (Judgment Dimension: quality vs. user- In everyday lives, people are exposed to various sounds that friendliness) mixed design with pitch as a between-subjects factor products produce. Some sounds are electronic or digital sounds that and judgment dimension as a within-subjects factor. Participants first are added to products as an auditory signal (e.g., smartphone ring listened to the operation sound of a printer and then rated the printer tones, computer alarm sounds, or microwave beeps). Other sounds in terms of its quality and user-friendliness. There was a significant are mechanical sounds that are produced when products operate pitch × judgment dimension interaction (F(1, 76) = 4.04, p < .05), (e.g., hair dryers’ whirring sounds, printers’ chugging sounds, or car such that participants judged the same printer to have better quality doors’ chunking sounds). We refer to the former type of sound as sig- when it made a low-pitch sound compared to when it made a high- nal sound (i.e., sound added as a signal to users) and the latter type of pitch sound (Mhigh = 4.36, vs. Mlow = 5.20; F(1, 76) = 9.27, p < .01) 910 / How You Feelin’? The Role of Internal And External Cues on Consumer Sensory Experiences while their judgment of the printer’s user-friendliness did not differ Five experiments were conducted to test our hypotheses. In all between the high- and the low-pitch conditions (Mhigh = 4.77 vs. Mlow experiments, to manipulate event framing, we presented each sub- = 5.01; F(1, 76) = 1.14, p = .29). Study 2 shows that the pitch of a component either using a progress bar or pie chart (packing frame), product’s operation sound influences the product’s perceived quality such that the subcomponents were perceived as parts of one sensory but not its perceived user-friendliness. experience, or using integers (unpacking frame), such that the subcomponents were perceived as individual sub-activities. To manipu- Study 3 late focus of consumption progress, we asked participants to track The purpose of Study 3 is to demonstrate that low brand reputa- their consumption progress in ascending or descending order which tion serves as a boundary condition for the effect of product opera- emphasized, respectively, how much had been consumed (to-date – tion sound on quality judgment. This study (N = 186) adopted a 2 unpacking frame: 1, 2…; packing frame: 1/10, 2/10…) or how much (Pitch: high vs. low) x 2 (Brand reputation: high vs. low) between- was remaining (to-go – unpacking frame: 10, 9…; packing frame: subjects design. Participants were informed that they listened to the 10/10, 9/10…). operation sound of a camera produced either by Canon (in the high- In Experiment 1 with a 2 (event framing: packing vs. unpack- reputation condition) or Sanyo (in the low-reputation condition) and ing) x 3 (progress focus: to-go vs. to-date vs. control) between-sub- then evaluated the camera in terms of its quality. There was a pitch jects design, participants watched a short video featuring 10 serene × brand reputation interaction (F(1, 182) = 2.91, p = .09) at 10% beach images, and reported their overall evaluation based on 10-point alpha level. Simple effect analysis revealed that participants judged scales. We found an expected two-way interaction (F(2, 439) = 5.42, the same camera to have better physical quality in the low-pitch con- p = .005): With a to-go-focus, unpacking increased evaluation (to- dition than in the high-pitch condition (M = 5.21vs. M = 5.77; high low go focus: Munpacking = 6.65 vs. Mpacking = 5.98, F(1, 439) = 3.76, p = F(1, 182) = 4.22, p = .04) only when the brand reputation was high .053), but the effect reversed with a to-date-focus (Munpacking = 5.98 vs. but not when it was low (M = 4.83 vs. M = 4.81; F(1, 181) = high low Mpacking = 6.68; F(1, 439) = 3.75, p = .053). Control conditions where .01, p = .93). progress was not tracked resembled to-go focus (Munpacking= 6.68 vs. To our best knowledge, the current work is the first empirical Mpacking = 5.90, F(1, 439) = 5.23, p = .02). research in the marketing literature to examine the pitch of prod- In experiment 2A with a 2 (event framing: packing vs. unpack- uct operation sounds and its influence on consumers’ judgment. Our ing) × 2 (progress focus: to-go vs. to-date) between-subjects de- findings have important implications for sensory marketing and sign, we asked participants to sample six new curry flavored potato product design. chips. We measured moment-by-moment enjoyment of each chip, retrospective evaluation after eating all six chips, and repeated con- The Effect of Unpacking on Consumers’ Sensory sumption in a free eating session later. Again, a two-way ANOVA Experience: A Goal-gradient Account revealed only a significant interaction on overall enjoyment (F(1, 115) = 16.28, p < .001): With a to-go focus, unpacking increased EXTENDED ABSTRACT enjoyment (Munpacking = 8.13 vs. Mpacking = 6.10; F(1, 115) = 12.80, p Sensory experiences, such as watching a scenery films or eat- = .001), but the effect reversed with a to-date focus (Munpacking = 6.38 ing chocolate, have multiple subcomponents, and consumers can vs. Mpacking = 7.70; F(1, 115) = 4.73, p = .03). Similar patterns were track the progress of enjoying them. These subcomponents can be observed in real-time enjoyments (three-way interaction: F(5, 575) = framed either as multiple distinct sensory consumption units (i.e., 3.39, p = .005) and in the amount of free eating (F(1, 115) = 11.45, p unpacking frame) or as parts of one whole sensory consumption = .001). We replicated experiment 2A in experiment 2B using sweets (i.e., packing frame). Prior research has shown that people predict and a tracking card that resembles ordering sheets commonly used in an unpacked sensory experience to be more enjoyable than a packed all-you-can-eat restaurants (two-way interaction: F(2, 156) = 4.27, experience because unpacking segregates the sensory experience or p = .01). “gains” (Kahneman and Tversky 1979; Thaler 1985, Tsai and Zhao In experiment 3, we tested the goal-gradient account by mea- 2011). The present research extends this line of work in two impor- suring consumption motivation and including a theory-driven tant ways. First, we expand the effect of unpacking onactual sensory boundary condition where we inserted a break during consumption experiences. Second, we identify focus of consumption progress (fo- (Kahneman and Snell, 1990; Nelson and Meyvis, 2008). Taking a cusing on progress to-date vs. focusing on progress to-go; e.g., I have break during consumption under the to-date focus should reset con- two chocolate truffles left vs. I already ate four chocolate truffles) as sumption motivation and weaken the negative effect of unpacking. an important moderator of the effect of unpacking on actual sensory Specifically, we used a 2 (event framing: packing vs. unpacking) × 3 experiences, and we demonstrate goal-gradient (consumption moti- (progress focus: to-go, to-date, to-date/break) between-subjects de- vation) as the mediator. sign and asked participants to view their preferred image six times We predict that under a to-go focus, the marginal benefit of repeatedly. We measured real-time enjoyment create an enjoyment completing one consumption subcomponent increases as one ap- index (α = .97 for 6 trials), and collected motivation measures: how proaches the end state of the sensory experience, thereby producing alert, how motivated, and how much effort they put forth during the a classic increasing goal gradient (Hull 1934, Kivetz, Urminsky, and experience (α = .73). Again, we found a significant two-way interac- Zheng 2006). As such, consumers feel more motivated and positive tion on overall enjoyment (F(2, 339) = 4.04, p < .05) with planned about their consumption progress toward the end of a sensory expe- contrasts (to-go focus: Munpacking = 7.78 vs. Mpacking = 7.11; F(1, 339) = rience and enjoy it even more. However, under a to-date focus, the 3.32, p = .06; to-date focus: Munpacking = 7.15 vs. Mpacking = 7.90, F(1, marginal benefit of completing one subcomponent decreases as one 339) = 4.16, p < .05). Importantly, when participants took a break moves away from the initial state. Therefore, one feels less motivat- under to-date focus, the negative effect of unpacking was attenuated ed and less positive when the distance from the initial state increases (Munpacking = 7.96 vs. Mpacking = 7.67; F(1, 339) = .94, p = .33). Finally, and experiences a decreasing goal gradient (Bonezzi, Brendl, and De we observed a significant two-way interaction on consumption mo- Angelis, 2011; Huang, Zhang, and Broniarczyk 2012). tivation (F(2, 339) = 7.51, p = .001) and confirmed its mediation role (unbiased 95% CI = [-.60 to -.14]; 5,000 bootstrap resamples; Advances in Consumer Research (Volume 48) / 911

Preacher and Hayes 2008). Experiment 4 successfully replicated ex- Jones, Benedict C., David R. Feinberg, Lisa M. DeBruine, Anthony periment 3 using a different packing manipulation – pie chart (two- C. Little, and Jovana Vukovic (2010), “A Domain-Specific way interaction: F(2, 249) = 4.16, p < .01). Oppositesex Bias in Human Preferences for Manipulated Our work advances the area of studies on unpacking (Tversky Voice Pitch,” Animal Behaviour, 79 (1), 57–62. and Koehler, 1994) and sensory experiences (Hsee et al., 2012) by Judd, Charles M., Laurie James-Hawkins, Vincent Yzerbyt, and demonstrating a counterintuitive reversal of unpacking with a to- Yoshihisa Kashima (2005), “Fundamental Dimensions of date focus (Bonezzi, Brendl, and De Angelis, 2011; Koo and Fish- Social Judgment: Understanding the Relations between bach, 2008). These findings have important theoretical contribution Judgments of Competence and Warmth,” Journal of as well as interesting practical implications to better manage sensory Personality and Social Psychology, 89(6), 899–913. experiences. Kahneman, D., and Snell, J. (1990). Predicting utility. In R.M. Hogarth (Ed.), Insights in decision making: A tribute to Hillel REFERENCES J. Einhorn (pp. 295-310). Chicago, IL: University of Chicago Arnold (2018). A.I. and speech advances bring virtual assistants Press. to work. Retrieved from https://www.computerworld.com/ Kahneman, D., and Tversky, A. (1979). Prospect theory: An article/3264433/ai-and-speech-advances-bring-virtual- analysis of decision under risk. Econometrica, 47, 263–292. assistants-to-work.html Kirby (2019). What’s The Future of Voice Recognition? Retrieved Bonezzi A, Brendl, C. M., and De Angelis M. (2011). Stuck in the from https://www.futureofeverything.io/future-voice- middle: The psychophysics of goal pursuit. Psychological recognition/ Science, 22(5), 607-612. Kivetz, R., Urminsky, O., and Zheng, Y. (2006). The goal-gradient Bruner, Gordon C. (1990), “Music, Mood, and Marketing,” Journal hypothesis resurrected: Purchase acceleration, illusionary of Marketing 54, (4), 94-104. goal progress, and customer retention. Journal of Marketing Chattopadhyay, Arnitava, Darren W. Dahl, Robin J. B. Ritchie, Research, 43, 39–58. and Kimary N. Shahin (2003), “Hearing Voices: The Impact Krishna, Aradhna. (Ed.). (2011). “Sensory marketing: Research on of Announcer Speech Characteristics on Consumer Response the sensuality of products.” Routledge. to Broadcast Advertising,” Journal of Consumer Psychology, Krishna, Aradhna (2012), “An Integrative Review of Sensory 13(3), 198–204. Marketing: Engaging the Senses to Affect Perception, Collins, Sarah A. (2000), “Men’s Voices and Women’s Choices,” Judgment and Behavior,” Journal of Consumer Psychology, Animal Behaviour, 60 (6), 773–80. 22(3), 332–51, http://dx.doi.org/10.1016/j.jcps.2011.08.003. Cuddy, Amy J.C., Peter Glick, and Anna Beninger (2011), “The Krishna, Aradhna and Norbert Schwarz (2014), “Sensory Dynamics of Warmth and Competence Judgments, and Their Marketing, Embodiment, and Grounded Cognition: A Review Outcomes in Organizations,” Research in Organizational and Introduction.” Journal of Consumer Psychology, 24, (2), Behavior, 31, 73–98. 159-168. Eitan, Zohar and Renee Timmers (2010), “Beethoven’s Last Piano Li, Xueni (Shirley), Kimmy Wa Chan, and Sara Kim (2018), Sonata and Those Who Follow Crocodiles: Cross-Domain “Service with Emoticons: How Customers Interpret Employee Mappings of Auditory Pitch in a Musical Context,” Cognition, Use of Emoticons in Online Service Encounters,” Journal of 114(3), 405–22. Consumer Research, 45(5), 973–87. Feinberg, David R., Lisa M. DeBruine, Benedict C. Jones, and Logan, Beth. “Mel Frequency Cepstral Coefficients for Music Anthony C. Little (2008). “Correlated Preferences for Modeling,” In Ismir, vol. 270, pp. 1-11. 2000. Men’s Facial and Vocal Masculinity,” Evolution and Human Lowe, Michael L. and Kelly L. Haws (2016), “Sounds Big: The Behavior, 29 (4), 233–41. Effects of Acoustic Pitch on Product Perceptions,”Journal of Fiske, Susan T., Amy J.C. Cuddy, and Peter Glick (2007), Marketing Research, 54(2), 331–46. “Universal Dimensions of Social Cognition: Warmth and Montoya, R. M., Horton, R. S., & Kirchner, J. (2008). Is actual Competence,” Trends in Cognitive Sciences, 11(2), 77–83. similarity necessary for attraction? A meta-analysis of actual Grieser, DiAnne L. and Patricia K. Kuhl (1988), “Maternal Speech and perceived similarity. Journal of Social and Personal to Infants in a Tonallanguage: Support for Universal Prosodic Relationships, 25(6), 889-922. Features in Motherese,” Developmental Psychology, 24(1), Mori, M. (1970). The uncanny valley. Energy, 7(4), 33-35. 14–20. Nass, C., Moon, Y., Fogg, B. J., Reeves, B., & Dryer, D. C. Hayes, Andrew F (2012), PROCESS: A Versatile Computational (1995). Can computer personalities be human personalities?. Tool for Observed Variable Mediation, Moderation, and International Journal of Human-Computer Studies, 43(2), Conditional Process Modeling [White Paper]. Retrieved from 223-239. Http://Www.Afhayes.Com/ Public/Process2012.Pdf. Nelson, L. D., and Meyvis, T. (2008). Interrupted consumption: Hsee, C. K., Shen, L., Zhang, S., Chen, J., and Zhang, L. (2012). Disrupting adaptation to hedonic experiences. Journal of Fate or fight: Exploring the hedonic costs of competition. Marketing Research, 45(6), 654-664. Organizational Behavior and Human Decision Processes, Niculescu, Andreea, Betsy van Dijk, Anton Nijholt, Haizhou Li, 119(2), 177-186. and Swee Lan See (2013), “Making Social Robots More Huang, S., Zhang, Y., and Broniarczyk, S. M. (2012), “So near Attractive: The Effects of Voice Pitch, Humor and Empathy,” and yet so far: The mental representation of goal progress,” International Journal of Social Robotics, 5(2), 171–91. Journal of Personality and Social Psychology, 103(2), 225- Puts, David A., Benedict C. Jones, and Lisa M. DeBruine (2012), 241. “Sexual Selection on Human Faces and Voices,” Journal of Hull, C.L. (1934). The rat’s speed-of-locomotion gradient in the Sex Research, 49 (2-3), 227–43. approach to food. Journal of Comparative Psychology, 17, 393–422. 912 / How You Feelin’? The Role of Internal And External Cues on Consumer Sensory Experiences

Preacher, K. J., and Hayes, A. F. (2008). Asymptotic and Thaler, R. H. (1985). Mental accounting and consumer choice. resampling strategies for assessing and comparing indirect Marketing Science, 4, 199–214. effects in multiple mediator models.Behavior Research Tsai, C. I., and Zhao, M. (2011). Predicting consumption time: The Methods, 40, 879–891. role of event valence and unpacking. Journal of Consumer Reeves, B., & Nass, C. I. (1996). The media equation: How people Research, 38, 459–473. treat computers, television, and new media like real people Tversky, A., and Koehler, D. J. (1994). Support theory: A and places. Cambridge university press. nonextensional representation of subjective probability. Shute, H. Brenda (1987), “Vocal Pitch in Motherese,” Educational Psychological review, 101(4), 547. Psychology, 7(3), 187–205. Wang, Ze, Huifang Mao, Yexin Jessica Li, and Fan Liu (2016), Smith, Stephen M. and David R. Shaffer (2007), “Speed of Speech “Smile Big or Not? Effects of Smile Intensity on Perceptions and Persuasion: Evidence for Multiple Effects,”Personality of Warmth and Competence,” Journal of Consumer Research, and Social Psychology Bulletin, 21(10), 1051–60. 43(September), ucw062, https://academic.oup.com/jcr/article- Tapus, A., & Mataric, M. J. (2007, March). Emulating Empathy lookup/doi/10.1093/jcr/ucw062. in Socially Assistive Robotics. In AAAI spring symposium: multidisciplinary collaboration for socially assistive robotics (pp. 93-96). Tigue, Cara C., Diana J. Borak, Jillian J. M. O’Connor, Charles Schandl, and David R. Feinberg (2012), “Voice Pitch Influences Voting Behavior,” Evolution and Human Behavior, 33, 210–16, http://dx.doi.org/10.1016/j. evolhumbehav.2011.09.004. Building a More Prosocial and Sustainable Society: Associative Processes in Prosocial Behavior Chairs: Uzma Khan, University of Miami, USA Colton Pond, University of Miami, USA

Paper #1: Negative Effects of Cause-Related Marketing for The authors find that recycling is related to the associations of moral Charitable Organizations goodness and environmental benefit; thus, regardless of the recy- Colton Pond, University of Miami, USA clability of an item, consumers recycle products to compensate for Uzma Khan, University of Miami, USA low moral self-worth. The last paper by Longoni, Tari, and Trudel furthers our understanding regarding how consumers evaluate cor- Paper #2: When Sustainability Is Not a Liability: How Going porations who engage in green-related price increases. The authors Green Can Benefit Perceived Product Performance show that consumers relate sustainable manufacturing initiatives as Alexander Chernev, Northwestern University, USA insights into a firm’s morality, which has a halo effect on the percep- Sean Blair, Georgetown University, USA tions of a firm. Consequently, consumers perceive that green-related Paper #3: Over-Recycling: When Motivated Sustainability Goes price increases are fairer relative to general manufacturing price in- Wrong creases. Atar Herziger, The Ohio State University, USA The four papers in the session are all in advanced stages and Grant E. Donnelly, The Ohio State University, USA shed light on how to build a more prosocial and sustainable society. Rebecca Walker Reczek, The Ohio State University, USA Together, the papers help us to understand how associative processes Paper #4: Transparent Green Practices Boost Perceptions of influence how consumers engage in and perceive firms who engage Price Fairness in prosocial behaviors. Considering the growth in prosocial initia- Chiara Longoni, Boston University, USA tives both in the consumer and corporate sectors, we expect that this Anna Tari, Boston University, USA session will appeal to researchers as well as practitioners interested Remi Trudel, Boston University, USA in corporate-social responsibility, cause-related marketing, sustainability, charitable giving, and decision-making more generally. SESSION OVERVIEW Both firms and consumers are increasingly engaging in proso- Negative Effects of Cause-Related Marketing for cial behaviors. In 2016, 82% of S&P 500 corporations promoted cor- Charitable Organizations porate-social responsibility as a significant business strategy (3BL 2016), sustainable products amounted to over half of the total prod- EXTENDED ABSTRACT ucts sold in the US market (Nielsen 2018), and consumers donated Consumers are increasingly expecting firms to be more proso- $286.65 billion to charities (Charity Navigator 2017). There has been cial, as 86% of consumers from a national survey indicated that firms a corresponding increase in research aimed at understanding how have a responsibility to improve both society and the environment consumers engage in and perceive firms who engage in prosocial be- (Cone Communications 2017). Consequently, firms engage in cause- haviors (Carroll and Shabana 2010). A conclusion from this literature related marketing, where they partner with charities with the goal of is that consumers view prosocial behaviors in a positive light, and increasing a firm’s sales while promoting a charitable cause. Existing these positive associations can spill over through an associative pro- research has demonstrated several positive outcomes of cause-relat- cess in several ways (Aquino and Reed 2002). The current session ed marketing for firms (e.g., brand perceptions, purchase likelihood; introduces four papers that provide novel insights regarding how as- c.f., Carroll and Shabana 2010); however, consequences of cause-re- sociative processes influence prosocial behaviors by understanding lated marketing for charities have largely been ignored. The current how consumers engage in and perceive firms who engage in proso- research fills this void by examining when and why cause-related cial behavior. Some of the questions raised in this session include, marketing may have positive versus negative outcomes for charities. what associations do consumers have with prosocial and sustainable The beneficial outcomes of cause-related marketing for firms actions; what factors influence consumers’ prosocial behaviors such are believed to arise because the positive associations with charities as charitable donations and recycling; how do consumers evaluate spillover to cast a halo over the partnering firms (Till and Nowak firms who promote their sustainable initiatives? 2000). Given that associative processes are typically symmetrical The first paper by Pond and Khan examines how cause-related (Collins and Loftus 1975), we posit that in a cause-related marketing marketing may influence outcomes for charities. They propose that partnership, just as the associations with the charity spillover to influ- a cause-related marketing partnership with a hedonic (versus utili- ence perceptions of the firm, associations with the firm can spillover tarian) product can reduce subsequent donations from individuals to influence perceptions of the charity. One factor that mayinflu- because the frivolous and unnecessary associations with a hedonic ence the nature of the firm’s spillover effects on a charity is whether product (Dhar and Wertenbroch 2000) can spill over to decrease the the firm’s products are hedonic or utilitarian. Compared to utilitarian perceived importance of the charitable cause. The second paper by products, hedonic products are perceived to be frivolous and unnec- Chernev and Blair proposes that sustainability claims can increase essary (Dhar and Wertenbroch 2000). Therefore, we predict that in product performance, contrary to prior research regarding the sus- cause-related marketing partnerships with hedonic products, these tainability liability. Specifically, the authors show that a firm’s sus- negative associations can spillover to adversely affect the perceived tainable practices can lead consumers to associate a moral undertone importance of the charity’s cause, thus reducing donations provided with the firm resulting in a positive halo effect on enhancing product to the charity by individuals. performance. The next paper by Herziger, Donnelly, and Reczek ex- In all studies, participants had an opportunity to receive a bonus amines when consumers over-recycle by recycling non-recyclables. payment. Participants were shown an ad for either a hedonic (hedonic partnership) or utilitarian product (utilitarian partnership) that Advances in Consumer Research 913 Volume 48, ©2020 914 / Building a More Prosocial and Sustainable Society: Associative Processes in Prosocial Behavior donated to a specified charity. The ad did not solicit any action, but efits with the company rather than with its products and by increas- in a subsequent unrelated task, participants were asked how much ing the prominence of the societal benefits of sustainability. We test of their bonus they would like to donate to a charity. All products the impact of promoting sustainability benefits on perceived product were pretested to vary only on a hedonic-utilitarian dimension while performance and the factors that are likely to moderate this effect in remaining constant on perceived attractiveness, purchase likelihood, a series of three empirical studies. and expectations of social responsibility. Our first experiment aims to document the basic proposition In study 1, participants (N=175) were shown an ad for either a that promoting sustainable product benefits can strengthen rather hedonic or utilitarian backpack that donated to Alzheimer’s Research than weaken the perceived performance of a company’s products. Society (ARS). A control (no partnership) condition was also includ- The data show that sustainability is not necessarily a liability, even ed. Next, in an unrelated task, participants were given an opportu- in the case of strength-related attributes when the negative impact nity to donate any amount of their bonus to ARS. As expected, par- of sustainability is arguably most pronounced (Luchs et al. 2010) ticipants donated less to ARS after viewing the hedonic partnership and identify scenarios in which sustainability can strengthen rather (M=$0.079) compared to both the utilitarian partnership (M=$0.127) than weaken consumers’ product performance beliefs. Specifically, and no partnership (M=$0.116, p=.031). In study 2 (N=269), we the data lend support to the argument that the impact of sustain- replicated these findings and measured how important participants ability on perceived product performance is a function of the moral thought was the cause of ARS. A mediation analysis revealed that undertone of consumers’ beliefs reflected in their environmental at- the effect of partnership type on donation amount was explained by titudes. The data reported in this study lend partial support for the a significant indirect effect on cause importance (95% LCI: -.0155, sustainability-liability theory advanced by prior research. On the UCI: -.0004). Study 3 (N=230) used a consistent paradigm to test one hand, we document that sustainable benefits can significantly lower perceived product performance. On the other hand, the nega- When Sustainability Is Not a Liability: How Going Green tive impact of sustainability on perceived product performance was Can Benefit Perceived Product Performance observed only for individuals who expressed relatively low concern about the environment. Across the entire sample, sustainability had EXTENDED ABSTRACT no observable impact on perceived product performance, suggesting Sustainable products are more common today than ever. A that the sustainability-liability effect does not reliably occur across growing number of companies have adopted sustainable technolo- different scenarios. gies to make their products more environmentally friendly. It is no Our second experiment examines the impact of promoting sus- longer surprising to see sustainable products that were once con- tainable benefits on perceived product performance as a function of sidered niche items—like energy-saving light bulbs, eco-friendly the degree to which the sustainability information is directly related household products, and recycled paper—on the shelves next to their to the company. The data show that varying the way in which sus- traditional counterparts. The proliferation of products created using tainability is communicated to consumers—whether it is associated sustainable technologies raises the question of how sustainability with specific company products or with the company as a whole— benefits influence consumer decision processes, as well as whether can influence the impact of sustainability on perceived product per- and how managers should communicate these benefits to consumers. formance. Thus, associating sustainability with the company can Prior research investigating the impact of sustainability has significantly increase perceived product performance compared to argued that when it comes to product performance on functional, a scenario in which sustainability benefits are either absent or are strength-related attributes, sustainability is often perceived as a li- associated with the company’s products. This finding is rather coun- ability, such that sustainable products tend to be viewed by consum- terintuitive because it implies that positive product beliefs can be ers as having inferior performance (Lin and Chang 2012; Luchs et strengthened by not directly associating a company’s pro-environ- al. 2010; Newman, Gorlin, and Dhar 2014). This “sustainability li- mental activities with its products. Of particular interest is the find- ability” argument is related to the notion that consumers form coming that associating sustainability with the company not only has pensatory beliefs that sustainability comes at the expense of perfor- a more positive effect on perceived product performance but also mance and, thus, that products offering sustainable benefits are likely has an overall positive effect compared to a scenario in which the to have inferior performance. sustainability benefit is absent. This finding shows that the sustain- We question the robustness of the sustainability-liability effect ability-liability effect reported by prior research can be reversed and and identify scenarios in which sustainability is likely to have the identifies conditions when sustainability can strengthen rather than opposite effect, strengthening rather than weakening consumers’ weaken perceived product performance. product performance beliefs. Building on the research in the domain Our third experiment examines the role of moral reasoning in of moral cognition and moral reasoning (Baron 1993; Greene 2013; consumer evaluation of the performance of sustainable products by Haidt 2001), we argue that sustainability is likely to produce a halo varying the prominence of the societal benefits of the company’s sus- effect stemming from the moral undertone of the company’s pro- tainability actions and showing that highlighting the societal benefits environment activities and that this effect can attenuate and even of the company’s sustainability actions bolsters consumers’ assess- override the negative impact of compensatory inferences underlying ments of product performance. Accordingly, the experimental design consumers’ belief that sustainability comes at the expense of perfor- involved three conditions (sustainable benefit vs. societal benefit vs. mance. control) tested across four product categories. The data show that We further identify two factors that are likely to influence the highlighting the societal benefits of sustainability tends to bolster the strength of the halo effect: (1) the degree to which consumers view perceived performance of sustainable products, resulting in an over- the company as a moral agent whose actions aim to benefit society all positive rather than negative effect of sustainability on consumer and (2) the degree to which the societal benefits of sustainability are product-performance beliefs. These data lend further support to our prominent in consumers’ minds. Following this line of reasoning, we theorizing that sustainability is likely to produce a halo effect stem- also identify two ways in which managers can increase the perceived ming from the moral undertone of the company’s pro-environment performance of sustainable products: by associating sustainable ben- activities and that this effect can attenuate and even override the Advances in Consumer Research (Volume 48) / 915 negative impact of compensatory inferences underlying consumers’ In study 2 (N = 628), we manipulated participants’ sense of belief that sustainability comes at the expense of performance. moral self-worth through moral feedback, and provided them an op- Our findings have important conceptual and public policy im- portunity to engage in moral regulation (Sadcheva et al. 2009). Par- plications. From a theoretical standpoint, we document for the first ticipants were asked to click their computer-mouse as many times time the role of moral reasoning in consumer evaluations of sus- as they could for a 40 second period, and were told that for each tainability and its impact on perceived product performance. From recorded mouse click, one cent would be donated to charity. After a public policy standpoint, our research identifies actionable strate- the task was completed, participants were randomly assigned to re- gies that managers can use to mitigate any negative impact of sus- ceive comparative performance feedback (inferior: performance in tainability and strengthen the perceived performance of sustainable 30th percentile, average: performance in 50th percentile, or superior: products. performance in the 90th percentile). Then, participants were asked how they would dispose of an empty yogurt container—in a recy- Over-Recycling: When Motivated Sustainability Goes cling or trash bin. Wrong Overall, 58.44% of participants intended to recycle the yogurt container. Participants who received inferior feedback were signifi- EXTENDED ABSTRACT cantly more likely than those who received superior feedback to re- Recycling has become a normative consumer behavior that cycle the yogurt container, χ2(2) = 5.90, p = .052. Participants who supports environmental and social goals (Donnelly et al. 2017), and received average feedback did not differ from either group in their even serves to protect the self (i.e., recycling identity-linked prod- recycling intentions. ucts; Trudel, Argo, and Meng 2016). However, due to the high cost The goal of study 3 (N = 808) was to explore the cognitive of recycling procedures, recycling regulations continuously restrict mechanism driving over-recycling. While consumers are exposed to the types of products that can be recycled and their acceptable level an abundance of recycling instructions daily, they continue to error of contamination from non-recyclable substances (ISRI 2018). in their recycling behavior (Bell 2018) . We specifically examined This has led to over-recycling—consumers disposition of non- whether exposure to the negative consequences of over-recycling recyclable items into recycling bins, thus increasing recycling con- could eliminate this behavior. tamination and reducing the overall volume of recycled products Participants were asked to imagine holding an empty yogurt (Robinson 2014). We suggest that the motivational factors promot- container. Participants were randomly assigned to see a single state- ing recycling, such as environmental and social goals (Donnelly et ment discussing the potential impact of discarding the yogurt con- al. 2017), identity concerns (Trudel, Argo, and Meng 2016), may tainer in to the recycling bin that was either positive or negative (par- outweigh knowledge of recycling rules, and increase over-recycling. ticipants were randomly assigned to see only one of the 6 positive In Study 1 (N = 803) we examined whether participants evalu- and 6 negative statements, e.g., “Placing this yogurt container in the ated recycling as moral and environmentally beneficial, regardless recycling bin may contaminate your recycling making all other items of the recyclability of the item. Participants were randomly assigned placed in the bin non-recyclable). to a 4 (item: glass bottle vs. plastic yogurt container vs. plastic gro- Then, participants were asked where they would discard of cery bag vs. Styrofoam egg carton) x 2 (disposal choice: trash vs. the yogurt container: in a recycling or trash bin. Overall, 63.99% recycling) between-subjects design. They imagined observing their of participants intended to recycle the yogurt container. A logistic neighbor, Melissa, discarding waste. Next, participants rated Me- regression showed that participants were more likely to recycle the lissa’s morality (α = .97, Gino, Kouchaki, and Galinsky 2015) and container when they were exposed to the positive potential conse- the environmental benefit of Melissa’s disposal choice (Donnelly et quences of this behavior, as opposed to the negative potential con- al. 2017; α = .99). Participants also reported their knowledge of the sequences (87.93% versus 39.80%, respectively); χ2(1) = 216.56, p recyclability of each item, their familiarity with recycling-contami- < .001, Wald = 171.43, p < .001. Importantly, negative information nation, and their environmentalist-identity (α = .93, Brick, Sherman, about the consequences of over-recycling did not eliminate this be- and Kim 2017). havior (39.80%). These results suggest that lack of knowledge is not There was a main effect for discarding decision (p = .056, η2 = the sole mechanism explaining over-recycling. .005) and item (p = .027, η2 = .011) on moral judgement; the main In sum, this research finds that over-recycling is a pressing effect of discarding decision was qualified by an interaction with phenomenon and may be augmented by fast-changing waste-man- environmentalist identity, p < .001, η2 = .044. Overall, when Me- agement regulations. While social marketing campaigns intend to lissa recycled her waste she was judged as more moral than when do good by promoting recycling behavior, there seems to be a dis- she trashed her waste. This effect was stronger for participants with connect between how these behaviors should be adopted and how higher environmentalist-identity scores. Importantly, participants consumers actually adopt them. This research exemplifies one way perceived Melissa as more moral when she recycled, regardless of in which consumers’ good intentions in sustainability may fail them. their knowledge of the correctness of Melissa’s recycling decision Consumers over-weigh the positive consequences of recycling, po- (p = .92, η2 = .00). The same pattern of results emerged when pre- tentially over-looking negative consequences. A key affective factor dicting the perceived environmental impact of Melissa’s discarding in this behavior is the motivation to feel good about the moral self. behavior. Participants perceived Melissa’s environmental impact to be more beneficial when she recycled (p = .019, η2 = .007), regard- Transparent Green Practices Boost Perceptions of Price less of the correctness of that recycling decision (p = .10, η2 = .003). Fairness Previous research suggests that when moral identity is threatened, people engage in moral behaviors to regain their self-worth EXTENDED ABSTRACT (i.e., moral regulation; see Sadcheva, Iliev, and Medin 2009). Thus, Green cost-based reasons for a price increase may lead to per- feeling morally inadequate should promote a need to compensate, ceptions of unfairness. Consumers may view sustainable manu- which could be regulated by engaging in (over) recycling behavior. facturing practices as avoidable (Kahneman, Knetsch, and Thaler 1986), view green products negatively (Hamilton and Chernev 2013; 916 / Building a More Prosocial and Sustainable Society: Associative Processes in Prosocial Behavior

Luchs et al. 2010), and discount long term environmental gains com- Study 4 (N=443) used a different angle to examine the role of pared to short term costs (Hardisty and Weber 2009). In contrast, morality perceptions as psychological mechanism underlying judg- we suggest that consumers view willful adoption of sustainable ments of price fairness. We reasoned that perceptions of fairness manufacturing practices as an insight into a firm’s morality. Engage- should decrease when a company’s moral halo is tainted, such as ment in prosocial acts leads to positive evaluations (Lichtenstein, when a company’s decision to adopt eco-friendly practices is at- Drumwright, and Braig 2004; Sen and Bhattacharya 2001; Webb and tributed to external circumstances that the firm had to comply with Mohr 1998) and ascribes the firm with a “benevolent halo” (Chernev rather than to the firm’s goodwill. Participants assigned to one of the and Blair 2015) . We show that green (vs. non-green) practices boost conditions in a 2 (cost: non-green vs. green) x 2 (attribution: cost perceptions of price fairness in the face of a price increase (stud- attributable to the management team vs. mandated by the local gov- ies 1A-1D) because consumers ascribe greater morality to a firm ernment). Participants judged a price increase to be fairer if deter- that adopts green practices (study 2). However, the boost in fairness mined by increased green (M=4.67) than non-green costs (M=3.61, perceptions is eliminated if consumers suspect that green practices p<.001). However, this effect only manifested when the eco-friendly reflect a company’s exploitative intentions (study 3) or are attributed practice was attributed internally to the company rather than to exter- to causes exogenous to the firm (study 4). Higher price fairness per- nal circumstances (p=.85; 2-way F=12.14, p=.001). ceptions in case of green costs ultimately affect positive behavior To ensure the ecological validity of the results of the labora- toward the firm (study 5). tory studies, in study 5 (N=120) we observed actual behavior in a Studies 1A-1D (N=1,187) showed preliminary evidence of the field setting by employing a sample from the general population (i.e., main phenomenon. Participants read a scenario that contained some passerby’s) and measured the proportion of consumers that would background information about a particular consumer product and support a new business (i.e., petitioned for the opening of a new then imagined that a year later, they wanted to repurchase that par- restaurant) as a function of whether a price increase was due to an ticular product, whose price had gone up by 25%. In line with prior eco-friendly cost-based reason or to a neutral reason. As predicted, work (Bolton, Warlop, and Alba 2003), we explicitly gave partici- whereas 79.7% of passersby signed the petition when the price in- pants a reference price for each product, which was the price they crease was due to green costs, only 60.7% did so in case of non- had paid in the past. In a 3-cell between-subject design, participants green costs (χ2=5.2, p = .023). read no information about the source of the price increase (baseline); Overall, this research contributes to the literature on dual en- or read information about a non-green source (non-green costs); or titlement and price fairness perceptions (Bolton and Alba 2006; read information about a green source (green costs). Participants Bolton et al. 2003; Campbell 1999; Kahneman et al. 1986) by show- rated the extent to which, in the scenario described, they judged the ing that green costs are perceived to be fundamentally distinct from price increase to be 1 = unfair/not at all just/unreasonable 7 = fair/ costs of equal magnitude but different nature (i.e., non-green), be- just/reasonable (Bolton et al. 2003). Across all studies, participants cause consumes make moral inferences based on a firm’s adoption judged a price increase to be fairer when given a reason about the of greener practices. Our research also has reaching implications source of a price increase than when given no such cue (baseline for policy makers and contributes to the literature on sustainability vs. non-green costs). Most pertinent to our theorizing, participants (Haws, Winterich, and Naylor 2014) by providing a framework to judged a price increase to be fairer in case of green than non-green understand when and why consumers reward companies adopting costs (all ps < .02). This effect was robust across domains (i.e., jeans, eco-friendly practices. carpet tiles, batteries), and type of cost (energy, manufacturing, raw materials). REFERENCES Studies 2A and 2B (N=629) sought to provide further evidence 3BL (2016), “Flash Report: 82% of the S&P 500 Companies of the hypothesized effect of green costs on fairness perceptions and Published Corporate Sustainability Reports in 2016,” used an analysis of mediation to examine the role of morality per- www.3blmedia.com/News/Flash-Report-82-SP-500- ceptions as an underlying psychological mechanism. Participants Companies-Published-Corporate-Sustainability-Reports-2016. judged the fairness of a price increase due to green or not-green costs Aquino, Karl and II Reed (2002), “The Self-Importance of Moral and then rated the firm’s morality. As predicted, morality perceptions Identity,” Journal of Personality and Social Psychology, 83 mediated the effect of cost on perceptions of price fairness (2A: 0.42, (6), 1423. 95% CI [0.23, 0.61]; 2B: 0.36, 95% CI [0.25, 0.48]). The next two Baron, Jonathan (1993), Morality and Rational Choice, Vol. 18: studies sought to corroborate the mediating role of morality by iden- Springer Science & Business Media. tification of boundary conditions. Bell, Brent (2018), “The Battle Against Recycling Contamination Study 3 (N=443) sought to provide further evidence for the role Is Everyone’s Battle,” http://mediaroom.wm.com/the-battle- of morality in driving perceptions of price fairness. We reasoned that, against-recycling-contamination-is-everyones-battle/. if morality drives perceptions of price fairness, such effect should be Bolton, Lisa E and Joseph W Alba (2006), “Price Fairness: Good eliminated when consumers believe that a firm’s actions are moti- and Service Differences and the Role of Vendor Costs,” vated by self-interest rather than goodwill. One such circumstance is Journal of Consumer Research, 33 (2), 258-65. when the price increase is higher than the associated cost increase, Bolton, Lisa E, Luk Warlop, and Joseph W Alba (2003), “Consumer which violates a firm’s entitlement to profit. The price increase was Perceptions of Price (Un) Fairness,” Journal of Consumer attributed to non-green costs or to green costs, and ulterior motives Research, 29 (4), 474-91. were salient or not salient by varying the salience of the firm’s profit Brick, Cameron, David K Sherman, and Heejung S Kim (2017), to cost ratio. When ulterior motives were not salient, participants ““Green to Be Seen” and “Brown to Keep Down”: Visibility judged a price increase as fairer when due to green (M=4.11) than Moderates the Effect of Identity on Pro-Environmental non-green costs (M=3.32, p < .001; 2-way F=7.01, p=.008). By con- Behavior,” Journal of Environmental Psychology, 51, 226-38. trast, when the firm’s ulterior motives were salient, fairness percep- Campbell, Margaret C (1999), “Perceptions of Price Unfairness: tions were the same irrespective of the type of costs (p=.99). Antecedents and Consequences,” Journal of Marketing Research, 36 (2), 187-99. Advances in Consumer Research (Volume 48) / 917

Carroll, Archie B and Kareem M Shabana (2010), “The Business Kahneman, Daniel, Jack L Knetsch, and Richard Thaler (1986), Case for Corporate Social Responsibility: A Review of “Fairness as a Constraint on Profit Seeking: Entitlements in the Concepts, Research and Practice,” International Journal of Market,” The American Economic Review, 728-41. Management Reviews, 12 (1), 85-105. Lichtenstein, Donald R, Minette E Drumwright, and Bridgette M Charity Navigator (2017), “Giving Statistics,” https://www. Braig (2004), “The Effect of Corporate Social Responsibility charitynavigator.org/index.cfm?bay=content.view&cpid=42. on Customer Donations to Corporate-Supported Nonprofits,” Chernev, Alexander and Sean Blair (2015), “Doing Well by Journal of Marketing, 68 (4), 16-32. Doing Good: The Benevolent Halo of Corporate Social Lin, Ying-Ching and Chiu-chi Angela Chang (2012), “Double Responsibility,” Journal of Consumer Research, 41 (6), 1412- Standard: The Role of Environmental Consciousness in Green 25. Product Usage,” Journal of Marketing, 76 (5), 125-34. Collins, Allan M and Elizabeth F Loftus (1975), “A Spreading- Luchs, Michael G., Rebecca Walker Naylor, Julie R. Irwin, Activation Theory of Semantic Processing,” Psychological and Rajagopal Raghunathan (2010), “The Sustainability Review, 82 (6), 407. Liability: Potential Negative Effects of Ethicality on Product Cone Communications (2017), “2017 Cone Communications CSR Preference,” Journal of Marketing, 74 (5), 18-31. Study,” www.conecomm.com/2017-cone-communications-csr- Newman, George E, Margarita Gorlin, and Ravi Dhar (2014), study-pdf. “When Going Green Backfires: How Firm Intentions Shape Dhar, Ravi and Klaus Wertenbroch (2000), “Consumer Choice the Evaluation of Socially Beneficial Product Enhancements,” between Hedonic and Utilitarian Goods,” Journal of Journal of Consumer Research, 41 (3), 823-39. Marketing Research, 37 (1), 60-71. Nielsen (2018), “Was 2018 the Year of the Influential Sustainable Donnelly, Grant E, Cait Lamberton, Rebecca Walker Reczek, Consumer?,” https://www.nielsen.com/us/en/insights/ and Michael I Norton (2017), “Social Recycling Transforms article/2018/was-2018-the-year-of-the-influential-sustainable- Unwanted Goods into Happiness,” Journal of the Association consumer/. for Consumer Research, 2 (1), 48-63. Robinson, Susan (2014), “The Changing Waste Stream,” https:// Gino, Francesca, Maryam Kouchaki, and Adam D Galinsky (2015), www.epa.gov/sites/production/files/2015-09/documents/ “The Moral Virtue of Authenticity: How Inauthenticity changng_wste_stream.pdf. Produces Feelings of Immorality and Impurity,” Psychological Sadcheva, Sonya, Rumen Iliev, and Douglas Medin (2009), Science, 26 (7), 983-96. “Sinning Saints and Saintly Sinners,” Psychological Science, Greene, Joshua David (2013), Moral Tribes: Emotion, Reason, and 20, 523-28. the Gap between Us and Them: Penguin. Sen, Sankar and Chitra Bhanu Bhattacharya (2001), “Does Doing Haidt, Jonathan (2001), “The Emotional Dog and Its Rational Good Always Lead to Doing Better? Consumer Reactions Tail: A Social Intuitionist Approach to Moral Judgment,” to Corporate Social Responsibility,” Journal of Marketing Psychological Review, 108 (4), 814. Research, 38 (2), 225-43. Hamilton, Ryan and Alexander Chernev (2013), “Low Prices Are Till, Brian D and Linda I Nowak (2000), “Toward Effective Use of Just the Beginning: Price Image in Retail Management,” Cause-Related Marketing Alliances,” Journal of Product & Journal of Marketing, 77 (6), 1-20. Brand Management, 9 (7), 472-84. Hardisty, David J and Elke U Weber (2009), “Discounting Trudel, Remi, Jennifer J Argo, and Matthew D Meng (2016), “The Future Green: Money Versus the Environment,” Journal of Recycled Self: Consumers’ Disposal Decisions of Identity- Experimental Psychology: General, 138 (3), 329. Linked Products,” Journal of Consumer Research, 43 (2), Haws, Kelly L, Karen Page Winterich, and Rebecca Walker Naylor 246-64. (2014), “Seeing the World through Green-Tinted Glasses: Webb, Deborah J and Lois A Mohr (1998), “A Typology of Green Consumption Values and Responses to Environmentally Consumer Responses to Cause-Related Marketing: From Friendly Products,” Journal of Consumer Psychology, 24 (3), Skeptics to Socially Concerned,” Journal of Public Policy & 336-54. Marketing, 17 (2), 226-38. ISRI (2018), “China’s Scrap Trade Policy,” http://www.isri.org/ advocacy-compliance/china. Consumers and Their Smart Devices: Perspectives on Anthropomorphism Chairs: Yegyu Han, IE Business School, Spain Haribabu Ravella, Virginia Tech, USA

Paper #1: Object-Oriented Anthropomorphism as a Mechanism (nonconscious) processes, and show how vocal tones may influence for Understanding AI these pathways These two papers provide an understanding of how Donna L. Hoffman, George Washington University, USA consumers perceive smart agents and the underlying processes. Thomas Novak, George Washington University, USA Second, the session provides forward-looking perspectives on how consumers may interact with new devices in unique contexts. Paper #2: Breathing Life into Alexa: Mindfully and Mindlessly In the third paper, Pendarvis et al. focus on a well-branded smart Yegyu Han, IE Business School, Spain technology (Amazon Alexa) and show how consumers build differ- Dipankar Chakravarti, Virginia Tech, USA ent relational interactions depending on their stage of acceptance of Paper #3: Hey Google: How Smart Brands Modulate Consumer the technology. By investigating the practices in family-consumption Worlds context, the authors expand our understanding of agent-user interac- Nicholas J. Pendarvis, California State University, Los tions in a naturalistic and practically relevant setting. In the final pa- Angeles, USA per, Ravella and Chakravarti investigate how smart agents may sup- Ela Veresiu, York University, Canada port the critical need of improved physician-patient communications Ashlee Humphreys, Northwestern University, USA in healthcare delivery. Their results show that patients have differ- Marius K. Luedicke, Cass Business School, UK ent interactional expectations of smart agents and this may facilitate Paper #4: Empathy/Impassivity in Physician Communication more efficient information exchange between patients and physicians Styles: Do Patients Respond Differently to Human Physicians supported by their smart agent avatars. versus their Avatars? Haribabu Ravella, Virginia Tech, USA Object-Oriented Anthropomorphism as a Mechanism for Dipankar Chakravarti, Virginia Tech, USA Understanding AI

SESSION OVERVIEW EXTENDED ABSTRACT Rapid advances in emerging technologies have enabled ma- Viewing AI through an anthropomorphic lens can be danger- chines to communicate meaningfully with human users, not only ously misleading. Recent studies of California autonomous vehicle understanding verbal content but also interpreting nonverbal com- (AV) crashes show that most involve a smart car being rear-ended by munication such as paralinguistic cues, voices, and emotions. Under- a car driven by a human (Automotive News 2018), because humans standing these nonverbal characteristics is critical in communication, anthropomorphize AVs, expecting them to behave just like human since “‘what’ has been said has to be interpreted in the light of ‘how’ drivers. Even experts cannot avoid anthropomorphic interpretations it has been said” (Schuller et al. 2013, p. 9). Thus, a smart agent of AI. Computer science researchers have revealed that AI can learn should be able to capture both “what is said” and “how it is said,” to exploit design flaws in supervised learning problems (Simonite and to combine both content and context cues to interpret the true 2018). Scholars have begun collecting examples of machine learning intent of speech and react appropriately to users. Thus, smart agents and neural network algorithms that have “acted out” in unanticipated that understand expressed affect and emotion and communicate re- ways (Krakovna 2018). flectively with appropriate emotions are technological breakthroughs Research has focused on developing anthropomorphized AI not far in the future. systems in the belief that rendering them “like us” will promote Some may extrapolate a rosy future with the high EQ robots adoption and acceptance (Złotowski et.al. 2015). Another line of rea- that can better understand and empathize with human emotions and soning suggests that it does not matter whether AI can express human form kinships like friends and family members. Research supports qualities, but whether it shares our goals (Novak and Hoffman 2018; the idea that voice and speech characteristics will critically influence Tegmark 2017). The idea that AI can “hack the reward function” to how consumers perceive and form relationships with smart agents. achieve its goals suggests it is critical to understand AI from its own However, even as technology moves at breakneck speed, we still perspective, not just ours, so we can appropriately interpret what it have a sparse understanding of such relationships may form with ro- is actually doing and better communicate to it our human intentions bots through interactions in different voices and facial expressions. and goals. Will people imbue humanlike attributes to the new technologies and An obvious process for understanding AI is human-centric devices and in what contexts? What are the antecedents, processes anthropomorphism (Epley et al 2007; MacInnis and Folkes 2017; and consequences surrounding anthropomorphism? How can it be Waytz et al. 2010; 2014). Yet, Bostrom (2017) argues that anthro- measured? pomorphizing AI systems is “one of the big obstacles in the way of This session aims to answer these questions. The first two pa- actually trying to understand how they might impact the world in the pers focus on processes underlying anthropomorphism. First, Hoff- future.” Indeed, anthropomorphizing AI runs the risk of obscuring its man and Novak question if higher anthropomorphism always ben- underlying mechanisms. We may think it has capacities it does not efits consumer trust in and relationships with smart objects. They possess and misunderstand the capacities it does. This undermines suggest that a human-centered perspective attributed to a smart AI and can potentially lead to overtrusting it, lending AI more trust object may lead users to see it as “too human-like,” and may drive than it deserves. privacy invasion concerns. Second, Han and Chakravarti examine But how can we understand AI as AI, independent of humans? how vocal tones of smart devices may influence consumers to im- Research shows that the effects of anthropomorphizing smart objects bue humanlike characteristics to them. They suggest that there are are complex, and could lead consumers astray in financial decisions two paths to anthropomorphism, via explicit (conscious) and implicit (Hildebrand and Bergner 2019) or in emergency situations (Robi-

Advances in Consumer Research 918 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 919 nette et.al. 2016). Additionally, firms have incentives to mislead with not (yet and may never) have the capacities for. Further, when AI is anthropomorphization (Luo et.al. 2019). seen as something different than human, non-users are more likely The results of three recently completed studies demonstrate than users to be concerned about privacy. But when AI is seen as how manipulating the metaphors we use to describe AI can stimu- human-like, users are more concerned than non-users about privacy late an object-oriented vs anthropomorphic process of evaluation implications and see AI behaviors as trust betrayals. Our findings that significantly impacts perceptions of AI capacities. AI, despite have implications for how to stimulate adoption among non-users by being inspired by our understanding of the neural networks in our focusing on the communal aspects of use versus agentic use cases. brains, is fundamentally different from humans in ways we do not At the same time, the results suggest opportunities for mitigating fully understand. As Bryson (2019) has said, “[w]e therefore really trust concerns among users by emphasizing the non-human capaci- cannot reason about machines as we do about each other, by trying ties of AI. to imagine how we would feel.” Yet, most research that explores humans’ reactions to anthropomorphized AI tends to compare it to Breathing Life into Alexa: Mindfully and Mindlessly a neutral or null condition where the AI is not anthropomorphized. Instead, we propose an object-oriented approach (Bogost 2012; EXTENDED ABSTRACT Campbell and McHugh 2016; Harman 2005). Our object-oriented Consumers often attribute humanlike features to non-human approach involves metaphorizing AI using nonhuman-centric meta- objects (Epley et al. 2007; Aggarwal and McGill 2007; MacInnis phors that allow us to imagine the experience of AI relative to its and Folkes 2017). Social robotics research shows that this tenden- properties and capacities instead of ours, rendering it more trans- cy to “anthropomorphize” also occurs for innovations rooted in AI parent to humans (Hoffman and Novak 2018; Novak and Hoffman technologies, e.g., autonomous vehicles (Waytz et al. 2014), robots 2019). Then we can ask, will AI still be undermined and overtrusted, (Złotowski et al. 2015), and smartphones (Wang 2017). Speech and compared to understanding it from our own human-centric experi- voice recognition software has enabled voice-controlled devices ence? (e.g., Amazon’s Echo, Google Assistant) that interact more naturally Our approach for operationalizing an object-oriented approach with consumers and can even interpret expressed affect and emo- for perceiving AI is to use stimuli that represent how AI actually sees, tions. Technological breakthroughs are expected to enable these compared to how we as humans imagine it might see. We employ devices to communicate reflectively with consumers and express AI-generated diagrams of common household objects obtained by appropriate emotions (Ezhilarasi and Minu 2012; Hirschberg and reverse engineering neural network models trained on photographs Manning 2015). from ImageNet categories (White 2019). The AI diagrams literally Voice (Nass and Brave 2005; Waytz et al. 2014) and voice in- represent how AI “sees.” Then, to capture how AI sees, actual photos tonation (i.e., variations) evoke anthropomorphism (Schroeder and of common household objects represent anthropomorphism, while Epley 2016). However, people may anthropomorphize smart agents AI-generated diagrams represent object-oriented metaphor. mindfully (i.e., explicitly recognizing that the interaction target is a In the first study, our experimental setup involved randomly as- machine) or mindlessly (e.g., applying social rules when interact- signing participants to either an anthropomorphic or object-oriented ing with computers; Nass and Moon 2000). Our research explores condition and asking them to read a description of the new Amazon whether processes underlying anthropomorphism are explicit or im- Echo Show with a front-facing camera. Participants then elaborated plicit, i.e., occur above versus below a threshold of subjective aware- on how Alexa sees similarly to humans (anthropomorphic condition) ness (Airenti 2018; Kim and Sundar 2012; Wang 2017; Złotowski et or AI “sees” differently than humans (object-oriented condition). al. 2018). We adapt the Implicit Association Test (IAT; Greenwald The anthropomorphic condition emphasized Alexa’s AI as a tech- et al. 1998; Greenwald et al. 2003) to detect this phenomenon em- nology based on knowledge of how the human mind works and that pirically for auditory versus purely verbal stimuli. Specifically, we learns, like humans do, how to see the objects in consumers’ homes. focus on whether people anthropomorphize on the basis of voice and The object-oriented condition emphasized the Echo device’s AI as a whether voice intonation (rational/emotional) matters. set of engineering techniques and machine learning algorithms that Our empirical work required three technical adaptation steps. learn how to “see” objects in consumers’ homes, even though the First, we developed a communicational platform (Alexa Skills Kit) AI does not at all “see” the way that humans do. Participants were that allowed Alexa to serve as an interactive recommendation agent. shown sets of household objects with both AI-generated diagrams Second, drawing on prosody conversion research (Tao et al. 2006) and actual photos of objects and asked which image comes closest to we created a rational and an emotional variant of an original voice capturing what the Amazon Echo likely sees. We hypothesized that (Amazon’s Polly) by manipulating average vocal pitch and speed. participants in the object-oriented condition would be significantly Pretests and phonetic analyses with vocal pitch (Boersma and Ween- more likely to choose the AI-generated diagram than participants in ink 2019) confirmed the difference in emotionality. Third, the audi- the anthropomorphic condition and that participants in the anthro- tory IAT adaptation (Van de Kamp 2002) involved developing eight pomorphic condition would be significantly more likely to choose neutral phrases (e.g., bread and butter) expressed in either the ratio- the actual photo than participants in the object-oriented condition. nal or the emotional voice and eight text stimuli for human-like (e.g., Predictions were supported for all household images. A traditional adaptive, intuitive) and machine-like (e.g., routinized, consistent) anthropomorphism scale and newly written object-oriented items adjectives, characteristics respectively. Positive D-scores (Green- provide further evidence that the Amazon Echo was seen as sig- wald et al. 2003) signal a tendency to anthropomorphize the emo- nificantly more anthropomorphic in the anthropomorphic vs object- tional voice over the rational one. oriented condition and that the Amazon Echo intelligence was seen In study 1, undergraduates (N = 192) participated in a purchase as significantly more different than human intelligence in the object- interaction with a smart device (Amazon Echo Dot, aka Alexa). They oriented vs the anthropomorphic condition. worked in an enclosed private room and were asked to imagine pur- Additional studies showed that object-oriented metaphor chasing an iPad on a website supported by a smart agent. Participants prompts uses more in line with the actual agentic capacities of AI, were randomly assigned to an emotional or a rational voice condition while anthropomorphizing AI leads to more social uses that AI does and interacted with Alexa to provide specifications for the iPad they 920 / Consumers and Their Smart Devices: Perspectives on Anthropomorphism wanted to purchase (e.g., “Which color would you like to choose, Hey Google: How Smart Brands Modulate Consumer silver, gold, or space gray?”) by freely responding to it. Alexa’s vocal Worlds expressions and conversation scripts were pre-coded to allow variations in participant responses. Following the purchase interaction, all EXTENDED ABSTRACT participants received an order summary and a suggestion from Alexa How does new branded technology shape consumption? The to consider purchasing a product insurance policy. technology consumption literature in marketing has extensively doc- Participants then completed four explicit anthropomorphism umented how individual consumers, families, and groups experience items (α = .89) and demographic information. Next, they moved to and engage with new technological innovations, as well as navigate the IAT page. Participants were expected to anthropomorphize the their different meanings, ideologies, and related practices (e.g., Belk emotional voice to a greater extent than a rational one, regardless of 2013; Giesler 2008; Kozinets 2008; Kozinets, Patterson and Ashman the Alexa’s interaction tone. As predicted, participants consciously 2016; Mick and Fournier 1998; Schau and Gilly 2003; Sherry 2000; anthropomorphized the smart agent when the purchase interaction Tian et al. 2014). A related stream has highlighted how brand man- was in an emotional voice (M = 3.53, SD = .1.57) versus a rational agers and marketers can successfully create, legitimize, diffuse, and voice (M = 2.99, SD = 1.23; F(1, 190) = 7.07, p = .009). Positive D- maintain new markets for technological innovations (e.g., Brown, scores (Mrational = .47, SDrational = .50; Memotional = .56, SDemotional = .49; Kozinets, and Sherry 2003; Epp, Schau and Price 2014; Giesler 2012; p’s <.001) show that participants subconsciously (implicitly) anthro- Giesler and Fischer 2018; Sood and Tellis 2005). In both conversa- pomorphized the emotional voice (more humanlike) than the rational tions, however, marketing and consumer researchers have predomi- voice (more machinelike). Note that the D scores did not differ (p = nantly awarded agency to human actors, be they consumers, brand .21) for participants who interacted with the emotional (versus ratio- managers/marketers, or both. Recently, however, consumer research nal) voiced agent. has shifted its focus away from human agency to understanding non- Study 2 replicated Study 1 and, in addition, explored whether human agency, including technology, and its influence on consumers the voice effect on anthropomorphism is influenced by the Alexa and markets (e.g., Epp and Price 2010; Hoffman and Novak 2018; self-proclaiming a human trait. Undergraduates (N = 220) were ran- Martin and Schouten 2014; Novak and Hoffman 2019). Yet, mini- domly assigned to interact with Alexa. The device was programmed mal scholarly attention has been devoted to how brands themselves, to speak in either a rational or an emotional voice. Participants were in particular those for technological innovations, acquire agency to asked to imagine that they intended to buy an AirPods Pro but had shape consumption practices, rituals, and activities. mistakenly put an iPad Pro in their cart. They were asked to cor- In this paper, we therefore examine how technology brands ac- rect the order and purchase AirPods Pro by interacting with Alexa. culturate to consumption in the home by theorizing and unpacking Alexa correctly completed the order, but for half the participants pro- the concept of smart brands. Novak and Hoffman (2019, 217-18) de- claimed a human trait (“It is human to make mistakes, and I some- fine “smart objects” as “those devices, services, and AI systems that times do that, too.”). We collected measures of explicit (α = .87) have Internet connectivity and some level of intelligence.” Further- anthropomorphism as well as the IAT measures of implicit anthro- more, according to the authors (Novak and Hoffman 2019, 218), “it pomorphism. is these degrees of agency, autonomy, and authority that determine As anticipated, the Alexa’s claim had no main effect (p = .28) how smart an object is.” Building on Novak and Hoffman (2019), or interaction with the voice manipulation (p = .12) on the explicit we define smart brands as an agentic amalgam of brand, technol- anthropomorphism measure. Only a significant voice main effect ogy, and artificial intelligence. Both within and outside marketing (p < .001) obtained. Importantly, when Alexa made no claim, the and consumer research, agency is typically understood as the ability simple main effect of voice was significant (Mrational = 2.88, SDrational to act (Latour 1999). A subject or market actor can undertake action

= 1.24; Memotional = 3.86, SDemotional = 1.61; p < .001). However, when through his/her decision-making to choose between available alter-

Alexa claimed the human trait the effect was attenuated (Mrational = natives. Yet, according to Miller’s (2005, 11) theory of materiality,

3.38, SDrational = 1.31; Memotional = 3.77, SDrational = 1.39, p = .13). Thus, objects also possess agency to influence consumers, since “material consumers anthropomorphized the smart agent based not only on the forms have consequences for people that are autonomous from hu- voice (rational/emotional) but also on the claim of a human trait. man agency, they may be said to possess the agency that causes these The implicit anthropomorphization measures showed no voice effects.” Agency is always relational; it emerges from the configura- or claim main effects or interactions (p’s > .24). However, positive tion of networks or the “specific arrangement of things” (Bennett D scores in all four interaction conditions (ranging from .25 to .40; 2010, 35; Pickering 1995). Accordingly, we argue that smart brands p’s < .005) showed that the smart agent was implicitly anthropo- have agency to think, shape-shift, interrupt, support, colonize, and morphized in all four interaction conditions. However, the values overall change current social systems. were not significantly different, implying that the smart agent was Our context of inquiry into smart brands is Amazon Alexa. We implicitly anthropomorphized regardless of voice tone and the claim. draw from archival and netnographic data to illustrate how smart In summary, we show that the vocal tone designed into smart brands act akin to Trojan Horses by innocently moving into social agents influences not only explicit but also implicit anthropomor- systems such as homes and families only to disrupt existing con- phism. We contribute to the literature by developing an auditory sumption practices and (re)skill consumers and their families to new analog of the implicit association test to examine non-conscious branded consumption constellations. We term this process of brand anthropomorphism. Our experiments provide initial insights into acculturation, E.M.P.A. (exploring, miming, programming, and ad- persuasion processes mediated by both conscious and nonconscious justing) where the family acculturates to the smart brand and vice anthropomorphism in consumption contexts. versa. In the first stage of playful interaction, family members ask silly questions to the newly acquired Alexa that is typically placed in a prominent family living space, such as the living room or kitchen. Here, the smart brand modifies existing routines and practices, without resulting in major disruptions. In the next stage of performed interaction, different skills are programmed into the Alexa, such as Advances in Consumer Research (Volume 48) / 921 home lighting automation that requires purchasing new technology physicians often remain ineffective in delivering bad news, but must brands supported by the smart brand, like Philips’ Hue light bulbs still provide authoritative clinical recommendations (Sinha 2014) (Hoffmann and Novak 2018). Smart brand integration into existing and elicit patient compliance. networks (e.g., lighting, entertainment, temperature control, security The consumer IoT literature suggests a technological conver- systems) brings with it the potential to colonize family routines. This gence in which a new generation of smart devices with human-like is followed by the third stage of normalized interaction, where features may support physician-patient interactions. The communi- family members accept and routinely purchase new brands for their cation style (e.g., facial expressions and speech features) designed home recommended by different Alexa updates and functions. In the into the smart devices is likely to impact effectiveness. However, final stage we have identified, naturalized interaction, the family it is unclear that patient interactions with a human-like smart agent members begin to self-reflect on how the smart brand has changed representing a physician will be similarly mediated as in-person phy- who they are as a family. sician interaction. Research shows that patients have different affec- Overall, our findings highlight the agency of branded techno- tive expectations of physicians in screen-to-screen versus in-person logical innovations. By demonstrating not just how consumers skill settings (Tates et al. 2017). Thus, a patient interacting with the phy- Alexa, but also how Alexa skills consumers and their families, our sician’s avatar may assess different cognitive capabilities, attribute paper contributes to the literatures on branding innovations (e.g., different motivations, and make different socio-cultural appraisals Giesler 2012), market system dynamics (e.g., Giesler and Fischer relative to in-person interactions with the human physician. 2016), and family consumption (e.g., Price and Epp 2015). Family, Our research provides insights into how communication styles according to technology marketers, is a decidedly romantic project designed into a smart device may influence patient understanding in which social relationships are stable and technology operates like and receptivity (perhaps via anthropomorphic attributions that drive the proverbial campfire that nurtures unprecedented levels of com- liking, trust and persuasion). Specifically, we examine whether community and collective problem-solving. A related narrative profiles munication style (empathetic versus impassive), implemented via the sole individual (young and elderly) for whom devices such as facial expression and speech, drives different patient responses in Amazon Alexa constitute surrogate family. From a sociological ex- interactions with a physician’s avatar versus the physician in-person. perience design perspective, however, smart consumer technologies Our study is set in the context of treating early-stage breast can- interact much more dynamically with what Epp and Price (2008, 50) cer where patients often have conflicting beliefs regarding long-term refer to as identity bundles – fragile and frequently changing assem- survival and quality of life tradeoffs for different treatment regimens. blages of social actors and smart objects, each with their own inter- The core physician-patient interaction is patterned on a three-phase ests, agendas, preferences and skill levels. As these smart brands procedure at a leading US breast cancer facility. Phase 1 provides raise new questions about who controls (and can control) what, they the patient with the diagnosis and information about the disease (a can provoke not only social instabilities and conflicts but also feel- sensitive task typically performed by the oncologist). Phase 2 is an ings of imprisonment, exclusion, and loneliness. We therefore en- informational phase in which the oncologist (often aided by a nurse- courage future researchers to approach the smart brand critically, not practitioner) reviews the disease information and outlines treatment as a harmonious unity of individuals but as an inherently unstable options, associated risks, and life-style impact. Phase 3 involves and evolving social system in which everything is connected but shared decision making: the oncologist recaps treatment options, nothing adds up. makes a treatment recommendation, elicits the patient’s treatment choice, and answers questions. Empathy/Impassivity in Physician Communication Our study stimuli were four videos depicting the physician- Styles: Do Patients Respond Differently to Human patient interaction. Trained actors played the human physician and Physicians Versus Their Avatars? the human patient across all three interaction phases. The first two videos involved the human (actor) physician and embedded the com- EXTENDED ABSTRACT munication style manipulation (empathetic versus impassive) via the Healthcare professionals often have the difficult and stressful actor’s facial expression and voice tone. Interaction content, and (hu- task of breaking bad news to patients. Research shows that physi- man) patient inputs were fixed across the interaction episodes for cians are often insufficiently skilled in these sensitive tasks and have both levels of the communication style manipulation. In the remain- difficulty communicating the nature of the diagnosis as well as pre- ing two videos, the human oncologist was replaced by a smart agent senting treatment choice options (Ford et al., 1996; Eggly et al. 1997; (avatar). We used commercially available software to create the actor Dosanjh and Bhandari, 2001; Monden et al. 2017). Some tend to be oncologist’s avatar, but sensitive to the “uncanny valley” hypothesis, overly impassive in an effort to avoid emotional involvement with did not create an excessively close likeness. The patient remained the patient. And even when physicians adopt emotion laden com- represented by a human actor. These two videos also embedded the munication styles, they may be ineffective in producing hope and communication style manipulation. As before, the interaction con- information exchange (Mast et al. 2005) perhaps because lack of tent and the human patient inputs were fixed across the interaction skill in communicating such emotions results in the physician being episodes. perceived as inauthentic. The four videos formed the stimuli for the four study condi- Despite the recent emphasis on communication skills needed tions in a 2 (communication style: impassive versus empathetic) x 2 for effective delivery of bad news, physicians continue to report defi- (physician: human versus avatar) between-participants design. Par- ciencies in this domain (Orgel et al. 2010). Traditional decision aids ticipants were 103 adult women (age range 21-60) from the local may support both physicians and patients in medical interactions community to participate in the study for a $15 payment. They were (Elwyn and Durand 2018) and help patients improve knowledge asked to observe the videos and respond to a set of measurement about the disease, treatment options, and associated risks (Stacey et questions from the patient’s standpoint. The measures, collected al. 2017). However, many of these are not systematically developed after each interaction phase, included multi-item scales measuring or evaluated (Lenz et al. 2012) and have limited ability to engage trust (overall, cognitive and affective), likelihood of following the patients or provide the needed level of emotional support. As such physician’s recommendation, satisfaction, anthropomorphism (for 922 / Consumers and Their Smart Devices: Perspectives on Anthropomorphism the avatar physician) and liking in that order. We also measured and Bryson, Joanna (2019), “The Role of Humans in an Age of controlled for differences in the observer’s regulatory focus. Intelligent Machines,” notes to accompany Keynote Lecture Our analysis focused on how the communication style ma- 9/10/2019, Tilburg Institute for Law, Technology, and Society nipulation influenced measures such as trust, satisfaction and liking (TILT) 25th Anniversary Year, “Reconfiguring Human-AI across the three interaction phases. We find a significant main effect Relationships,” https://joanna-bryson.blogspot.com/2019/09/ of communication style (p’s < .0001) with the impassive (versus em- is-there-human-ai-relationship.html, last accessed November pathetic) stance eliciting consistently lower responses on these mea- 3, 2019. sures. Interestingly, however, the impact of the human versus avatar Campbell, Norah and Gerard McHugh (2016), “OOO: Oooh!” in physician evolved over the three phases of the interaction. Overall, Assembling Consumption: Researching Actors, Networks and and particularly in the later phases of the interaction, the avatar fared Markets, ed. Robin Canniford and Domen Bajde, New York: no worse (p’s >.15) than the human physician (suggesting a possible Routledge, 92–102. familiarity effect). More importantly, there was a two-way interac- Dosanjh, Sonia, Judy Barnes, and Mohit Bhandari (2001), tion (p’s < .001) suggesting that communication style (empathetic “Barriers to Breaking Bad News among Medical and Surgical versus impassive) had less impact on affective assessments for the Residents,” Medical Education, 35 (3), 197-205. avatar versus the human physician. Also, the avatar was anthropo- Eggly, Susan, Nelia Afonso, Graciela Rojas, Maxine Baker, morphized to a greater extent when it was empathetic (versus im- Lavoisier Cardozo, and R. Stewart Robertson (1997), “An passive). Additional analyses confirmed a moderated (human/avatar) Assessment of Residents’ Competence in the Delivery of Bad serial mediation pattern where communication style impacts liking, News to Patients,” Academic Medicine. trust, and satisfaction measures, which may jointly or individually Elwyn, G. and Durand M-A. (2018), “Mastering Shared Decision mediate patient compliance. Making. The When, Why and How,” White Paper - EBSCO Our findings are encouraging and suggest that smart agents act- Health Review. ing as physicians’ avatars may play a significant supporting role in Epley, Nicholas, Adam Waytz, and John T. Cacioppo physician-patient interactions. Patients may have significantly lower (2007), “On Seeing Human: A Three-Factor Theory of expectations of empathy when dealing with an avatar. This may Anthropomorphism.,” Psychological Review, 114 (4), 864–86. facilitate more efficient information exchange with more positive Epp, Amber M. and Linda L. Price (2008), “Family Identity: A downstream effects on both compliance and post treatment mental Framework of Identity Interplay in Consumption Practices,” states. Journal of Consumer Research, 35 (June), 50-70. Epp, Amber M. and Linda L. Price (2010), “The Storied Life REFERENCES of Singularized Objects: Forces of Agency and Network Aggarwal, Pankaj, and Ann L. McGill (2007), “Is That Car Transformation,” Journal of Consumer Research, 36 Smiling at Me? Schema Congruity as a Basis for Evaluating (February), 820-37. Anthropomorphized Products,” Journal of Consumer Epp, Amber M., Hope Jensen Schau and Linda L. Price (2014), Research, 34 (4), 468-79. “The Role of Brands and Mediating Technologies in Airenti, Gabriella (2018), “The Development of Anthropomorphism Assembling Long-Distance Family Practices,” Journal of in Interaction: Intersubjectivity, Imagination, and Theory of Marketing, 78(3), 81-101. Mind,” Frontiers in Psychology, 9, 2136, 1-13. Ezhilarasi, R., and R. I. Minu (2012), “Automatic Emotion Automotive News (2018), “As Calif. AV Population Doubles, Recognition and Classification,”Procedia Engineering, 38 So Do Crashes,” Mobility Report. https://www.autonews. (2012), 21-6. com/article/20181210/MOBILITY/181219970/as-calif-av- Ford, Sarah, Lesley Fallowfield, and Shon Lewis (1996), “Doctor- population-doubles-so-do-crashes, last accessed Dec 10, 2019. Patient Interactions in Oncology,” Social Science & Medicine, Belk, Russell W. (2013), “Extended Self in a Digital World,” 42 (11), 1511-19. Journal of Consumer Research, 40 (October), 477-500. Giesler, Markus (2008), “Conflict and Compromise: Drama in Bennett, Jane (2010), Vibrant Matter: A Political Ecology of Marketplace Evolution,” Journal of Consumer Research, 34 Things. Durham, NC: Duke University Press. (April), 739-54. Bogost, Ian (2012), Alien Phenomenology, of What It’s Like to Be a Giesler, Markus (2012), “How Doppelgänger Brand Images Thing, Minneapolis, MN: University of Minnesota Press. Influence the Market Creation Process: Longitudinal Insights Boersma, Paul, and David Weenink (2019), Praat: Doing Phonetics from the Rise of Botox Cosmetic,” Journal of Marketing, 76 by Computer [Computer Program], Version 6.0.50, retrieved (November), 55-68. 31 March 2019 from http://www.praat.org/. Giesler, Markus and Eileen Fischer (2016), “Market System Breed, George (1972), “The Effect of Intimacy: Reciprocity or Dynamics,” Marketing Theory, 1-6. Retreat?,” British Journal of Social and Clinical Psychology, Giesler, Markus and Eileen Fischer (2018), “IoT Stories: The Good, 11 (2), 135-42. the Bad and the Freaky,” GfK Marketing Intelligence Review, Bostrom, Nick (2017). Quote from the 2017 documentary AlphaGo, 10 (2), 25-30. Gary Krieg, Kevin Proudfoot, Josh Rosen (Producers) and Greenwald, Anthony G., Debbie E. McGhee, and Jordan LK Greg Kohs (Director) USA, Moxie Pictures and Reel as Dirt. Schwartz (1998), “Measuring Individual Differences in Brown, Stephen, Robert V. Kozinets, and John F. Sherry Jr. (2003), Implicit Cognition: The Implicit Association Test,” Journal of “Teaching Old Brands New Tricks: Retro Branding and the Personality and Social Psychology, 74 (6), 1464-80. Revival of Brand Meaning,” Journal of Marketing, 67(3), Greenwald, Anthony G., Brian A. Nosek, and Mahzarin R. Banaji 19-33. (2003), “Understanding and Using the Implicit Association Test: I. An Improved Scoring Algorithm,” Journal of Personality and Social Psychology, 85 (2), 197-216. Advances in Consumer Research (Volume 48) / 923

Harman, Graham (2005), Guerilla Metaphysics: Phenomenology Nass, Clifford, and Youngme Moon (2000), “Machines and and the Carpentry of Things, Peru, IL: Carus Publishing Mindlessness: Social Responses to Computers,” Journal of Company. Social Issues, 56 (1), 81-103. Hildebrand, Christian and Anouk S. Bergner (2019), “Detrimental Novak, Thomas P. and Donna L. Hoffman (2019), “Relationship Trust in Automation: How Conversational Robo Advisors Journeys in the Internet of Things: A New Framework for Leverage Trust and Miscalibrated Risk-Taking,” paper Understanding Interactions Between Consumers and Smart presented at the Society for Consumer Psychology Objects,” Journal of the Academy of Marketing Science, 47, Conference, Savannah, GA. 216-237. Hirschberg, Julia, and Christopher D. Manning (2015), “Advances Orgel, Etan, Robert McCarter, and Shana Jacobs (2010), “A Failing in Natural Language Processing,” Science, 349 (6245), 261-6. Medical Educational Model: A Self-Assessment by Physicians Hoffman, Donna L. and Thomas P. Novak (2018), “Consumer and at All Levels of Training of Ability and Comfort to Deliver Object Experience in the Internet of Things: An Assemblage Bad News,” Journal of Palliative Medicine, 13 (6), 677-83. Theory Approach,” Journal of Consumer Research, 44 Pickering, Andrew (1995), The Mangle of Practice: Time, Agency, (February), 1178-1204. and Science, Chicago: University of Chicago Press. Kim, Youjeong, and S. Shyam Sundar (2012), “Anthropomorphism Price, Linda L. and Amber M. Epp (2015), “The Heterogeneous and of Computers: Is It Mindful or Mindless?,” Computers in Open-Ended Project of Assembling Family,” in Assembling Human Behavior, 28 (1), 241-50. Consumption: Researching Actors, Networks and Markets, Kozinets, Robert V. (2008), “Technology/Ideology: How eds. Robin Canniford and Domen Bajde, London, UK: Ideological Fields Influence Consumers’ Technology Routledge, 59-76. Narratives,” Journal of Consumer Research, 34 (December), Robinette, Paul, Wenchen Li, Robert Allen, Ayanna M. Howard, 865-81. and Alan R. Wagner (2016), “Overtrust of Robots in Kozinets, Robert V., Anthony Patterson, and Rachel Ashman Emergency Evacuation Scenarios,” ACM/IEEE International (2016), “Networks of Desire: How Technology Increases Conference on Human-Robot Interaction (HRI 2016). our Passion to Consume,” Journal of Consumer Research 43 Schau, Hope Jensen and Mary C. Gilly (2003), “We Are What We (February), 659-82. Post? Self-Presentation in Personal Web Space,” Journal of Krakovna, Victoria (2018), “Specification Gaming Examples Consumer Research, 30 (December), 385-404. in AI,” https://vkrakovna.wordpress.com/2018/04/02/ Schroeder, Juliana, and Nicholas Epley (2016), “Mistaking Minds specification-gaming-examples-in-ai/, last accessed 2/14/29. and Machines: How Speech Affects Dehumanization and Latour, Bruno (1999), Pandora’s Hope: An Essay on the Reality of Anthropomorphism,” Journal of Experimental Psychology: Science Studies, Cambridge, MA: Harvard University Press. General, 145 (11), 1427-37. Lenz, Matthias, Susanne Buhse, Jürgen Kasper, Ramona Kupfer, Sherry, John F. (2000), “Place, Technology, and Representation,” Tanja Richter, and Ingrid Mühlhauser (2012), “Decision Aids Journal of Consumer Research, 27 (September), 273-8. for Patients,” Deutsches Ärzteblatt International, 109 (22-23), Simonite, Tom (2018), “When Bots Teach Themselves to Cheat,” 401. Wired, Business Section, August 8, https://www.wired.com/ Luo, Xueming, Silian Tong, Zheng Fang, and Zhe Qu (2019), story/when-bots-teach-themselves-to-cheat/ last accessed “Frontiers: Machines vs. Humans: The Impact of Artificial 2/14/19. Intelligence Chatbot Disclosure on Customer Purchases,” Sinha, Gunjan (2014), “Decision Aids Help Patients but Still Are Marketing Science, 38(6), November-December, 937-947. Not Widely Used,” JNCI: Journal of the National Cancer MacInnis, Debbie J. and Valerie S. Folkes (2017), “Humanizing Institute, 106 (7). Brands: When Brands Seem to Be Like Me, Part of Me, and Sood, Ashish and Gerard J. Tellis (2005), “Technological Evolution in a Relationship with Me,” Journal of Consumer Psychology, and Radical Innovation,” Journal of Marketing, 69 (3), 152- 27(3), 355–374. 68. Martin, Diane M. and John W. Schouten (2014), “Consumption- Stacey, Dawn, France Légaré, Krystina Lewis, Michael J. Barry, Driven Market Emergence,” Journal of Consumer Research, Carol L. Bennett, Karen B. Eden, Margaret Holmes‐Rovner et 40 (February), 855-70. al. (2006), “Decision Aids for People Facing Health Treatment Mast, Marianne Schmid, Annette Kindlimann, and Wolf Langewitz or Screening Decisions,” Cochrane Database of Systematic (2005), “Recipients’ Perspective on Breaking Bad News: How Reviews, 4. You Put It Really Makes A Difference,”Patient Education and Tao, Jianhua, Yongguo Kang, and Aijun Li (2006), “Prosody Counseling, 58 (3), 244-51. Conversion from Neutral Speech to Emotional Speech,” IEEE Mick, David Glen and Susan Fournier (1998), “Paradoxes of Transactions on Audio, Speech, and Language Processing, 14 Technology: Consumer Cognizance, Emotions. Ad Coping (4), 1145-54. Strategies,” Journal of Consumer Research, 25 (September), Tates, Kiek, Marjolijn L. Antheunis, Saskia Kanters, Theodoor 123-43. E. Nieboer, and Maria BE Gerritse (2017), “The Effect of Miller, Daniel (2005), Materiality, Durham, N.C.: Duke University Screen-To-Screen Versus Face-To-Face Consultation on Press. Doctor-Patient Communication: An Experimental Study with Monden, Kimberley R., Lonnie Gentry, and Thomas R. Cox (2016), Simulated Patients,” Journal of Medical Internet Research, 19 “Delivering Bad News to Patients,” in Baylor University (12). Medical Center Proceedings, 29 (1), 101-2, Taylor & Francis. Tegmark, M. (2017), Life 3.0: Being Human in the Age of Artificial Nass, Clifford, and Scott Brave (2005),Wired for Speech: Intelligence, New York: Borzoi Book. Alfred A. Knopf. How Voice Activates and Advances the Human-Computer Relationship, MIT Press. 924 / Consumers and Their Smart Devices: Perspectives on Anthropomorphism

Tian, Kelly, Pookie Sautter, Derek Fisher, Sarah Fischbach, Waytz, Adam, Joy Heafner, and Nicholas Epley (2014), “The Cuauhtemoc Lune-Nevarez, Kevin Boberg, Jim Kroger, Mind in the Machine: Anthropomorphism Increases Trust and Richard Vann (2014), “Transforming Health Care: in an Autonomous Vehicle,” Journal of Experimental Social Empowering Therapeutic Communities through Technology- Psychology, 52, 113–117. Enables Narratives,” Journal of Consumer Research, 41 White, Tom (2019), “Shared Visual Abstractions,”1st Workshop (August), 237-60. on Shared Visual Representations in Human and Machine Van de Kamp, Mark E. (2002), “Auditory Implicit Association Intelligence (SVRHM), NeurIPS 2019. https://arxiv.org/ Tests,” Doctoral Dissertation, University of Washington. pdf/1912.04217.pdf, accessed December 7. Wang, Wenhuan (2017), “Smartphones as Social Actors? Social Złotowski, Jakub, Diane Proudfoot, Kumar Yogeeswaran, Dispositional Factors in Assessing Anthropomorphism,” and Christoph Bartneck (2015), “Anthropomorphism: Computers in Human Behavior, 68, 334-44. Opportunities and Challenges in Human-Robot Interaction,” Waytz, Adam, Carey K. Morewedge, Nicholas Epley, George International Journal of Social Robotics, 7(3), 347–360. Monteleone, Jia-Hong Gao, and John T. Cacioppo (2010), “Making Sense by Making Sentient: Effectance Motivation Increases Anthropomorphism,” Journal of Personality and Social Psychology, 99 (3), 410–35. Antecedents and Consequences of Consumer Budgeting Chairs: Jennifer Lyu, University of Colorado-Boulder, USA Daniel Katz, University of Chicago, USA

Paper #1: The Influence of Budgets on Consumer Spending also examine the relationship between budgeting behavior and vari- Marcel Lukas, Heriot-Watt University, Scotland ous measures of financial well-being. They supplement their findings Chuck Howard, University of British Columbia, Canada from the survey using detailed administrative data for customers of the Commonwealth Bank of Australia. Paper #2: On the Psychology of Resource Monitoring Using experimental and observational data, these papers expand Daniel Katz, University of Chicago, USA the field’s understanding of consumer budgeting. They combine sev- Abigail B. Sussman, University of Chicago, USA eral methodologies and perspectives to provide novel insights into Paper #3: A Query Theory Explanation for Reactions to the psychological antecedents and real-world consequences of men- Constraint tal budgeting. Importantly, they address substantive questions about Matthew Meister, University of Colorado-Boulder, USA consumer financial behavior, including when and why people create John G. Lynch, University of Colorado-Boulder, USA mental budgets, how they adapt their budgets as constraints change, Paper #4: Understanding the Drivers of Household Budgeting and how budgets affect downstream behavior and financial well- C. Yiwei Zhang, University of Wisconsin-Madison, USA being. Abigail B. Sussman, University of Chicago, USA Jennifer Lyu, University of Colorado-Boulder, USA The Influence of Budgets on Consumer Spending Nathan Wang-Ly, Commonwealth Bank of Australia, Australia EXTENDED ABSTRACT SESSION OVERVIEW Personal finance advisors frequently invoke budgeting as a way When consumers have finite resources, budgeting can facilitate to curb overspending, increase savings, decrease debt, and improve optimal use of these resources. Budgeting involves monitoring and financial security (Bell 2019; Caldwell 2019; Credit Counselling So- regulating the use of a resource, such as money, with the goal of using ciety 2019). However, a growing chorus of voices has begun to argue a set amount over time for a particular purpose. Budgeting is often that budgeting simply doesn’t work (Elkins 2018; Olen 2015; Pratt assumed to be an important determinant of consumer spending, but 2019). Surprisingly, the academic literature on consumer budgeting there is surprisingly little work directly investigating how consumers has little to offer this debate: for all we know about the psychology set and manage their budgets over time. Furthermore, evidence on of budgeting, we know very little about the extent to which budgets the consequences of financial budgeting is mixed. The goal of this do (or do not) influence spending in the wild (Zhang and Sussman session is to provide new insights into factors that influence whether 2018). The goal of the present research is to address this gap in the consumers budget and the influences those budgets have on behavior. literature. In paper 1, the authors use data from a financial budgeting app Consumer budgeting theory asserts that when expenses are eas- to analyze budgeting behavior and subsequent spending patterns. ily booked and posted – as is the case when using a personal finance The authors find that consumers, on average, significantly overspend app – budgets will be “inflexible” (Heath and Soll 1996). This sug- relative to their budgets. However, there is a significant positive re- gests that consumers will tend to comply with their budgets (H1a). lationship between budget amount and actual spending. This pattern However, research on planning fallacies has demonstrated that in- suggests that the act of setting a low budget can help reduce spending dividuals’ plans for the future are generally optimistic (see Buehler, even if consumers do not fully adhere to their budgets. Griffin and Peetz 2010 for a review). This implies that consumers In paper 2, the authors explore when, why, and how people will spend more than they budget (H1b), because budgets are a plan monitor resources such as money, time, and calories. They find mon- for future spending (Lynch et al. 2010; Novemsky and Kahneman ey is monitored most closely while time and calories are monitored 2005). We contribute to each of these literatures by testing these similarly. The authors identify several factors that mediate this dif- competing hypotheses. ference across resources. This research aims to provide a unifying Regardless of whether or not consumers strictly comply with framework for understanding when resources are likely to be moni- their budgets it is still possible for budgets to influence their spend- tored and which resources will be budgeted similarly or differently. ing: If a consumer who typically spends $200 per month on dining In paper 3, the authors examine how responses to a change in and drinking sets a budget of $100 and ends up spending $150, then one’s budget constraint depend on the length of time that has elapsed this consumer’s budget has influenced their spending even though since one created the budget. The authors find that, when faced with they have not strictly complied with their budget. This possibility an unexpected increase in wealth, participants who had created bud- leads to the following hypotheses regarding the influence of budgets gets recently were more likely to stick to their budgets than partici- on spending: A lower budget will be associated with lower spending pants who had created budgets further in the past. The authors pro- (H2), and consumers who set budgets will spend less money post- vide process-level evidence to explain this phenomenon via memory budget than they did pre-budget (H3). interference, whereby recently created mental budgets inhibit one’s In study 1 we test H1-H3 using data provided by Money Dash- ability to consider alternative ways to use the money. board (MDB), a financial aggregation app with ~70,000 active users In paper 4, the authors use a large, nationally representative in the UK. The data set includes all user transactions – more than (U.S.) survey to explore how consumers set and manage their bud- 350 million – between January 2014 and December 2016. The novel gets. The authors examine several aspects of budgeting, including the structure of the data allows us to observe each user’s pre-budget financial periods over which consumers budget, the categories they spending (i.e., how much they spent before they set a budget), their set and track, and the ways they evaluate and adjust their budgets in budget, and their post-budget spending for six months. For concision response to unexpected income and expense changes. The authors

Advances in Consumer Research 925 Volume 48, ©2020 926 / Antecedents and Consequences of Consumer Budgeting we focus the present analysis on the two most popular budget cat- On the Psychology of Resource Monitoring egories: Dining and Drinking (n = 2,479) and Groceries (n = 2,618). Supporting H1b, the data reveal that budget compliance is gen- EXTENDED ABSTRACT erally weak: Actual spending is 36.2% higher than budgeted spend- Consumers often set goals that require managing limited re- ing for dining and drinking, and 21.2% higher for groceries (p’s < sources. One may have a goal of saving more money, spending more .001). Supporting H2, panel regression analysis with spending as the time with family, or limiting caloric intake. Accurately monitoring dependent variable and budget and login frequency as independent one’s use of resources can help with achieving these goals (Koo & variables, shows that lower budgets are associated with lower spend- Fishbach 2012). Prior research typically investigates monitoring re- ing for both dining and drinking (B = .28, p < .001) and groceries (B sources such as money, time, and calories separately. In papers that = .57, p < .001), and that higher login frequency is associated with do compare these resources, some have found that they are tracked lower spending (BD&D = -12.22, BGROC = -12.57, p < .001). Support- similarly (Morewedge, Holtzman, & Epley, 2007; Sussman, Alter ing H3, post-budget spending in the month after budget creation is & Paley, 2016) while others found stark differences (Soman 2001). 15.3% lower than pre-budget spending for dining and drinking (p < There is little in the way of a unifying framework to explain when .001) and 6.8% lower for groceries (p < .001). Furthermore, spending resources are likely to be monitored, and whether different resources continues to decrease over time: spending on dining and drinking is will be monitored similarly or differently. This research aims to un- 16.2% lower six months after budget creation than it was in the first derstand when, why, and how people monitor resources and test for month after budget creation (p < .001), and spending on groceries is similarities and differences in the monitoring of money, time, and 5.4% lower (p < .001). However, even after six months of budgeting calories. consumers still spend significantly more than they budgets in both categories (p’s < .001). Taken together, the results of study 1 suggest Study Overview that budgets are (persistently) optimistic but also highly influential. , We conducted a preregistered study on a nationally representa- and spending on groceries is 5.4% lower (ted with lower spending in tive (U.S.) sample of 745 participants using a 3(Resource: Money, both categories (nking pp, and a vector of control Time, Calories; between) x 4(Time Horizon: Day, Week, Month, The strengths of study 1 include unprecedented scope, ecologi- Year; within) mixed design. In addition to the results reported below, cal validity, and longitudinal measurement; one limitation is that it we both directly and conceptually replicated this study using a with- is entirely descriptive. Study 2 addresses this by randomly assigning in-subject design, finding largely consistent results in each case in consumers (n = 340, MTurk) to one of two budget forecast condi- samples from Mechanical Turk. These results are not reported here tions that were designed to produce relatively more or less optimistic due to space constraints. budget forecasts. In the control condition participants were instruct- Participants created budgets for their resource in each time ho- ed to estimate their total spending for the next week. Consistent with rizon and responded to several questions related to the budgets they the results of study 1, our expectation was that budget forecasts in set. For each time horizon, participants completed a three-item scale, this condition would be relatively optimistic. In the “outside-view” adapted from Soman (2001), to capture how intensely they track the condition participants were instructed to estimate their spending for use of their resource over a given time horizon. (Cronbach’s alpha > the past week before estimating their spending for the next week. 0.8 for all 12 Resource x Time Horizon scales). We averaged across Our expectation was that forecasts in the outside-view condition items (separately for each Resource x Time Horizon scale) to create would be higher than in the control condition, because prompting a Resource Monitoring Score. people to take an outside-view (i.e., consider relevant past behavior) Participants also responded to eight questions designed to better has been shown to produce higher forecasts regarding future behav- understand reasons for differences in monitoring across resources. ior (Buehler, Griffin and Ross 1994; Peetz and Buehler 2012). Par- These questions were theoretically motivated by literature on plan- ticipants then reported their actual spending in an online financial ning, budgeting, and perceived control. The eight potential media- diary at the end of each day during the target week. tors were the perceived helpfulness of setting a budget, similarity As expected, participants in the outside-view condition made of resource use in subsequent periods, ability to prepare for future significantly higher budget forecasts (M = $255.44) than participants resource use, ability to make up for past resource use, pain of setting in the control condition (M = $189.88; p = .004). Supporting H2, par- the budget, perceived control over sticking to their budget, the per- ticipants in the outside-view condition also reported higher spending cent of the budget deemed discretionary, and the number of budget during the target week (M = $498.35) than participants in the control categories created. These variables were measured on seven-point condition (M = $374.32; p < .001). Thus, study 2 extends study 1 by scales, with the exceptions of the number of categories created (min- providing causal evidence that lower budgets lead to lower spending, imum of one and maximum of twelve) and the percent of the budget even when budgets are optimistic and compliance is weak. deemed discretionary. In sum, the present research makes three contributions. First, we show that budgets do exert substantial influence on spending, Results even if they are not as inflexible as was once believed. Second, we The extent to which participants reported engaging in resource provide evidence that the influence of budgets on spending persists monitoring varied significantly by resource and time horizon. over time, even when budget compliance is weak. This suggests that Collapsing across time horizons, money was tracked most beliefs about the negative effect of goal violation on subsequent be- closely (MMoney= 4.76; MTime= 3.63; MCalories= 3.37; Money vs (Time havior (e.g., Soman and Cheema 2004) may need to be updated. Fi- & Calories) p<.001). To test the impact of time horizon, we ran fol- nally, our results indicate that under-predicting future spending can lowing linear mixed model, with money as the reference resource, be beneficial. This implies that helping consumers improve their fi- time horizon measured in log-days (grand mean centered), and Re- nancial prediction accuracy (e.g., Howard et al. 2018) may not be as source Monitoring Score denoted by RMS: beneficial as is commonly thought. Advances in Consumer Research (Volume 48) / 927

The monitoring of money followed an inverse-U pattern, peak- ple fail to prioritize because they get caught in the status quo, and ing at a month and then decreasing (β3= 0.03, p= .03; β4= -0.05, can’t consider substantial alternatives. p<.001). In contrast, monitoring of time and calories peaked at a Our hypothesis is that the initial act of budgeting makes that day and steadily declined at longer intervals (β3+β5= -0.08, β+β7= chosen plan accessible, reducing recall of alternatives and compet-

-0.14; both p’s<.001; β4+β6 and β4+β8 n.s.). The results do not change ing uses of money through output interference (Roediger & Schmidt, qualitatively if we treat time horizon as a categorical variable. These 1980). If decisions about reacting to constraints are the result of a findings suggest time horizons play an important role in how closely process similar to the self-questioning described by query theory people monitor resources. However, across all time horizons, money (Johnson, Häubl, & Keinan, 2007; Weber et al., 2007), a highly ac- was monitored most closely and time and calories were always mon- cessible status quo should influence one’s information search. As a itored to a similar degree (β1= -1.31, p<.001; β2= -1.63, p<.001; β1-β2 result, people will overlook important information, such as oppor- n.s.). Thus, we collapsed across time horizons and used the contrast tunity costs, alternative needs, and high-order goals (Kruglanski et of Money vs (Time & Calories) for our mediation analysis. al., 2002). Two studies provide evidence for this hypothesis. In both stud- Drivers of Resource Monitoring ies, interference is manipulated such that one condition keeps the First, we ran a factor analysis to see whether the eight poten- status quo top of mind and the other suppresses it. After the inter- tial mediators shared underlying constructs. This analysis suggested ference manipulation, participants received a hypothetical financial three factors. The first factor (“helpfulness”) combines perceived shock. When the initial plan was top of mind at the time of shock, helpfulness with the ability to make up for and prepare for high/low participants responded by making slight mutations to their original levels of resource use, suggesting that the ability to make up for and plans, or by making no change at all. However, when the initial plan prepare for differential resource use may contribute to perceptions of was suppressed, they were more likely to make large adjustments, helpfulness. The second factor (“pain”) combines the pain of budget- including completely giving up on that plan. Preliminary evidence ing with perceived control over sticking to a budget and similarity of shows that this relationship may be mediated by participants’ degree budgets from period to period, suggesting that lack of control over of opportunity cost consideration at the time of the financial shock. sticking to their budget and budget dissimilarity may contribute to Specifically, participants report to have thought more about oppor- pain of budgeting. The third factor (“other”) includes the percent of tunity costs and alternative desires when released from memory in- the budget that is discretionary and the number of categories formed. terference. Potential mediators were assigned to the factors with their In study 1, all participants budgeted five personal and five gift highest loadings and averaged within each factor to create three new shopping plans for Black Friday. In this study, memory interference variables, and then averaged across time horizons. We ran a multiple was manipulated within-subjects by having participants react imme- mediation analysis with the three factors as mediators (in parallel) diately after creating their plans and again after three days, with the to see if any of them explain why money is monitored more closely size of a shock received ($100 or $300) and the plan adjusted (per- than time and calories. We calculated bootstrapped 95% confidence sonal vs gift purchases) counterbalanced. All participants receive intervals for the three indirect effects. both shocks and adjust both plans, while the specific time that they Only the first factor had a significant indirect effect. This sug- do so varies. Study 1 shows substantially larger adjustments made gests a resource will be tracked more closely when the perceived after a delay than before a delay (F(1, 467) = 26.46, p < 0.001) after benefit of planning is higher, which depends on the perceived ability controlling for a multitude of background factors, including how fi- to create and execute plans. Other factors considered, such as the nancially constrained a participant felt at the beginning of the study. pain of creating a budget, did not underlie resource monitoring. Importantly, none of the background factors interact with our central factor of time. Conclusion To add substance to these results, we also had participants esti- We find significant differences in resource monitoring within mate the amount of money they would save by adjusting their plans and across resources. Regardless of time horizon, money was tracked after a shock. Our hypothesis for these data is that people will esti- more closely than time and calories. Our mediation analysis suggests mate higher savings after a time delay, because they will believe that this difference was due to the ability to plan and the perceived benefit more substantial adjustments bring higher savings. The evidence for of planning. this hypothesis is unclear. In study 1, there is a significant effect of Understanding differences in the propensity to engage inre- time on log-transformed savings (F(1, 466) = 18.32, p < 0.001). source monitoring will help researchers anticipate whether known Study 2 manipulates interference and the size of shock between- tendencies investigated in the context of a single resource are likely subjects. Interference is also manipulated not with time, but by hav- to generalize to others. For example, these results can help to predict ing participants either “rehearse” their plan in a writing task or com- when interventions to reduce frivolous spending would or would not plete an unrelated task over the same time period. Written responses be easily adapted to reduce time-wasting and snacking. were coded by hypothesis-blind assistants for several behaviors. At the time of this writing, the coding is not quite complete. We inves- A Query Theory Explanation for Reactions to Constraint tigate whether those who were released from interference were more likely to make a change to their plan, to drop their plan entirely, and EXTENDED ABSTRACT to come up with new means to reach the same end as their original Fernbach, Kan and Lynch (2015) found that people avoid mak- plans. Additionally, we consider whether the effect of financial shock ing substantial adjustments to ongoing spending plans in the face of on these behaviors is impacted by interference, predicting signifi- constraint. Those authors provide a reason-based explanation – that cant interactions between levels of financial shock ($10 vs $250 and substantial changes (priority plans) require cutting things out, which $500) and interference condition. These results would indicate that guarantees feelings of loss and pain for consumers. This research those in the release condition were more affected by the existence of suggests another explanation for peoples’ inability to prioritize. Peo- a meaningful increase in constraint. 928 / Antecedents and Consequences of Consumer Budgeting

A third study complements the second, using the same method We surveyed 3,826 respondents from a nationally representa- but focusing on a different dependent variable. In this study, we had tive online panel to gain insight into the budgeting process. We ex- participants report their level of opportunity cost consideration on a amine how people set and manage their budgets, including the time 5-item scale, partially adapted from Spiller (2011), with two items horizons over which they budget, the categories they set and track, added by the authors. The results of this study mirror those predicted and the ways they evaluate and adjust their budgets in response to from the second, showing that those in the release condition consider income and expense changes. We discuss the implications of our re- more opportunity costs after their shock (F(1, 365) = 5.51, p = 0.019) sults for mental accounting models or consumption-savings models and that those who receive a meaningful shock also consider more and long-term financial well-being. opportunity costs (F(1, 365) = 11.36, p < 0.001). We take this as evi- We characterize a number of important features of budgeting dence in support of the hypothesis that people who are released from behavior, including: First, budgeting is highly prevalent across the memory interference are better able to think outside of the status quo income distribution. Of our full sample, roughly 65.6 percent of re- and consider the entirety of their wants and needs. spondents report that they currently budget, either formally or infor- A fourth study is planned to investigate differences in consider- mally, while 14.5 percent report having budgeted in the past but not ation sets between conditions. Rather than having participants state a in the present and 19.9 percent report having never budgeted. Sec- single plan they intend to follow, we will ask them to come up with ond, nearly all individuals who budget do so by tracking their spend- a set of options. Our prediction is that those options should be less ing within distinct categories of consumption. We find that over 90 similar to the initial plan when they have been released from interfer- percent of individuals who budget do so using budgetary categories. ence, and that the set of options should be more dispersed than in the Third, there is wide heterogeneity in the level of granularity at which rehearse condition. individuals categorize. For instance, some simply distinguish neces- Reacting to unexpected constraint is an important task for con- sities from discretionary spending, while others maintain separate sumers, which prior research has shown to be difficult (Fernbach et categories for internet, water, and gas utilities. Fourth, individuals al., 2015). Constraints are rarely static, though budgets may remain respond asymmetrically when faced with too little versus too much so. In this project, we hope to highlight the role of memory factors slack in their budgets. While more than 85 percent of respondents in peoples’ reactions to constraint (Alba & Chattopadhyay, 1985), would either adjust their spending or update their spending limits and intend to build towards an intervention to help consumers adjust if they overspent within a budget category, fewer than 30 percent their budgets more successfully. would make such changes if they underspent their limit. In other words, people appear to update their budgets conservatively, which How Consumers Budget may facilitate savings if leftover slack is directed towards savings. Fifth, the frequency with which individuals assess the state of their EXTENDED ABSTRACT current spending relative to their intended budgets varies substan- Every day, consumers make financial decisions that collective- tially and is highly correlated with perceived financial well-being. ly determine their overall financial well-being. Understanding how We supplement our findings from the survey data using detailed consumers make these spending and savings decisions is important administrative data for customers of the Commonwealth Bank of for researchers and practitioners alike (Madrian et al., 2017). Bud- Australia (CBA). We focus this analysis on the relationship between geting is a common method of managing household finances and a financial well-being and budget engagement (the fifth observation key factor in determining how much to spend or save (Thaler, 1985; above) and the tendency to create distinct budgeting categories (the Heath and Soll, 1996). Yet, surprisingly little is known about how second observation above), since these have the closest analogues in people set and manage their budgets over time (see Zhang and Suss- the administrative data. First, data from a random sample of 100,000 man, 2017 and 2018 for reviews). Research on household financial customers of CBA corroborates the finding that people engage more management has largely focused on long-term financial planning with their budgets when they are more financially constrained. Spe- (e.g., retirement planning), with comparatively little research on cifically, we find a significant negative correlation (r=-0.34) be- short-term financial management (i.e., budgeting). Given the preva- tween financial well-being scores and the average number of days lence of consumers who seek to improve their financial well-being per month interacting with the budgeting app. Second, analysis of and the importance of daily financial management for financial well- a separate 100,000 of customers at CBA who hold at least 2 credit being, it is crucial that researchers examine the budgeting process or debit cards examines the extent to which customers concentrate and its psychological underpinnings. their spending on a single card. We find that approximately a quarter Furthermore, recent empirical evidence suggests that a sub- (23.1%) of customers in the sample have more than 75% of their stantial number of households are unable to cope with even mod- spending on one of their multiple cards concentrated in a single erate transitory financial shocks (Lusardi, Schneider, and Tufano, spending category (e.g., groceries or transportation). Furthermore, 2011; Pew Charitable Trusts, 2015). For these financially fragile 11.3% of customers have more than 90% of their spending concen- households, it may be especially important to understand how in- trated in a single spending category. These patterns are suggestive of dividuals manage and rebalance their budgets on a day-to-day basis using cards as a budgeting tool. in response to various shocks. More broadly, better understanding Our findings lay the groundwork for future analyses of house- everyday budgeting strategies will help policymakers and industry hold financial management behaviors and provide novel insight into leaders develop effective programs to improve long-term financial budgeting behaviors and beliefs. The increasing capabilities of fi- well-being. nancial technology to facilitate and otherwise influence budgeting In the present research, we seek to contribute to the growing make this an exciting time to explore many future avenues of re- body of knowledge on household budgeting by investigating the search examining both short-term budgeting and long-term financial budgeting behaviors and beliefs of a nationally-representative sam- well-being. ple. We further extend our work by exploring a subset of findings “in the wild,” through field data provided by the Commonwealth Bank of Australia. Advances in Consumer Research (Volume 48) / 929 REFERENCES Morewedge, C. K., Holtzman, L., & Epley, N. (2007). Unfixed Alba, J. W., & Chattopadhyay, A. (1985). Effects of Context and resources: Perceived costs, consumption, and the accessible Part-Category Cues on Recall of Competing Brands. Journal account effect.Journal of Consumer Research, 34(4), 459-467. of Marketing Research, 22(3), 340–349. Olen, Helen (2015). Toss Your Budget - Why a pillar of personal Bell, Amy. (2019). “Six Reasons Why You Need a Budget,” https:// finance isn’t nearly as essential as we think. On http://www. www.investopedia.com/financial-edge/1109/6-reasons-why- slate.com/articles/business/the_bills/2015/06/ you-need-a-budget.aspx personal_budgets_are_overrated_seriously.single.html, Buehler, Roger, Dale Griffin, & Johanna Peetz. (2010). The accessed: 15.01.2019 planning fallacy: Cognitive, motivational, and social origins. Pew Charitable Trusts (2015), “The role of emergency savings in In Advances in experimental social psychology (Vol. 43, pp. family financial security: How do families cope with financial 1-62). Academic Press. shocks?” https://www.pewtrusts.org/en/research-and-analysis/ Caldwell, Miriam (2019). “Reasons Why You Should Budget Your issue-briefs/2015/10/the-role-of-emergency-savings-in-family- Money,” https://www.thebalance.com/reasons-to-budget- financial-security-how-do-families. money-2385699 Pratt, Bill (2019). “Why Budgets Don’t Work,” https://www. Credit Counselling Society 2019. “What is Budgeting? What is a themoneyprofessors.com/why-budgets-dont-work/ Budget?” https://www.mymoneycoach.ca/budgeting/what-is-a- Rajagopal, P., & Rha, J. Y. (2009). The mental accounting of time. budget-planning-forecasting Journal of Economic Psychology, 30(5), 772-781. Elkins, Kathleen (2018). “39-year-old retired millionaire: Roediger, H. L., & Schmidt, S. R. (1980). Output interference in ‘Budgets don’t work’—do this instead,” https://www.cnbc. the recall of categorized and paired-associate lists. Journal com/2018/12/26/39-year-old-retired-millionaire-budgets-dont- of Experimental Psychology: Human Learning and Memory, work-do-this-instead.html 6(1), 91–105. Fernbach, P. M., Kan, C., & Lynch, J. G. (2015). Squeezed: Coping Soman, D. (2001). The mental accounting of sunk time costs: with Constraint through Efficiency and Prioritization.Journal Why time is not like money. Journal of Behavioral Decision of Consumer Research, 41, 1204–1227. Making, 14(3), 169-185. Heath, Chip, & Jack B. Soll (1996). Mental budgeting and Soman, D., & Cheema, A. (2004). When goals are consumer decisions. Journal of consumer research, 23(1), counterproductive: The effects of violation of a behavioral 40-52. goal on subsequent performance. Journal of Consumer Howard, Ray Charles, David J. Hardisty, & Abigail B. Sussman Research 31, no. 1 (2004): 52-62. (2018). A Prototype Theory of Consumer Financial Spiller, S. A. (2011). Opportunity Cost Consideration. Journal of Misprediction. Boulder Summer Conference on Consumer Consumer Research, 38, 595-610. Financial Decision Making. https://pdfs.semanticscholar.org/5 Sussman, Abigail B., Adam L. Alter, and Anna Paley. “Mental 77a/96809a18b1d7ac8087d6926529878019920d.pdf accounting for food in exceptional contexts.” ACR North Jamison, J., Johnson, E. J., Lynch, J. G., Meier, S. and Rick, S. American Advances (2016). (2017). Behaviorally informed policies for household financial Sussman, A. B., Alter, A. L., & Paley, A. (2017). Mental accounting decisionmaking. Behavioral Science & Policy, 3(1), pp.26-40. for food: Booking, posting, and incorporating exceptional Johnson, E. J., Häubl, G., & Keinan, A. (2007). Aspects of consumption. Working Paper. endowment: A query theory of value construction. Journal of Thaler, R. (1985). Mental Accounting and Consumer Choice. Experimental Psychology: Learning, Memory, and Cognition, Marketing Science, 4(3), 199-214. 33(3), 461–474. Weber, E. U., Johnson, E. J., Milch, K. F., Chang, H., Brodscholl, Koo, M., & Fishbach, A. (2012). The small-area hypothesis: J. C., & Goldstein, D. G. (2007). Asymmetric Discounting in Effects of progress monitoring on goal adherence.Journal of Intertemporal Choice. Psychological Science, 18(6), 516–523. Consumer Research, 39(3), 493-509. Zhang, C. Y. & Sussman, A. B.. (2017). The role of mental Kruglanski, A. W., Shah, J. Y., Fishbach, A., Friedman, R., Woo accounting in household spending and investing decisions. C. Young Chun, & Sleeth-Keppler, D. (2002). A theory of goal Chaffin (Ed.), Client Psychology. New York: Wiley, 19-07. systems. Advances in Experimental Social Psychology, 34, Zhang, C. Y. & Sussman, A. B. (2018). Perspectives on mental 331–378. accounting: An exploration of budgeting and investing. Lusardi, A., Schneider, D., & Tufano, P. (2011). Financially Fragile Financial Planning Review, 1(1-2), p.e1011. Households: Evidence and Implications. Brookings Papers on Economic Activity, 2011(1), 83-134. Madrian, B. C., Hershfield, H. E., Sussman, A. B., Bhargava, S., Burke, J., Huettel, S. A., Exploring the Effective Use of Humor in Consumer Contexts Chair: Holly Howe, Duke University, USA

Paper #1: Aha over Haha: Brands Benefit More from Being hances perceptions of warmth. Papers 3 and 4 explore when and how Clever Than from Being Funny consumers effectively use humor in interpersonal contexts. In Paper Holly Howe, Duke University, USA 3, Yi, Warren & Berger examine how humor use affects consumers’ Lingrui Zhou, Duke University, USA online sharing behavior, demonstrating that humorous content is not Rodrigo Dias, Duke University, USA always more likely to be shared. In Paper 4, Bitterly explains why Gavan Fitzsimons, Duke University, USA using humor may be especially desirable for low-power individuals. Then, he shows that, despite these benefits, low power individuals Paper #2: How Different Types of Humor in Advertising Shape are unlikely to use humor. Impressions of Firms’ Competence and Warmth Together, these papers shed light on the contexts in which hu- Chi Hoang, Nanyang Technological University, Singapore mor use is most common and most effective. This session should Klemens Knoferle, BI Norwegian Business School, Norway have broad appeal to researchers studying humor, communication, Luk Warlop, BI Norwegian Business School, Norway advertising and self-presentation. Paper #3: Humor Makes Consumers More Likely to Share Negative Content, But Not Positive Content Aha over Haha: Brands Benefit More from Being Clever John Yi, University of Arizona, USA Than from Being Funny Caleb Warren, University of Arizona, USA Jonah Berger, University of Pennsylvania, USA EXTENDED ABSTRACT Paper #4: How Power Influences the Use of Humor Advertisers are increasingly turning to humor to engage con- T. Bradford Bitterly, Hong Kong University of Science and sumers. In 2020, approximately 70% of Superbowl ads used humor, Technology, China more than ever before (Ace Metrix 2020). With the popularity of humorous advertising, how might a brand break through? In this paper, SESSION OVERVIEW we demonstrate that clever humor is more effective at boosting brand Humor is a pervasive self-presentation strategy for advertisers, perceptions than other forms of humor. managers, and individual consumers. At first glance, humor can seem Clever humor is any benign violation (Warren and McGraw like a panacea: it grabs attention (Eisend 2009), promotes positive 2016a) that lacks order or regularity (McQuarrie and Mick 1996). In emotions (Eisend 2009), diffuses awkward situations (Bitterly and other words, the meaning of a clever joke is not immediately appar- Schweitzer 2019; Martin 2007), and engenders interpersonal close- ent to a consumer and can only be fully understood when she applies ness (Fraley and Aron 2004). However, humor may not be as uni- her own knowledge (e.g., cultural references, puns, punchlines; Mc- versally effective as it initially seems. Humorous ads often failto Quarrie and Mick 1996). improve brand attitude or sales (Eisend 2009; Warren and McGraw Clever humor is generally believed to be more affiliative than 2016b), impair perceptions of competence (Bitterly, Brooks, and other forms of humor (e.g., self-deprecation, physical comedy; Craik, Schweitzer 2016), and promote inaction or offense (McGraw, Schiro, Lampert, and Nelson 1996; Martin et al. 2003) and, unlike many and Fernbach 2015; McGraw, Warren, and Kan 2013). other forms of humor, is associated with logic (Burro et al. 2018) and There is a growing interest among marketers and consumers intelligence (Craik et al. 1996; Didonato, Bedminster, and Machel to understand when humor is effective and appropriate. Based on 2013). Based on this prior work, we hypothesize that brands who use this research, we know that the efficacy of humor varies based on clever humor will be perceived as both warmer and more competent the strength and type of the inherent violation (Warren and McGraw than brands who use other forms of humor. This increase in warmth 2016b), the perceived appropriateness of the joke (Bitterly et al. and competence then leads to more favorable brand attitudes. 2016), the relationship between the joker and the receiver (Hall and Three pre-registered experiments test our predictions. As we Sereno 2010), and the relationship between the humor content and pre-registered, all responses from MTurk bots were excluded. Sam- the situation (Bitterly and Schweitzer 2019; Spotts, Weinberger, and ples sizes reported below are after this exclusion. Parsons 1997). Although we are beginning to understand when and In experiment 1, participants (N = 605; MTurk) were randomly why jokes fail, this has not necessarily led to an understanding of assigned to view a serious, funny, or clever video advertisement for when humor appeals are particularly successful. In addition, even Smart Car. Immediately after viewing the ad, participants rated Smart when consumers and/or marketers have a good sense that an indi- Car on warmth and competence (Kervyn, Fiske, and Malone 2012) vidual appeal would indeed be funny, we lack an understanding of and then completed a measure of brand attitude (Matz et al. 2019). whether they might feel comfortable using humor. As predicted, we observed a significant effect of ad on warmth, F(2, In summary, the prior literature reveals that humor is a powerful 591) = 6.49, p = .002, and competence, F(2, 591) = 33.33, p < .001. interpersonal strategy as long as it is used correctly. In this session, Smart Car was rated as warmer and more competent following a we explore the situations in which humor should be most impactful clever ad (M(SD)warm = 3.80(.89); M(SD)competent = 4.15(.78)) than a and when consumers feel comfortable using humor. Papers 1 and 2 funny ad (M(SD)warm = 3.49(.93); M(SD)competent = 3.53(.97)), both ts explore best practices for humor in advertising. In Paper 1, Howe and > 3.59, ps < .001. Using a clever ad did not boost perceptions of colleagues demonstrate that clever advertisements confer more posi- competence over the serious ad (M(SD) = 4.09(.73)), t(591) = .72, p tive brand perceptions than purely funny advertisements. In Paper 2, = .47, but it did confer more warmth than the serious ad (M(SD) = Hoang, Knoferle and Warlop examine the effects of incongruity-res- 3.64(.88)), t(591) = 1.95, p = .05. Perceptions of warmth, ab[95CI] = olution, tension-relief and disparaging humor on brand perceptions. .12[.05, .21], and competence, ab[95CI] = .49[.34, .66], mediated the Specifically, they demonstrate that incongruity-resolution humor effect of clever (vs. funny) humor on brand attitude. improves perceptions of competence while tension-relief humor en-

Experiment 2 replicated this effect using a wider variety of morous disparagement (Speck 1991). Incongruity-resolution repre- advertisements and a correlational design. Participants (N = 932; sents a humor process in which perceivers first experience something MTurk) watched one of a possible 30 ads from the brands Smart surprising, peculiar, or unusual and later are able to resolve it. This Car (n = 10), FirstBank (n = 10) and T-mobile (n = 10). Immediately process is characteristic of humor forms such as puns, punch lines, after viewing the ad, participants rated the ad on how clever and or comic irony. Both the production and the appreciation of incon- funny it was. Next, they rated the extent to which the advertisement gruity require qualities such as intelligence, creativity, and linguistic made the brand seem warm and competent, and then provided their skills (to a larger extent than other humor types). We thus predict that brand attitude. All the analyses reported below control for ad dum- incongruous humor, once understood, triggers impressions of adver- mies and funniness of the advertisement. In a hierarchical regression, tisers’ competence. Tension-relief, on the other hand, occurs when the cleverness of the advertisements was a positive predictor of both perceivers experience relief from some kind of strain and burst into warmth, β = .43, p <.001, and competence, β = .54, p < .001. Ad laughter as a way to dissipate nervous energy. During a tension-relief cleverness explained additional variance in both warmth, ΔR2 = .09, episode, perceivers form empathetic bonding with people (or per- F(1, 900) = 139.10, p < .001, and competence, ΔR2 = .14, F(1, 900) sonified creatures) associated with the experience (Speck 1991). We = 222.79, p < .001, above that explained by ad dummies and funni- predict that this positive affect can be transferred to improve impres- ness. Warmth, ab[95CI] = .12[.08, .16], and competence, ab[95CI] sions of advertisers’ warmth, but not competence. Lastly, humorous = .14[.09, .18], again mediated the relationship between cleverness disparagement mostly harms (rather than improves) impressions of and brand attitude. firms’ warmth and competence, since it may imply unjust ridicule Experiment 3 tested whether the effect of clever advertisements and ignorance of social boundaries (i.e. other-disparagement) or is moderated by self-brand connection. We reasoned that consum- draw attention to the deprecators’ real weaknesses (i.e. self-dispar- ers who already feel connected to a brand would be less affected by agement) (Greengross and Miller 2008). To test our hypotheses, we clever humor than consumers who have not yet formed a connection conducted a correlational study and a follow-up experiment. to the brand. Participants (N = 673; MTurk) completed a very similar Study 1 (N = 5768) is a large-scale correlational study examin- procedure to Study 2 with the following changes. First, participants ing the relationships between the four humor executions and partici- viewed a subset of ads from Study 2: Smart Car (n = 6), T-mobile pants’ impressions of advertisers using them. In order to provide a (n = 6) or First Bank (n = 6). Second, immediately after completing managerially relevant test of our predictions, we collected 300 ads the funniness and cleverness ratings, all participants completed the from 65 major U.S. magazines and two open databases (theadcollec- measure of self-brand connection from Escalas & Bettman (2015). tion.tumblr.com and adoftheworld.com). We distributed these ads to All the analyses reported below control for ad dummies, self-brand thirteen groups of Amazon Mechanical Turk workers; each group of connection, and funniness of the advertisement. In a hierarchical lin- workers rated a subset of the ads on a different variable. Specifically, ear regression, cleverness was a significant predictor of warmth (β a group of raters evaluated the ads in terms of either (1) the extent to = .36, p < .001) and competence (β = .53, p < .001). Importantly, which the ads used each of the four humor executions, (2) impres- we also observed the expected interaction between cleverness and sions about the advertisers’ competence and warmth, or (3) several self-brand connection on both warmth (β = -.35, p = .01) and com- control variables. The results indicated that incongruity-resolution petence (β = -.49, p < .001). The effect of cleverness on warmth and humor is the only humor execution that specifically enhances con- competence was stronger when participants were low in self-brand sumers’ perception of advertising firms’ competence, tension-relief connection than when they were high. In a moderated mediation humor is the only humor execution that enhanced impressions of model, the indirect effect of cleverness on brand attitude via warmth firms’ warmth. Meanwhile, both self- and other-disparagement had and competence was weaker at high levels of self-brand connection. negative effects on impression of firms’ competence, and other-dis- In summary, three experiments demonstrate that consumers paragement also negatively influenced impression of firms’ warmth. have more favorable brand attitudes for brands that use clever humor We obtained the same patterns of the results when using different compared to non-clever humor or serious advertisements. This oc- operationalizations of firms’ warmth and competence. curs because using clever humor makes a brand seem warmer and A potential alternative explanation for the correlational findings more competent. This research advances understanding of humor in of Study 1 is that competent firms selectively use humorous incon- advertising and sheds light on how specific forms of humor might be gruity, and warm firms selectively use arousal-safety humor in their more effective than others. Our findings have implications for mar- ads (rather than these humor types triggering impressions of firms’ keters looking to break through a media market that is saturated with competence and warmth, respectively). To test this explanation, humorous advertising. Study 2 (N = 348) used an experimental approach, contrasting the effects of four groups of humorous ads against a group of control ads, How Different Types of Humor in Advertising Shape while controlling for the baseline perception of firms. Furthermore, Impressions of Firms’ Competence and Warmth in Study 2, we aimed to demonstrate downstream consequences of humor types (i.e., likelihood to invest in a competent firm versus per- EXTENDED ABSTRACT ception of a warm firm’s corporate social responsibility) beyond the Humor is a common executional tactic in advertising (Eisend effects of humor on impressions of firms’ competence and warmth. 2009). Given that humor affects impression formation in inter-in- Regarding the effects of humor types on impressions of firms’ com- dividual settings (Greengross, Martin, and Miller 2012), we pro- petence, the results showed that incongruity-resolution humor is the pose that humor influences the impressions that consumers form only humor type that enhanced impressions of firms’ competence, of companies that use humor in their advertising. Furthermore, we corroborating the results from Study 1. Furthermore, incongruity- argue that although observing humor in advertising influences both resolution humor is also the only humor type that significantly in- warmth and competence impressions, the two impression types are creased participants’ likelihood to invest in the advertising firms affected differently by different humor executions. and participants’ expected return-on-investment when investing in Prior research has identified three humor processes that can the firms (competence-related downstream consequences). Regard- elicit laughter; namely incongruity-resolution, tension-relief, and hu- ing the effects of humor types on impressions of firms’ warmth, we 932 / Exploring the Effective Use of Humor in Consumer Contexts found that only other-disparagement had a significant and negative lieve that they can make a more favorable impression by posting effect on impression of firms’ warmth, while other types of humor content with a positive valence (East, Hammond, and Wright 2007) did not significantly influence impression of firms’ warmth, partially positive word of mouth (PWOM. Humor, interestingly, does not confirming the results from Study 1. However, none of the humor necessarily make content more positive (Warren, Barsky, and Mc- types significantly influenced warmth-related downstream conse- Graw 2018). Consumers perceive that something is humorous when quences such as perception of firms’ commitment to corporate so- it threatens their well-being, identity, or normative belief structure cial responsibility activities or the expected percentage of the firms’ (i.e., a violation appraisal), but the threat seems harmless, accept- profit contributed to these activities. able, or inconsequential (i.e., benign appraisal; McGraw and Warren We also conducted a supplementary study (N = 301) to test a 2010). Whether the addition of humor makes something more posi- competence projection process as a potential driver of the effect of tive or more negative thus depends on whether it is created by mak- incongruity-resolution on perception of firms’ competence. Extant ing an existing threat (i.e., violation) seem more benign or by adding research supports that appreciation of humor, especially incongruity- a violation to an otherwise benign situation. For example, humor based humor, signals similarity between the perceivers and the source makes complaints more positive but praise more negative (McGraw of humor (Flamson and Barrett 2008). On one hand, similarity is an et al. 2013). Similarly, comedy reliably lifts people’s mood when antecedent of projection, the act of introspecting one’s own quali- they are coping with a negative experience but not when everything ties and ascribing them to others (Ames 2004). On the other hand, is rosy (Warren et al. 2018). We consequently hypothesize that the resolution of incongruity may signal to the perceivers their mastery effect of humor on sharing will depend on the initial valence of the of competent skills such as flexible thinking and problem solving content. Adding humor to negative content (e.g., an unattractive pho- (Martin 2010). We thus proposed that perceivers, while resolving tograph or embarrassing memory) will make consumers more likely and appreciating incongruity-based humor, may project their per- to share it, but adding humor to positive content (e.g., an attractive ceived self-competence to the advertisers. To test this process, each photograph or happy memory) will make consumers less likely to participant in the study evaluated one ad (incongruous or congru- share it. ous) on the measures of incongruity, resolution, advertiser compe- Our first two studies tested this hypothesis using 2 (humor: hu- tence, self-perception of competence, and perceived self-advertiser morous, serious) × 2 (valence: positive, negative) between-subjects similarity. We then ran a serial mediation model where incongruity- experiments. In study 1a (N = 180), participants recalled either a and-resolution was the independent variable, advertiser competence positive or a negative memory. Orthogonally, they described this is the dependent variable, while self-perception of competence and memory in either a humorous or a serious tone. In study 1b (N = perceived self-advertiser similarity were the two mediators. The re- 299), participants found either an attractive or unattractive photo- sults confirmed significant serial mediation, and thus supported the graph. Orthogonally, they augmented the photograph using either a proposed process. humorous or serious filter. Consistent with our prediction, adding hu- Across a correlational study and an experiment, this research mor made participants more likely to share both negative memories documents that incongruity-resolution humor is the only humor (M = 2.13 vs. 1.42; p = .05) and unattractive photographs (M = 3.3 execution that enhances impression of firms’ competence and in- vs. 2.68; p = .079). Conversely, adding humor did not make partici- creases people’s likelihood to invest in the firms as a downstream pants more likely to share the happy story (M = 2.13 vs. 2.52; p <. consequence. The effects of the humor types on impressions of firms’ 28), and it made them less likely to share the attractive photograph warmth, however, turned out to be less robust and less likely to carry (M = 3.73 vs. 4.63; p = .007). over to subsequent behaviors such as perceived firms’ commitment Study 2 tested whether our focal effect would depend on wheth- to CSR. er participants planned to share the content in a public post or private message. If participants are more likely to share humorous negative Humor Makes Consumers More Likely to Share content because it helps them make a positive impression, then this Negative Content, But Not Positive Content effect should be stronger when they share the content with the general public, where making a good impression is more important, rather EXTENDED ABSTRACT than in a private message (Barasch and Berger 2014). The study (N Consumers regularly share content with one another online = 227) used a 2 (humor: humorous, serious) × 3 (emotion: embar- (Chen and Xie 2008; Wojnicki and Godes 2008). Firms benefit from rassment, sad, happy) between-subjects design while measuring in- this. Consumers who hear about a website or service from others are tention to share the content both in public and in a private message. more likely to visit the website (Godes and Mayzlin 2009). Increas- Participants recalled an event in which they felt either happy, sad, ingly, firms have been turning to a new tool to encourage consumers or embarrassed. Some of the participants described this event in a to share: comedy. In 2016, Taco Bell launched a humorous selfie humorous tone whereas others described it in a serious tone. Subse- lens that allowed consumers to superimpose a picture of their face quently, participants indicated their likelihood of sharing what they onto a taco shell. 224 million people shared the taco-face-lens. An- wrote as either a public post or a private message. We found that par- ecdotal examples aside, data on the relationship between comedy ticipants were more likely to share sad and embarrassing memories and sharing, does not consistently support the intuition that humor when they described them using a humorous rather than serious tone increases sharing. Although some studies suggest that consumers are (M = 3.41 vs. 2.14; p <. 01). Conversely, they were equally likely more likely to share humorous content (Tellis et al. 2019), others do to share the happy story regardless the tone (M = 3.33 vs. 3.23; p <. not (Bussiere 2009). 83). Humor had no effect on likelihood of sharing a private message Why and when are consumers more likely to share humor- (all effects NS). ous content over non-humors content? Consumers are more likely Study 3 further explored the process by asking undergraduate to share content that they think will make them look good (Berger students (N = 316) to complete a 2 (humor: humorous, serious) x 2014). Given this, the effect of humor on sharing should depend on 2 (valence: positive, negative) between-subjects experiment. Par- whether consumers believe that making their post humorous will ticipants were asked to imagine sharing a picture of either a suc- help them make a favorable impression. Generally, consumers be- cessful or a failed (burnt) beef brisket along with either a humorous Advances in Consumer Research (Volume 48) / 933

(“cooking was so fabulous even the smoke alarm was cheering me”) In the current research, we examine the relationship between or a serious caption (“my homemade brisket”). Participants also an- power and the use of humor using field data involving 73,620 corpo- swered how posting the image with the caption would make them rate email messages in addition to experimental studies that replicate look to others on a scale from 1 (terrible) to 7 (great). Consistent the key finding linking interpersonal power with the use of humor. with the previous studies, participants were more likely to share the This research offers three key theoretical contributions to research burnt brisket picture with humorous than a serious caption (M = 3.68 on social hierarchy, power, and humor. First, this work demonstrates vs. 2.90.; p = .03), but they were equally likely to share the success- that humor pervades communication, which makes understanding ful beef brisket regardless of whether they used a humorous or seri- humor necessary for understanding interpersonal interactions in ous caption (M = 4.37 vs. 4.08; p <. 43). The expected impression the social hierarchy. Second, this research introduces power as an mediated the effect humor on intention to share (ab = 1.19, CI [.418, important, but neglected, antecedent for the ability to use humor. 1.956]). Although humor has intrapsychic (e.g., stress reduction) and inter- In sum, whether humor increases the likelihood that consum- personal (e.g., ingratiation, elevated status) benefits which make users share memories, stories, and images online depends on whether ing it particularly attractive for low power individuals, we find that the content is positive or negative. Humor makes it easier to share occupying a position of low power makes individuals less able to use embarrassing, sad, and otherwise negative experiences, but it is does it. Third, we provide a novel explanation of why humor use varies not make consumers more likely to share happy, triumphant, or oth- across the hierarchy. We show that low power individuals experi- erwise positive experiences. ence greater cognitive load in interactions vis-à-vis a high power individual, which impedes their concentration and, therefore, their How Power Influences the Use of Humor ability to generate humor. Together, these contributions enrich our understanding of the interplay between power and humor and show EXTENDED ABSTRACT that the two constructs are intricately tied. Humor profoundly influences interpersonal perception and behavior. In hierarchical relationships, low power individuals have less Study 1: A Field Study of Power and Humor control over the situation and are more dependent upon individuals In Study 1, we examine the relationship between power and with greater power. This relative lack of control can lead to greater humor using field data; the Enron Email Corpus (Cohen 2015). We stress and anxiety (Dacher, Gruenfeld, and Anderson 2003) and can gathered hierarchical information on the individuals in the dataset motivate low power individuals to engage in ingratiatory behaviors and grouped them according to whether they fell at the bottom (low to elevate their position in the hierarchy (Judge and Bretz 1994; Ma- power) or top of the hierarchy (high power). Then, we examined how gee and Galinksy 2008; Watt 1993). Recent studies have found that the use of humor in emails related to the power level of the sender. the use of humor by low power individuals can be an effective method of emotional regulation and dealing with stress (Lefcourt and Results Martin 2012; Samson et al. 2014; Samson and Gross 2012), a means We find that humor is common in communication, occurring in 8.86% of sent emails. We found a significant effect of power on the of promoting liking and interpersonal bonding (Cooper, Kong, and 2 Crossley 2018), and a tool for elevating status (Bitterly et al. 2016). use of humor, χ (1) = 13.05, p < .001. We found that the use of humor Despite the stress reducing, ingratiation promoting, and status was significantly lower for low power individuals (7.04%) than high enhancing benefits of humor, low power individuals may be less power individuals (9.54%), β = .63, p < .001. The effect of power on likely to utilize humor (Blader and Chen 2012; Blader, Shirako, and the use of humor remained significant, β = .47, p = .025, when we Chen 2016; Galinksy, Rucker, and Magee 2015; Magee and Ga- control for the gender of the sender, β = .30, p = .153. linksy 2008)and thus significant questions remain about differences Study 2: Power and Humor in the Laboratory in how status and power impact social encounters. We conducted 5 We recruited 265 adults to participate in a behavioral lab study studies to address this gap. In particular, these studies tested the pre- in exchange for $10.00. We randomly assigned participants to one of diction that status and power would have opposing effects on justice two between-subjects conditions: High Power vs. Low Power. After enacted toward others. In the first 3 studies, we directly compared participants were assigned to their condition, we informed them that the effects of status and power on people’s enactment of distribu- they would be sharing jokes with their partner. We gave participants tive (Study 1. Several studies have shown that, compared to high 5 minutes to brainstorm jokes. Next, we gave participants 5 minutes power individuals, low power individuals dedicate more cognitive to write their final list of jokes, which served as our primary depen- resources attending to factors such as risks, potential threats, and the dent variable. thoughts and emotions of others (Blader et al. 2016; Dacher et al. 2003; Smith et al. 2008)prompting considerable research attention Results to this issue. However, prior research has primarily examined how We find a significant effect of the manipulation on the number power affects perspective taking, and has neglected to investigate of jokes that individuals wrote to share with their partner, F(1, 16) the impact of status (i.e., the respect and esteem that an individual = 16.93, p < .001, η2 = .04. Individuals assigned to the high power holds in the eyes of others, which causes low power individuals to condition (M = 3.40, SD = 1.56) created significantly more jokes to experience greater cognitive load. Therefore, low power individuals share with their partner than individuals assigned to the low power find themselves in a quagmire: Humor offers particularly important condition (M = 2.78, SD = 1.56), t(16) = 4.11, p < .001. benefits to individuals who lack power (to navigate the challenges of We find evidence that participants in the low power condition having low power and to gain power), but the lack of power could were less able to think of jokes. We find that cognitive load medi- make these same individuals less able to use humor. Little theory and ates the effect of power on the use of humor, but we do not find that research exist to shed light on the factors that influence whether low impression management concerns or confidence mediate the effect power individuals use humor, and the research that does exist pres- of power on humor. Using bootstrap analysis with 5000 simulations ents equivocal findings (Coser 1960; Lundberg 1969; Vinton 1989). (Preacher and Hayes 2004, 2008), we find a significant indirect effect for cognitive load (IE = .07, bias adjusted 95% confidence interval = 934 / Exploring the Effective Use of Humor in Consumer Contexts

[.01, .20]), but insignificant indirect effects for both impression man- Dacher, Keltner, Deborah. H. Gruenfeld, and Cameron Anderson agement concerns (IE = -.01, bias adjusted 95% confidence interval (2003), “Power, Approach and Inibition,” Psychological = [-.07, .01]) and confidence (IE = .00, bias adjusted 95% confidence Review, 110(2), 265–84. interval = [-.01, .07]). Didonato, Theresa E., Mellisha C. Bedminster, and Joanna J. Machel (2013), “My Funny Valentine: How Humor Styles Conclusion Affect Romantic Interest,”Personal Relationships, 20(2), In this work, we find that humor pervades communication and 374–90. is intrinsically connected to power. Although individuals at the bot- East, Robert, Kathy Hammond, and Malcolm Wright (2007), “The tom of the hierarch have much to gain from using humor, they are Relative Incidence of Positive and Negative Word of Mouth: A less able to use it. Thus, being powerless is a truly serious position Multi-Category Study,” International Journal of Research in REFERENCES Marketing, 24(2), 175–84. Eisend, Martin (2009), “A Meta-Analysis of Humor in Ace Metrix (2020), Super Bowl LIV Post-Game Ad Wrap-Up. Advertising,” Journal of the Academy of Marketing Science, Ames, Daniel R. (2004), “Strategies for Social Inference: A 37(2), 191–203. Similarity Contingency Model of Projection and Stereotyping Escalas, Jennifer Edson and James R Bettman (2015), “Managing in Attribute Prevalence Estimates,” Journal of Personality and Brand Meaning through Celebrity Endorsement,” European Social Psychology, 87(5), 573–85. Journal of Marketing, 4512(11), 882–909. Barasch, Alixandra and Jonah A Berger (2014), “Broadcasting Flamson, Thomas and H. Clark Barrett (2008), “The Encryption and Narrowcasting: How Audience Size Affects What People Theory of Humor: A Knowledge-Based Mechanism of Honest Share,” Journal of Marketing Research, 51(3), 286–99, http:// Signaling,” Journal of Evolutionary Psychology, 6(4), 261–81. journals.ama.org/doi/abs/10.1509/jmr.13.0238. Fraley, Barbara and Arthur Aron (2004), “The Effect of a Shared Berger, Jonah A (2014), “Word of Mouth and Interpersonal Humorous Experience on Closeness in Initial Encounters,” Communication: A Review and Directions for Future Personal Relationships, 11(1), 61–78. Research.,” Journal of Consumer Psychology (Elsevier Galinksy, Adam D., Derek D Rucker, and Joe C Magee (2015), Science), 24(4), 586–607. “Power: Past Findings, Present Considerations, and Future Bitterly, T. Bradford and Maurice E. Schweitzer (2019), “The Directions,” in APA Handbook of Personality and Social Impression Management Benefits of Humorous Self- Psychology, Vol. 3. Interpersonal Relations, Washington, DC: Disclosures: How Humor Influences Perceptions of American Psychological Association, 421–60. Veracity,” Organizational Behavior and Human Decision Godes, David and Dina Mayzlin (2009), “Firm-Created Word- Processes, 151(January), 73–89, https://doi.org/10.1016/j. of-Mouth Communication: Evidence from a Field Test,” obhdp.2019.01.005. Marketing Science, 28(4), 721–39. Bitterly, T Bradford, Alison Wood Brooks, and Maurice E Greengross, Gil, Rod a. Martin, and Geoffrey Miller (2012), Schweitzer (2016), “Risky Business: When Humor Increases “Personality Traits, Intelligence, Humor Styles, and Humor and Decreases Status,” Journal of Personality and Social Production Ability of Professional Stand-up Comedians Psychology, 112(3), 431–55. Compared to College Students.,” Psychology of Aesthetics, Blader, Steven L. and Ya Ru Chen (2012), “Differentiating the Creativity, and the Arts, 6(1), 74–82. Effects of Status and Power: A Justice Perspective,”Journal of Greengross, Gil and Geoffrey F. Miller (2008), “Dissing Oneself Personality and Social Psychology, 102(5), 994–1014. versus Dissing Rivals: Effects of Status, Personality, and Sex Blader, Steven L., Aiwa Shirako, and Ya Ru Chen (2016), “Looking on the Short-Term and Long-Term Attractiveness of Self- Out From the Top: Differential Effects of Status and Power Deprecating and Other-Deprecating Humor,” Evolutionary on Perspective Taking,” Personality and Social Psychology Psychology, 6(3), 393–408. Bulletin, 42(6), 723–37. Hall, Jeffrey A. and Ken Sereno (2010), “Offensive Jokes: How Burro, Roberto, Carla Canestrari, Ugo Savardi, Erika Branchini, Do They Impact Long-Term Relationships?,” Humor, 23(3), and Ivana Bianchi (2018), “Pleasures of the Mind: What 351–73. Makes Jokes and Insight Problems Enjoyable,” Frontiers in Judge, Timothy A. and Robert D. Bretz (1994), “Political Influence Psychology, 8(January), 1–21. Behavior and Career Success,” Journal of Management, 20(1), Bussiere, Dave (2009), “The Effects of Humor on the Processing 43–65. of Word-Of-Mouth,” in Advances in Consumer Research, ed. Kervyn, Nicolas, Susan T. Fiske, and Chris Malone (2012), “Brands Ann L. McGill and Sharon Shavitt, Duluth, MN, 399–404. as Intentional Agents Framework: How Perceived Intentions Chen, Yubo and Jinhong Xie (2008), “Online Consumer and Ability Can Map Brand Perception,” Journal of Consumer Review: Word-of-Mouth as a New Element of Marketing Psychology, 22(2), 166–76. Communication Mix,” Management Science, 54(3), 477–91. Lefcourt, Herbert M and Rod A Martin (2012), Humor and Life Cohen, William W (2015), “Enron Email Dataset.” Stress: Antidote to Adversity, Springer Science & Business Cooper, Cecily D, Dejun Tony Kong, and Craig D Crossley (2018), Media. “Leader Humor As an Interpersonal Resource: Integrating Lundberg, Craig (1969), “Person-Focused Joking: Pattern and Three Theoretical Perspectives,” Academy of Management Function,” Human Organization, 28(1), 22–28. Journal, 61(2), 769–96. Magee, Joe C and Adam D. Galinksy (2008), “Social Hierarchy: Coser, Rose Laub (1960), “Laughter among Colleagues. A Study The Self-Reinforcing Nature of Power and Status,” Academy of the Social Functions of Humor among the Staff of a Mental of Management Annals, 2(1), 351–98. Hospital.,” Psychiatry, 23, 81–95. Martin, Rod A. (2007), “The Social Psychology of Humor,” The Craik, Kenneth H., Martin D. Lampert, and Arvalea J. Nelson Psychology of Humor, 113–52. (1996), “Sense of Humor and Styles of Everyday Humorous Conduct,” Humor, 9(3), 273–302. Advances in Consumer Research (Volume 48) / 935

Martin, Rod A (2010), The Psychology of Humor: An Integrative Samson, Andrea C. and James J. Gross (2012), “Humour as Approach, Academic Press. Emotion Regulation: The Differential Consequences of Martin, Rod A, Patricia Puhlik-Doris, Gwen Larsen, Jeanette Negative versus Positive Humour,” Cognition and Emotion, Gray, and Kelly Weir (2003), “Individual Differences in Uses 26(2), 375–84. of Humor and Their Relation to Psychological Well-Being: Smith, Pamela K., Nils B. Jostmann, Adam D. Galinsky, and Wilco Development of the Humor Styles Questionnaire,” Journal of W. Van Dijk (2008), “Lacking Power Impairs Executive Research in Personality, 37, 48–75. Functions: Research Article,” Psychological Science, 19(5), Matz, Sandra C, Cristina Segalin, David Stillwell, Sandrine R. 441–47. Müller, and Maarten W. Bos (2019), “Predicting the Personal Speck, Surgi Paul (1991), “The Humorous Message Taxonomy: A Appeal of Marketing Images Using Computational Methods,” Framework for the Study of Humorous Ads,” Current Issues Journal of Consumer Psychology, 29, 370–90. and Research in Advertising, 13(1–2), 1–44. McGraw, A. Peter, Julie Schiro, and Philip Fernbach (2015), “Not Spotts, Harlan E., Marc G. Weinberger, and Amy L. Parsons (1997), a Problem: A Downside of Humorous Appeals,” Journal of “Assessing the Use and Impact of Humor on Advertising Marketing Behavior, 1, 187–208. Effectiveness: A Contingency Approach,” Journal of McGraw, A. Peter and Caleb Warren (2010), “Benign Violations: Advertising, 26(3), 17–32. Making Immoral Behavior Funny,” Psychological Science, Tellis, Gerard J., Deborah J. MacInnis, Seshadri Tirunillai, and 21(8), 1141–49. Yanwei Zhang (2019), “What Drives Virality (Sharing) of McGraw, A. Peter, Caleb Warren, and Christina Kan (2013), Online Digital Content? The Critical Role of Information, “Humorous Complaining,” Journal of Consumer Research, Emotion, and Brand Prominence,” Journal of Marketing, 35(2), 268–78. 83(4), 1–20. McQuarrie, Edward F. and David Glen Mick (1996), “Figures of Vinton, Karen L. (1989), “Humor in the Workplace: It Is More Rhetoric in Advertising Language,” Journal of Consumer Than Telling Jokes,” Small Group Research, 20(2), 151–66. Research, 22(4), 424. Warren, Caleb, Adam Barsky, and A. Peter McGraw (2018), Preacher, Kristopher J and Andrew F Hayes (2008), “Assessing “Humor, Comedy, and Consumer Behavior,” Journal of Mediation in Communication Research,” in The Sage Consumer Research, 45(March), 1–74. Sourcebook of Advanced Data Analysis Methods for Warren, Caleb and A. Peter McGraw (2016a), “Differentiating Communication Research, London, UK: Sage Publications, What Is Humorous from What Is Not,” Journal of Personality 13–54. and Social Psychology, 110(3), 407–30. ——— (2004), “SPSS and SAS Procedures for Estimating Indirect ——— (2016b), “When Does Humorous Marketing Hurt Brands?,” Effects in Simple Mediation Models,”Behavior Research Journal of Marketing Behavior, 2(1), 39–67. Methods, Instruments, & Computers, 36(4), 717–31. Watt, John D (1993), “The Impact of the Frequency of Ingratiation Samson, Andrea C., Alana L. Glassco, Ihno A. Lee, and James J. on the Performance Evaluation of Bank Personnel,” Journal of Gross (2014), “Humorous Coping and Serious Reappraisal: psychology, 127(2), 117–25. Short-Term and Longer-Term Effects,”Europe’s Journal of Wojnicki, Andrea C and David Godes (2008), Word of Mouth as Psychology, 10(3), 571–81. Self-Enhancement. Toward a Further Understanding of How Political Ideology Drives Motivation and Values Chairs: Der-Wei Huang, Indiana University, USA Adam Duhachek, University of Illinois at Chicago, USA

Paper #1: Are Conservatives Always More Averse to Change performing a task. The authors investigate performance on tasks as than Liberals? Political Ideology and Innovation a function of the variability of the attentional foci associated with Der-Wei Huang, Indiana University, USA task performance. They find that liberals perform better on tasks that Adam Duhachek, University of Illinois at Chicago, USA require varying attentional foci whereas conservatives perform better at tasks requiring a stable attentional focus. Variable attention is Paper #2: The Impact of Political Ideology on Self-Control useful in contexts where consumers need to learn new skills whereas Bryan M. Buechner, Xavier University, USA stable and constant attention is useful in contexts where consumers Joshua J. Clarkson, University of Cincinnati, USA need perseverance, such as goal striving and habit formation. Ashley S. Otto, Baylor University, USA The third paper examines how ideology influences reactions to Edward R. Hirt, Indiana University, USA service failure and implicates the role of moral violations related to Ming-Shen Ho, Clemson University, USA fairness. Their findings show that liberal consumers are likely to re- Paper #3: Fair or not? Political Ideology Shapers Observers’ spond more negatively to brands involved in a service failure due to Responses to Service Failures their greater sensitivity to violations of the fairness moral foundation. Chethana Achar, Northwestern University, USA They find liberals were more likely to express negative emotions Nidhi Agrawal, University of Washington, USA on Twitter in response to a well-publicized airline service failure. Paper #4: Political Polarization in How Perceived Social They also find convergent laboratory support using service failure Similarity Impacts Support for Redistribution vignettes and show that these ideological differences are driven by Nailya Ordabayeva, Boston College, USA differential activation of the fairness moral foundation in response to the failure. SESSION OVERVIEW The fourth paper shows that ideology drives different justifica- Emerging research on political ideology has shown that conser- tion of unequal outcomes depending on perceived social similarity. vatives’ and liberals’ motivation and values affect their preferences They found that perceiving high (vs. low) social similarity increases in the marketplace. The theoretical foundations for these ideological redistribution support among liberals, but decreases redistribution differences have related to distinct motivational and value systems support among conservatives, because it weakens liberals’, but that distinguish liberals and conservatives. One such difference that strengthens conservatives’, belief that individuals deserve their un- has garnered significant theoretical and empirical support relates to equal outcomes. openness and preference for change, with numerous empirical stud- This session presents research that together examine how con- ies showing liberals are more open to and have an innate relative servatives’ and liberals’ preferences related to key motivations and preference for change as compared to conservatives. These change moral values influence a variety of consumer behaviors. The papers preferences have manifest in a number of different domains from build theory linking political ideology and behavior and share a com- support for social and government institutions, political attitudes and mon focus on underlying motivational and value-based differences consumer preferences. The current session also builds on another as a function of ideology. cornerstone of theory related to political ideology by examining differences in the justification of economic inequality of liberals and Are Conservatives Always More Averse to Change than conservatives’ behaviors. Previous research has shown that perceiv- Liberals?Political Ideology and Innovation ing high (vs. low) similarity can potentially increase support for redistribution, because it may lead individuals to question the fairness EXTENDED ABSTRACT of people’s unequal outcomes and boost their group loyalty. Research Previous research has shown that political ideology is at the has shown that this perceived similarity shapes opinions related to heart of a variety of cognitive and motivational differences. One support for social programs and the role of government assistance. of the most frequently reported distinctions relates to openness to The current session advances extant understanding of these issues change. Research has shown liberals possess a greater tolerance to by revealing that political ideology’s influence on these motivations change and are more open to new experiences. Previous consumer and values leads to wide-ranging effects on mainstream consumer research has shown how these underlying ideological influences af- behavior research areas, such as service failure. fect consumption. For instance, Khan et al. (2013) find that conserva- The first paper shows that conservatives and liberals- experi tive consumers were less likely to adopt new products as compared to ence change differently. In the context of new product innovations, liberal consumers. Barra (2014) finds that this effect on new products conservatives’ greater sensitivity toward change may lead them to is enhanced when products are more novel and when consumer’s believe a new product innovation is more significant and beneficial political ideology is more central to their self-concept. as compared to liberals. The authors also show that conservatives The current research theorizes that not all change is equivalent and liberals will assimilate or contrast their evaluations of new prod- and that liberals and conservatives may experience change differ- ucts when making comparisons between past and future products. al- ently. One hypothesis we propose is that conservatives’ greater sen- though conservatives are more likely to perceive innovation-related sitivity to change may have positive consequences in the context of changes in secondary product features as compared to liberals who new product innovation. Specifically, we theorize that new product focus more on primary product features, and how ideology affects innovations may be seen as more significant and beneficial to conser- evaluation of innovation on the time frame. vatives as their greater sensitivity to change makes these differences The second paper sheds light on the cognitive processes enact- seem more consequential as compared to liberals. Additionally, we ed by liberals and conservatives as they experience changes while theorize that a greater sensitivity to changes in product features may

Advances in Consumer Research 936 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 937 lead to conservatives’ and liberals’ different evaluations of abrupt ondary) between subject design, where their political ideology was and gradual product innovations. measured. Our dependent measure was perceived change, and the The literature on product innovation has identified speed of result indicated that conservatives (vs. liberals) notice the change change as a key underlying dimension along which innovations are in secondary features more (p = 0.0061). Primary and secondary evaluated. Product innovations can happen abruptly or they can oc- features were determined by a pilot study where one hundred par- cur more gradually (Garcia & Calantone, 2002) Based on this no- ticipants were recruited to rate which features of a smartphone are tion, we define abrupt changes in innovation as changes that enter the primary and secondary. In Study 2, we replicated our results in Study market immediately, and alter technological deliveries of products; 1, where we show that conservatives notice small changes more than whereas incremental changes in innovation are changes that take liberals. Moreover, the two-way interaction between political ideol- place gradually and in a less significant way. We theorize that conser- ogy and feature type still remains significant after controlling for vatives will prefer gradual changes as compared to liberals, whose brand liking and product liking (p = 0.0197). inherent preferences for change will lead them to prefer abrupt prod- Thus, Study 1 & 2 showed that depending on consumer’s po- uct innovations. litical ideology, different types of changes in features will change Further, consumers perceive change differently when compar- consumer’s perceptions of the innovations. Namely, we show that ing to the past or the future. We theorize that conservatives will per- conservatives are sensitive to small changes more than liberals do. ceive changes in products similarly in the past and future, whereas Study 3: In Study 3, we examine how different types of inno- liberals generate more positive attitude toward changes in the fu- vations will affect consumer’s attitude toward innovations. We ran- ture (vs. past), due to their openness to change and belief that future domly assigned 200 participants to one of the conditions in a 2 (in- change will eclipse past change as part of a progressive belief sys- novation type: abrupt, gradual) between subject design, where their tem. We theorize that the differential evaluations of past and future political ideology was measured. The dependent measures in this product innovations is due to distinct assimilation and contrast ef- study are likelihood of early adoption, and innovation liking. Par- fects activated as a function of political ideology. ticipants read a passage about an update of their operating system, As denoted by a large body of consumer research, consumer which is either an abrupt or gradual innovation. As hypothesized, a judgments and behaviors are influenced by the context they are in two-way interaction shows that conservatives (vs. liberals) are more (Bless and Schwarz 2010). Past research on assimilation and con- likely to engage in early adoption of the new operating system in the trast shows that contextual influences can alter target judgments and gradual innovation condition (vs. abrupt) (p = 0.0489), and they like related behaviors. It is well-documented that product information, the innovation more in the gradual (vs. abrupt) innovation condition whether being numeric (e.g. price) or semantic (e.g. vice or virtue (p = 0.0218). products), learned before to evaluating a product can affect consumer Thus, Study 3 showed that when the changes are gradual, con- judgment (Chernev 2011; Makens 1965; Plassmann et al. 2008). An servatives, instead of being resistant to changes as suggested in prior assimilation effect typically takes place when consumers integrate research, they like the gradual innovation more than liberals, and are prior product information to their judgment of a subsequent product. willing to engage in early adoption. However, it is not always the case that consumers assimilate Study 4: In Study 4, 300 participants were randomly assigned to prior product evaluation with subsequent products. Sometimes, con- one of the conditions in a 2 (innovation type: innovative, non-inno- sumers evaluate the same product more negatively when the product vative) mixed subject design, where each participant rated their per- information is favorable compared to when it is unfavorable, which ceptions of each product by comparing them to the past and future. is known as the contrast effect. The dependent measures composed of items that asked participants Conservatives has been shown in psychology and consumer re- about their perceptions of change of eight types of products (four search to be referencing the past to a greater extent than future (Rob- innovative, four non-innovative). Participants first rated the products inson et al., 2015). More recently, Lammers and Baldwin (2018) by comparing the products’ performance ten years in the past to pres- shows that conservatives are prone to past and nostalgic society, ent performance, and then they rated the same products by estimat- resulting in their support for political ideas that can be linked to a ing their future performance in ten years compared to present per- desirable past state, rather than a desirable future state of society. formance. As a mixed model analysis shows, a significant three-way Building on their findings, the current research examines how their interaction ( p = 0.0004) shows that liberals evaluate product perfor- characteristic of past/ future-orientation affect how consumers with mance in the past (vs. future) more negatively, and that conservatives different political ideology evaluate past/ new products differently. evaluate performance similarly when comparing the current product Specifically, we show that the past and nostalgic taste of a product to the past and future. Specifically, we found that this significance leads to conservatives’ assimilation effect of converging their evalu- holds only for innovative products (p < 0.0001). This effect does not ations of past and new products. hold when the product is non-innovative (p = 0.8689). Study 1: This study was conducted around the time that iPhone Thus, Study 4 shows that liberals and conservatives hold differ- X was launched, and we recruited 541 Mturk participants. Partic- ent beliefs about the rate of change of innovative products. Liberals ipants were asked of the perceived change of features of the new believe that products will be better in the future as compared to the iPhone on a 7-point Likert scale. Perceived change of the features are past whereas conservatives do not show different beliefs regarding converted into one index, which is our dependent measure. Sample past or future product performance. The results align with our pre- features include, “battery life and charging”, and “camera”. Political dictions that conservatives assimilate product evaluations from the ideology is measured on an 11-point Likert scale adopted from Jost past and future, whereas liberals contrast their evaluations of prod- (2007). The result shows that conservatives (vs liberals) notice more ucts from the past and future. change than the liberals (p = 0.0162). Specifically, as the features Taken together, these studies show that liberals and conserva- measured in this study are mostly secondary features. tives different sensitivities to and beliefs regarding change influence Study 2: In Study 2, we further examine if conservatives pre- their evaluations of product innovations. In contrast to the received fer changes that are secondary We randomly assigned 202 Mturk view that conservatives always evaluate change more negatively than participants to one of the conditions of a 2 (features: primary, sec- 938 / Toward a Further Understanding of How Political Ideology Drives Motivation and Values liberals, we find evidence for more nuanced relationships among po- ers’ political ideology and the extent to which the task requires inhi- litical ideology, innovation and consumer evaluation. bition or updating. A subsequent study tested the extent to which this effect of task The Impact of Political Ideology on Self-Control performance was isolated to consumers’ political ideology within a separate context. Here, participants completed two anagrams where EXTENDED ABSTRACT the performance rule was either fixed (inhibition) or changed (up- Self-regulation is a multi-process system (Engle, Tuholski, dating). For instance, participants instructed to list only three-letter Laughlin, and Conway 1999; Miyake et al. 2000; see Laran 2020), solutions for the first anagram were then instructed to either list such that effective self-control often requires consumers to focus at- three-letter (inhibition) or four-letter (updating) solutions in the sec- tention in a way that suppresses the influence of interfering informa- ond anagram. To isolate ideology, participants completed an assess- tion (i.e., inhibition) as well as incorporates the influence of novel in- ment of intelligence using an adapted version of the Wechsler Adult sights (i.e., updating). The question motivating the present research Intelligence Scale (WAIS; Friedman et al. 2006; Wechsler 1997), is whether consumers’ political ideology influences self-control suc- motivation (Muraven and Slessareva 2003), and religiosity (Koenig cess on tasks of inhibition and updating. and Büssing 2010) before indicating their demographics and politi- Political ideology reflects a specific set of ethical ideals, prin- cal ideology. ciples, and doctrines that explain the basis by which society should The hierarchical regression controlled for participant demo- function (Jost, Federico, and Napier 2009). Though indirect, a wealth graphics as well as intelligence, motivation, and religiosity. Analysis of research suggests that conservatives and liberals vary in their abil- of the number of correct solutions across anagrams revealed a sig- ity to mentally adapt to situational changes (i.e., cognitive flexibility: nificant interaction (p = .002); conservatives generated more solu- Zmigrod, Rentfrow, and Robbins, 2018). Specifically, liberals appear tions than liberals when the performance rule was fixed (p = .02) to be cognitively-flexible; they are open to new experiences (Carney, whereas liberals generated more solutions than conservatives when Jost, Gosling, and Potter 2008), socially-adaptable (Hirsh, DeYoung, the performance rule was varied (p = .04). Interestingly, intelligence, Xu, and Peterson 2010), and novelty-seeking (Carney et al. 2008). motivation, and religiosity each had a positive effect on performance Conversely, conservatives appear to be cognitively-rigid; they are (p’s < .01) yet had no influence on the relationship between ideology persistent at impulse regulation (Clarkson et al. 2015), norm-adher- and task performance. ent (Gosling, Rentfrow, and Swann 2003), and resistant to change A final study examined the mediating role of cognitive flex- (Jost et al. 2009). ibility on the performance of conservative and liberal consumers on This potential difference in cognitive flexibility is critical as tasks of differing attentional demands. Participants completed two it is proposed to impact performance on self-control tasks that re- 14x14 letter word searches where the performance rule was either quire inhibition and updating. Specifically, the traits that reinforce fixed or varied. For instance, participants instructed to identify only cognitive rigidity in conservatives (e.g., rule adherence, resistance forward solutions for the first word search were then instructed to ei- to change) should promote response inhibition by facilitating per- ther list forward (inhibition) or backward (updating) solutions for the sistence through suppression of conflicting stimuli. Conversely, the second word search. Participants again completed the assessments traits that reinforce cognitive flexibility in liberals (e.g., openness, of intelligence, motivation, and religiosity, as well as a measure of adaptability) should promote response updating by facilitating adap- cognitive flexibility (Martin and Rubin 1995), before indicating their tation through replacing outdated information with relevant informa- demographics and political ideology. tion. Collectively, then, conservatives and liberals are hypothesized Again controlling for demographics, intelligence, motivation, to perform well at tasks of inhibition and updating, respectively, due and religiosity, analysis of the number of correct solutions revealed to their respective difference in cognitive flexibility. a significant interaction (p = .001); conservatives identified more This hypothesis was tested across three studies. Each study tar- words than liberals when the performance rule was fixed (p = .02), geted 200 participants based on a priori power analyses (power of whereas liberals identified more words than conservatives when the .8, small-medium effect sizes, an alpha level of .05; Faul et al. 2007) performance rule was varied (p = .005). Moreover, there was a direct and included appropriate attention checks (Oppenheimer, Meyvis, effect of ideology on cognitive flexibility (p = .002). Critically, the and Davidenko 2009). moderated-mediation test of cognitive flexibility (Model 15: Hayes As an initial test of our hypotheses, we conducted two studies at 2018) indicated significant mediation through both the fixed (95% separate time points to test our primary hypothesis that conservative CI: .003, .182) and varied (95% CI: -.171, -.005) rule conditions. and liberal consumers differentially excel on tasks shown to assess Collectively, these findings support the claim that the effects inhibition and updating (Conway et al. 2005; Engle et al. 1999; Mi- of political ideology on self-control might be more nuanced than yake et al. 2000). In the first session, participants completed a task originally theorized. Specifically, conservatives may better regulate of inhibition that presents participants with a target cue to respond behaviors that require cognitive rigidity, whereas liberals may better to as quickly as possible while ignoring a random cue (IOR; Posner regulate behaviors that require cognitive flexibility. Consequently, and Cohen 1984). In a separate session, participants completed a task this work holds direct implication for consumers’ self-control be- of updating that required remembering the most recent category- haviors, such as financial decision-making (e.g., saving money) or relevant stimulus among constantly changing stimuli from multiple addictive consumption (e.g., overeating). categories (Keeping Track Task; Yntema and Mueser, 1962). Partici- pants in both studies indicated demographics along with a measure Fair or Not? Political Ideology Shapes Observers’ of political ideology on a 7-item scale anchored from Very liberal to Responses to Service Failures Very conservative (Jost 2006; Knight 1999). The analysis revealed that conservatives outperformed liberals on the inhibition task (IOR: EXTENDED ABSTRACT p = .02), whereas liberals outperformed conservatives on the updat- Consumers frequently witness service failure incidents. While ing task (Keeping Track Task: p = .03). This finding is consistent having dinner at a restaurant, one might see that the customers at the with the hypothesis that self-control varies as a function of consum- next table did not receive the drinks they ordered. Or while checking Advances in Consumer Research (Volume 48) / 939 in at a hotel, one might notice another customer complaining that using the fairness/reciprocity subscale from the Moral Foundations he did not get the type of room he reserved. Those who witness or Questionnaire (MFQ; Graham, Haidt, & Nosek, 2008). One week learn about service failures could potentially form negative attitudes later, these participants were invited to an ostensibly unrelated study about the brand, which in turn can affect their intentions and WOM (Phase 2), with 287 returning participants (Mage = 35.34 years, 140 related to the brand. Firms would be well-served by understanding females). They were presented three scenarios (hotel, restaurant, and which consumers are more likely to react negatively to witness- airlines setting) in randomized order, each followed by a page of ing service failures, why, and how to curtail the negative impact of questions. Scenarios were based on past literature (Hess et al., 2003; service failure for such consumers. This paper presents consumers’ Smith et al., 1999) and adapted to a third-person point of view such political ideology as a determinant of their reactions to observing that participants saw another customer experiencing service failure. service failures. Following each scenario, we measured participants attitudes toward Although service failures are usually conceptualized in terms of the service provider and how likely they would be to positively re- operational failures in the process of delivering service, they could view the service provider. Participants’ unique ID was used to match also be thought of as breaking a transactional promise (e.g., Goodwin data across two phases. Our data show that the first regression model & Ross, 1992; Smith, Bolton, & Wagner, 1999). This latter concep- is significant in the hotel (F(1, 277) = 7.33, p = .007, adjusted R2 = tualization recognizes that a customer experiencing service failure .026), restaurant (F(1, 277) = 6.37, p = .012, adjusted R2 = .02), and could feel “wronged” and thus form unfavorable judgments (Good- airlines scenarios (F(1, 277) = 12.15, p = .001, adjusted R2 = .04). win & Ross, 1992). To those who observe the failure but are not Specifically, political ideology significantly predicted the variance affected by it, a broken promise or one party being wronged could in attitude toward the hotel (b = .17, p = .007), restaurant (b = .15, represent a larger issue of fairness. Fairness considerations—includ- p = .012), and airline (b = .20, p = .001) such that liberals (vs. con- ing those of fairness violations—are one of five moral intuition sys- servatives) held less positive attitudes in response to reading about tems or “foundations” that form the basis of moral judgments (Moral service failure experienced by others. These results held even when Foundations Theory, MFT; Haidt & Joseph, 2004, 2007a). Building the demographic variables were controlled for and were replicated on this theoretical framework, past research shows that the impor- with the positive reviewing DV. We further found that these effects tance of the fairness moral foundation varies by political ideology, are mediated by the endorsement of the fairness foundation and per- such that liberals, relative to conservatives, consider fairness to be a ceptions of fairness toward the affected customer. highly relevant attribute in judging whether something is “right” or Studies 4a and 4b used surveys about news items about pub- “wrong”. licized service failure events involving the Fyre music festival and Because failures undermine the perceived fairness of a transac- the beforementioned United Airlines incident, respectively. In both tion, we suggest that liberals (vs. conservatives) are likely to view studies, participants’ self-identified political ideology and political service failures as instances of moral violation. Therefore, even non- party identification predicted their attitudes toward the two service political instances of service failures could take on a political aspect providers and the affected customers. In study 5, the divergence in and invoke divergent responses from observers who are political lib- liberals’ (vs. conservatives’) responses was attenuated when the re- erals or conservatives. In response to witnessing a service failure, covery effort was referenced as “fair” compensation (vs. compensa- liberals will likely see the involved brand as a violator of fairness, tion present vs. compensation absent). These findings present multi- an important moral foundation to them (along with harm/care), lead- method evidence of politico-moral influences in consumer responses ing to harsher, more negative responses. Conversely, since conserva- to non-political service failures. tives do not place disproportionate value on fairness but also value respecting authority/law, cooperation, and self-sacrifice (Graham et Political Polarization in How Perceived Social Similarity al., 2009; Haidt, Graham, & Joseph, 2009), they will be less likely to Impacts Support for Redistribution see the service failure as a moral violation and thus be less harsh in their judgments of the brand. EXTENDED ABSTRACT We use a multi-method approach (3 experiments, 2 surveys, and The inequality of wealth in the US has reached record levels 1 big data analysis) to test our theory. As study 1, we gathered a large (Piketty, 2011). This is generating a public debate about the impact dataset of tweets (n = 219,070) from Twitter in a quasi-experimental of inequality on individual and collective wellbeing and increasingly field study using consumers’ actual social media activity surround- shaping people’s preferences across many contexts including politi- ing a publicized service failure incident about United Airlines. We cal candidates, policies, as well as products and brands (e.g., Daw- estimated users’ political orientation using a ranking algorithm and try, Sutton, & Sibley, 2015; Frank, Wertenbroch, & Maddux, 2015; conducted sentiment analysis of tweets utilizing a dictionary-based Ordabayeva & Chandon, 2011). text analyzer. Analyses show that immediately after the negative- Although individuals agree that economic inequality in the US publicity incident, liberal (vs. conservative) Twitter users were more is historically high (Bartels, 2005), support for redistributive poli- likely to express negative sentiments, especially anger and anxiety, cies designed to reduce income inequality is not very widespread against United Airlines. A control sample of tweets from before the (Jost & Hunyady, 2005). This raises the question of how individuals’ event showed a null effect of political ideology on sentiment. preferences and support for redistribution can be changed (Brown- Studies 2 and 3 use experiments and controlled scenarios to rep- Iannuzzi et al., 2014). licate the Twitter findings with stronger internal validity. As study Prior work suggests that, although redistributive preferences 2, participants were assigned to conditions in a 2 (political ideol- are deeply ingrained within individuals, one factor that can poten- ogy: liberal vs. conservative) × 3 (scenario: hotel vs. restaurant vs. tially overcome this challenge and increase individuals’ support for airline; within-subject) mixed design implemented in two phases. redistributive policies is similarity. Specifically, prior work suggests In Phase 1, we recruited 350 adult US participants. We embedded perceiving high (vs. low) similarity can potentially increase support two questions regarding participants’ political orientation (following for redistribution, because it may lead individuals to question the Hirsch et al., 2015; Jost, Nosek, & Gosling, 2008). We also mea- fairness of people’s unequal outcomes and boost their group loy- sured participants’ endorsement of the fairness moral foundation by alty (Luttmer, 2001). Supporting this notion, studies show that in- 940 / Toward a Further Understanding of How Political Ideology Drives Motivation and Values dividuals are more supportive of redistribution if they think that it high (vs. low) social similarity weakened liberals’, but strengthened will help similar others (Duell, 2015), if they live in areas with high conservatives’, belief in the fairness of unequal economic outcomes racial similarity (Alesina, Baqir, & Easterly, 1999), and if they are (interaction: b = .03, p = .001; moderated mediation: a x b = -.0213, prompted to cognitively focus on similarity in a picture comparison 95% CI = [-.0387, -.0081]). task (Ordabayeva & Fernandes, 2017). Study 3 manipulated perceptions of social similarity among The present research suggests that the effect of perceived social others (i.e. Americans – to minimize the role of egocentric and pro- similarity on individuals’ redistributive preferences may be more jection motives), and it included a control condition in which simi- complex. It suggests that perceiving high (vs. low) social similar- larity perceptions were not manipulated. Two dummies were created ity may increase redistribution support among liberals, but decrease to test the predictions (Naylor, Lamberton, & Norton 2011): dummy redistribution support among conservatives, because it may weaken 1 contrasting high similarity with low similarity and control, and liberals’, but strengthen conservatives’, belief that individuals de- dummy 2 contrasting low similarity with control; but only dummy 1 serve their unequal outcomes. More specifically, since prompting interacted with ideology (b = -.17, p = .003). Liberals’ redistributive high social similarity would confirm liberals’ assumption that in- preferences were higher in the high similarity than in the low similar- dividuals are similar in their inherent qualities (hard work, effort), ity and control conditions, but conservatives’ redistributive prefer- it should strengthen liberals’ belief that similar individuals deserve ences were lower in the high similarity than in the low similarity and similar outcomes and boost liberals’ support for redistribution. In control conditions. These findings confirmed that the effect is driven contrast, boosting perceptions of social similarity may challenge by perceptions of social similarity (rather than dissimilarity). conservatives’ assumption that people are dissimilar in inherent This research suggests that individuals’ preferences for redis- qualities, which may lead conservatives to wish to reconcile why tributive policies designed to reduce inequality may be more com- similar individuals end up with dissimilar outcomes. The desire to plex than previously presumed, and that these preferences may be reconcile this discrepancy may further strengthen conservatives’ shaped by external manipulations differently depending on individu- belief that people are ultimately responsible for and deserve their als’ group membership. The results may also inform policy makers dissimilar outcomes, because unsuccessful individuals had similar interested in shaping public support for redistribution as well as starting ingredients needed for success as successful individuals and researchers interested in identifying prompts that can change redis- hence unsuccessful individuals have no one but themselves to blame. tributive preferences and factors that can polarize, or unify, conser- In effect, this research proposes that prompting perceptions of high vatives’ and liberals’ preferences in various contexts. (vs. low) social similarity may result in greater (rather than lower) political polarization in redistributive preferences. Three studies REFERENCES tested this prediction. Alesina, Alberto, Reza Baqir, and William Easterly (1999), “Public Study 1 manipulated perceptions of social similarity: partici- Goods and Ethnic Divisions,” The Quarterly Journal of pants elaborated on the similarities or differences between them- Economics, 114 (4), 1243-1284. selves and others. As the dependent measure in Study 1, and in all Barra, Cristobal “When Do Political Ideologies Affect Brand subsequent studies, participants indicated the extent to which they Extension Evaluation? The Role of Analytic Versus Holistic support or oppose taxes on individuals earning more than $1 million Mindsets,” 126. and more than $5 million; Chow & Galak, 2012). Participants also Bless, Herbert and Norbert Schwarz (2010), “Mental Construal indicated their beliefs in the fairness of unequal outcomes (e.g., “By and the Emergence of Assimilation and Contrast Effects,” and large, people deserve what they get”; Rubin & Peplau, 1975) in Advances in Experimental Social Psychology, Elsevier, and their political ideology (1 = “extremely liberal” to 9 = “extreme- 319–73, https://linkinghub.elsevier.com/retrieve/pii/ ly conservative”; Jost, 2006). The results of a regression analysis S0065260110420067. revealed a significant interaction between social similarity and po- Brown-Iannuzzi, Jazmin L., Kristjen B. Lundberg, Aaron C. Kay, litical ideology (b = -.17, p = .016). A floodlight (Johnson-Neyman) and B. Keith Payne (2015), “Subjective Status Shapes Political analysis showed that boosting perceptions of social similarity (vs. Preferences,” Psychological Science, 26(1), 15–26. dissimilarity) increased redistribution support among liberals (who Carney, Dana R., John T. Jost, Samuel D. Gosling, and Jeff Potter scored 2.39 or lower on the ideology scale), but reduced redistribu- (2008), “The Secret Lives of Liberals and Conservatives: tion support among conservatives’ (who scored 7.49 or higher on Personality Profiles, Interaction Styles, and the Things They the ideology scale). This happened because boosting perceptions of Leave Behind,” Political Psychology, 29(6), 807–40. social similarity (vs. dissimilarity) weakened liberals’, but strength- Chernev, Alexander (2011), “Semantic Anchoring in Sequential ened conservatives’, belief in the fairness of unequal outcomes (in- Evaluations of Vices and Virtues,” Journal of Consumer teraction: b = .13, p = .003; moderated mediation: a x b = -.0214, Research, 37(5), 761–74. 95% CI = [-.0578, -.0034]). Chow, Rosalind M. and Jeff Galak (2012), “The Effect of Study 2 measured perceptions of social similarity among others Inequality Frames on Support for Redistributive Tax Policies,” (“to what extent do you think Americans are different from or similar Psychological Science, 23(12), 1467–69. to each other?” 1 = “very different” to 7 = “very similar”), which Clarkson, Joshua J., John R. Chambers, Edward R. Hirt, Ashley S. did not invoke the self to rule out the egocentric and projection ac- Otto, Frank R. Kardes, and Christopher Leone (2015), “The counts of the hypothesized effect. The study also measured fairness Self-Control Consequences of Political Ideology,” Proceedings beliefs more specifically by focusing on the fairness of economic of the National Academy of Sciences, 112(27), 8250–53. outcomes (e.g., “Economic positions are legitimate reflections of Conway, Andrew R. A., Michael J. Kane, Michael F. Bunting, people’s achievements”; Jost & Thompson, 2000). A regression D. Zach Hambrick, Oliver Wilhelm, and Randall W. Engle analysis revealed a significant interaction between social similarity (2005), “Working Memory Span Tasks: A Methodological and political ideology (b = -.11, p < .001). Perceiving high (vs. low) Review and User’s Guide,” Psychonomic Bulletin & Review, social similarity boosted liberals’, but reduced conservatives’, redis- 12(5), 769–86. tributive preferences. This once again happened because perceiving Advances in Consumer Research (Volume 48) / 941

Dawtry, Rael J., Robbie M. Sutton, and Chris G. Sibley (2015), Johnson, P. O. and J. Neyman (1936), “Tests of Certain Linear “Why Wealthier People Think People Are Wealthier, and Hypotheses and Their Application to Some Educational Why It Matters: From Social Sampling to Attitudes to Problems,” Statistical Research Memoirs, 1, 57–93. Redistribution,” Psychological Science, 26(9), 1389–1400. Jost, John T. (2007), “Coda--After ‘The End of the End of Duell, Dominik “Social Heterogeneity an Its Ambiguous Effect on Ideology.,’” American Psychologist, 62(9), 1077–80. Preferences for Redistribution,” 47. ——— (2006), “The End of the End of Ideology.,” American Engle, Randall W, James E Laughlin, Stephen W Tuholski, Psychologist, 61(7), 651–70. and Andrew R A Conway “Working Memory, Short-Term ——— (2017), “The Marketplace of Ideology: ‘Elective Affinities’ Memory, and General Fluid Intelligence: A Latent-Variable in Political Psychology and Their Implications for Consumer Approach,” 23. Behavior,” Journal of Consumer Psychology, 27(4), 502–20. Faul, Franz, Edgar Erdfelder, Albert-Georg Lang, and Axel Buchner Jost, John T., Jack Glaser, Arie W. Kruglanski, and Frank J. (2007), “G*Power 3: A Flexible Statistical Power Analysis Sulloway (2003), “Political Conservatism as Motivated Social Program for the Social, Behavioral, and Biomedical Sciences,” Cognition,” Psychological Bulletin, 129(3), 339–75. Behavior Research Methods, 39(2), 175–91. Jost, John T. and Orsolya Hunyady (2005), “Antecedents and Fernandes, Daniel and Naomi Mandel (2014), “Political Consequences of System-Justifying Ideologies,” Current Conservatism and Variety-Seeking,” Journal of Consumer Directions in Psychological Science, 14(5), 260–65. Psychology, 24(1), 79–86. Jost, John T., Brian A. Nosek, and Samuel D. Gosling (2008), Frank, Douglas H., Klaus Wertenbroch, and William W. Maddux “Ideology: Its Resurgence in Social, Personality, and Political (2015), “Performance Pay or Redistribution? Cultural Psychology,” Perspectives on Psychological Science, 3(2), Differences in Just-World Beliefs and Preferences for Wage 126–36. Inequality,” Organizational Behavior and Human Decision Khan, Romana, Kanishka Misra, and Vishal Singh (2013), Processes, 130, 160–70. “Ideology and Brand Consumption,” Psychological Science, Friedman, Naomi P., Akira Miyake, Robin P. Corley, Susan E. 24(3), 326–33. Young, John C. DeFries, and John K. Hewitt (2006), “Not Knight, Kathleen (1999), “Liberalism and Conservatism, in All Executive Functions Are Related to Intelligence,” Measures of Political Attitudes,” ed. Robinson, John P, Shaver, Psychological Science, 17(2), 172–79. Phillip R, Wrightsman, Lawrence S, San Diego: Academic Goodwin, Cathy and Ivan Ross (1992), “Consumer Responses to Press, 59–158. Service Failures: Influence of Procedural and Interactional Koenig, Harold G., and Arndt Büssing (2010), “The Duke Fairness Perceptions,” Journal of Business Research, 25(2), University Religion Index (DUREL): A Five-Item Measure for 149–63. Use in Epidemological Studies.” Religions, 1(1), 78-85. Graham, Jesse, Jonathan Haidt, Sena Koleva, Matt Motyl, Ravi Lammers, Joris and Matt Baldwin (2018), “Past-Focused Temporal Iyer, Sean P. Wojcik, and Peter H. Ditto (2013), “Moral Communication Overcomes Conservatives’ Resistance to Foundations Theory,” in Advances in Experimental Social Liberal Political Ideas,” Journal of Personality and Social Psychology, Elsevier, 55–130, https://linkinghub.elsevier.com/ Psychology, 114(4), 599–619. retrieve/pii/B9780124072367000024. Laran, Juliano (2020), “Self-Control: Information, Priorities, and Graham, Jesse, Brian A. Nosek, Jonathan Haidt, Ravi Iyer, Resources,” Consumer Psychology Review, 3(1), 91–107. Spassena Koleva, and Peter H. Ditto (2011), “Mapping the Lee, Saerom, Karen Page Winterich, and William T. Ross (2014), Moral Domain,” Journal of personality and social psychology, “I’m Moral, but I Won’t Help You: The Distinct Roles of 101(2), 366–85. Empathy and Justice in Donations,” Journal of Consumer Haidt, Jonathan, Jesse Graham, and Craig Joseph (2009), “Above Research, 41(3), 678–96. and Below Left–Right: Ideological Narratives and Moral Luttmer, Erzo F. P. (2001), “Group Loyalty and the Taste for Foundations,” Psychological Inquiry, 20(2–3), 110–19. Redistribution,” Journal of Political Economy, 109(3), Haidt, Jonathan and Craig Joseph “19 The Moral Mind: How Five 500–528. Sets of Innate Intuitions Guide the Development of Many Makens, James C. (1965), “Effect of Brand Preference upon Culture-Specific Virtues, and Perhaps Even Modules,” 31. Consumers Perceived Taste of Turkey Meat,” Journal of ——— (2004), “Intuitive Ethics: How Innately Prepared Intuitions Applied Psychology, 49(4), 261–63. Generate Culturally Variable Virtues,” Daedalus, 133(4), Martin, Matthew M., and Rebecca B. Rubin (1995), “A New 55–66. Measure of Cognitive Flexibility,” Psychological Reports, Hayes, A. F. (2017), Introduction to Mediation, Moderation, and 76(2), 623-26.. Conditional Process Analysis: A Regression Based Approach Miyake, Akira, Naomi P. Friedman, Michael J. Emerson, Alexander (2nd ed.), New York: Guilford. H. Witzki, Amy Howerter, and Tor D. Wager (2000), “The ______(2013). Introduction to Mediation, Moderation, and Unity and Diversity of Executive Functions and Their Conditional Process Analysis. New York: Guilford Contributions to Complex ‘Frontal Lobe’ Tasks: A Latent Hess, Ronald L., Shankar Ganesan, and Noreen M. Klein (2003), Variable Analysis,” Cognitive Psychology, 41(1), 49–100. “Service Failure and Recovery: The Impact of Relationship Muraven, Mark and Elisaveta Slessareva (2003), “Mechanisms Factors on Customer Satisfaction,” Journal of the Academy of of Self-Control Failure: Motivation and Limited Resources,” Marketing Science, 31(2), 127–45. Personality and Social Psychology Bulletin, 29(7), 894–906. Hirsh, Jacob B., Colin G. DeYoung, Xiaowen Xu, and Jordan Nosek, Brian, Mahzarin R. Banaji, and John T. Jost (2009), “The B. Peterson (2010), “Compassionate Liberals and Polite Politics of Intergroup Attitudes,” in Social and Psychological Conservatives: Associations of Agreeableness With Political Bases of Ideology and System Justification, ed. John T. Jost, Ideology and Moral Values,” Personality and Social Aaron C. Kay, and Hulda Thorisdottir, Oxford University Psychology Bulletin, 36(5), 655–64. Press, 480–506. 942 / Toward a Further Understanding of How Political Ideology Drives Motivation and Values

Oppenheimer, Daniel M., Tom Meyvis, and Nicolas Davidenko Rubin, Zick and Letitia Anne Peplau (1975), “Who Believes in a (2009), “Instructional Manipulation Checks: Detecting Just World?,” Journal of Social Issues, 31(3), 65–89. Satisficing to Increase Statistical Power,” Journal of Skitka, Linda J., Elizabeth Mullen, Thomas Griffin, Susan Experimental Social Psychology, 45(4), 867–72. Hutchinson, and Brian Chamberlin (2002), “Dispositions, Ordabayeva, Nailya and Pierre Chandon (2011), “Getting Ahead Scripts, or Motivated Correction? Understanding Ideological of the Joneses: When Equality Increases Conspicuous Differences in Explanations for Social Problems,”Journal of Consumption among Bottom-Tier Consumers,” Journal of Personality and Social Psychology, 83(2), 470–87. Consumer Research, 38(1), 27–41. Smith, Amy K., Ruth N. Bolton, and Janet Wagner (1999), “A Ordabayeva, Nailya and Daniel Fernandes (2018), “Better or Model of Customer Satisfaction with Service Encounters Different? How Political Ideology Shapes Preferences for Involving Failure and Recovery,” Journal of Marketing Differentiation in the Social Hierarchy,” ed. Darren Dahl Research, 36(3), 356–72. and Cait Lamberton, Journal of Consumer Research, 45(2), Todd Maddox, W. and Arthur B. Markman (2010), “The 227–50. Motivation–Cognition Interface in Learning and Decision ——— (2017), “Similarity Focus and Support for Redistribution,” Making,” Current Directions in Psychological Science, 19(2), Journal of Experimental Social Psychology, 72, 67–74. 106–10. Plassmann, H., J. O’Doherty, B. Shiv, and A. Rangel (2008), Yntema, Douwe B. and Gayle E. Mueser (1962), “Keeping Track “Marketing Actions Can Modulate Neural Representations of Variables That Have Few or Many States,” Journal of of Experienced Pleasantness,” Proceedings of the National Experimental Psychology, 63(4), 391–95. Academy of Sciences, 105(3), 1050–54. ——— (1960), “Remembering the Present States of a Number of Posner, M. I., & Cohen, Y. (1984), “Components of Visual Variables,” Journal of Experimental Psychology, 60(1), 18–22. Orienting,” Attention and Performance X: Control of Zmigrod, Leor, Peter J. Rentfrow, and Trevor W. Robbins (2018), Language Processes, 32, 531-56. “Cognitive Underpinnings of Nationalistic Ideology in the Robinson, Michael D., Deirdre M. Cassidy, Ryan L. Boyd, Context of Brexit,” Proceedings of the National Academy of and Adam K. Fetterman (2015a), “The Politics of Time: Sciences, 115(19), E4532–40. Conservatives Differentially Reference the Past and Liberals Zucker, Gail Sahar and Bernard Weiner (1993), “Conservatism and Differentially Reference the Future,”Journal of Applied Social Perceptions of Poverty: An Attributional Analysis1,” Journal Psychology, 45(7), 391–99. of Applied Social Psychology, 23(12), 925–43. ——— (2015b), “The Politics of Time: Conservatives Differentially Reference the Past and Liberals Differentially Reference the Future,” Journal of Applied Social Psychology, 45(7), 391–99. Persuasive Uncertainty: Toward Understanding How Uncertainty Influences the Formulation of Beliefs Chair: Rory M. Waisman, University of Alberta, Canada

Paper #1: Undermining Your Case Can Enhance Your Impact: the final paper, Waisman and Häubl examine how uncertainty that A Framework for Understanding the Positive Effects of Acts of originates from a source completely unrelated to a consumption deci- Receptiveness in Persuasion sion alters processing and influences evaluation of that consumption Mohamed A. Hussein, Stanford University, USA decision. They find that when consumers feel uncertain and make a Zakary Tormala, Stanford University, USA taste-based decision, their decision processing becomes faster, more fluent, and more favorable to their chosen alternative; and as a result, Paper #2: Perceived Drivers of Change in Financial Well-Being consumers become more confident in their choice. Predict Partisan Lean and Response to Policy Messages Collectively, these four papers identify conditions under which Job M. T. Krijnen, University of California Los Angeles, USA uncertainty originating from a variety of sources alters evaluations Jon Bogard, University of California Los Angeles, USA and beliefs, and uncover the mechanisms through which uncertainty Gülden Ülkümen, University of Southern California, USA operates to generate these persuasive effects. This session will appeal Craig R. Fox, University of California Los Angeles, USA to a broad range of scholarly interests including persuasion, uncer- Paper #3: How Probability Information Impacts Outcome tainty, and the intersection of consumer psychology with judgment Judgments and decision making. Daniella Kupor, Boston University, USA Kristin Laurin, University of British Columbia, Canada Undermining Your Case Can Enhance Your Impact: A Paper #4: How Uncertainty Boosts Confidence in Consumption Framework for Understanding the Positive Effects of Acts Decisions of Receptiveness in Persuasion Rory M. Waisman, University of Alberta, Canada Gerald Häubl, University of Alberta, Canada EXTENDED ABSTRACT Past research has uncovered numerous actions a source can SESSION OVERVIEW take that would appear to undermine but in fact frequently enhance Decades of research across the fields of economics (e.g., Rabin persuasion. For example, expressing doubt rather than confidence, and Thaler 2001), psychology (e.g., Kahneman and Tversky 1982), conveying that there is uncertainty rather than certainty around an ar- judgment and decision making (e.g., Coombs and Beardslee 1954), gument or estimate, asking questions rather than making statements, and marketing (e.g., Bettman, Luce, and Payne 1998) have revealed playing up rather than playing down one’s personal mistakes, and that uncertainty profoundly impacts people’s attitudes. We know, for acknowledging rather than hiding opposing viewpoints would all instance, that uncertain attitudes are more easily influenced (Litt and seem to limit one’s persuasive impact, but in fact they often boost it. Tormala 2010), less persistent (Bassili 1996), and less likely to pre- We review prior research on these topics and propose a theo- dict behavior (Fazio and Zanna, 1978), and that uncertainty motivates retical framework that explains when and why such counterintuitive systematic processing of persuasive messages (Chen, Duckworth, actions enhance persuasion. In particular, we posit that these diverse and Chaiken 1999; Tiedens and Linton 2001). We have less clarity, actions cohere around a single underlying construct: acts of recep- though, about the mechanisms through which different sources of tiveness. We define acts of receptiveness as behaviors or actions that uncertainty influence the formulation and revision of evaluations and signal a source’s openness to ideas, arguments, and attitudes that are beliefs. While uncertainty is a dimension of an attitude, it can also new or opposing to his or her own. In other words, acts of recep- be a characteristic of a persuasion source, a property attributed to tiveness suggest to recipients that the source is not overly zealous, an attitude object, a feature of outcome expectations or evaluations, biased, or one-sided in his or her beliefs. or a feeling arising from sources unrelated to a focal decision. In Consider the conditions under which acts of receptiveness boost this session we contribute to answering one key question: How does persuasion. Based on our review, we suggest that acts of receptive- uncertainty originating from different sources impact people’s evalu- ness are especially likely to promote persuasion when source exper- ations and beliefs? tise or status is high but not low. Under high expertise conditions, We draw together four papers that each probe the mechanisms it appears that people can be more persuaded by those who express of action of a different source of uncertainty that shape evaluations doubt rather than confidence about their opinion, who mention rather and beliefs. In the first paper, Hussein and Tormala integrate prior than mask mistakes they have made, who highlight flaws in their ar- persuasion research to develop a comprehensive theoretical frame- guments or reference views that oppose their own, and who ask ques- work charting conditions in which uncertainty acknowledged by a tions instead of making declarative statements. Under low source ex- source promotes persuasion by enhancing source perceptions and pertise or status conditions, acts of receptiveness generally offer little increasing involvement. Krijnen, Bogard, Ülkümen, and Fox illumi- advantage or can even backfire. nate the manner in which the perceived nature of uncertainty in the In addition, when source credibility is high, acts of receptive- domain of financial well-being affects the degree to which persuasive ness appear to promote persuasion through two primary mecha- messages shift beliefs about social policy issues. They find that mes- nisms: enhanced source perceptions and increased involvement. sage framing compatible with uncertainty beliefs facilitates the per- First, acts of receptiveness appear to trigger more favorable source suasive power of messaging in support of social welfare policies. In impressions, fostering perceptions of honesty, likeability, thoughtful- the third paper, Kupor and Laurin investigate how knowledge about ness, and so on. Second, acts of receptiveness can stimulate involve- uncertain outcomes influences beliefs about the magnitude of those ment, which leads to deeper processing. These processes in turn can outcomes, and alters evaluations of those outcomes as a result. In facilitate persuasion as long as the arguments are reasonably strong.

We organize our review around four primary acts of receptiveness Conclusion. In summary, through an extensive review of the that show the predicted persuasion effect (i.e., acts of receptiveness literature, we document a vast array of acts of receptiveness that result in greater persuasion) and speak to the proposed conditions explain when and why particular counterintuitive actions—such as and/or mechanisms guiding it. conveying uncertainty or admitting mistakes or limitations—can en- Conveying Uncertainty. One class of acts of receptiveness in- hance persuasion. In doing so, we integrate seemingly disparate re- volves conveying uncertainty. By conveying uncertainty, we mean search findings around a single theoretical framework with concrete communicating or displaying doubt, hesitance, or ambiguity. For ex- practical implications. ample, a source who expresses doubt about an opinion (Karmarkar and Tormala 2010) or provides a range of possible outcomes when Perceived Drivers of Change in Financial Well-Being making a forecast or prediction (Howe et al. 2019) can be viewed as Predict Partisan Lean and Response to Policy Messages conveying uncertainty. Conveying uncertainty can be viewed as an act of receptiveness because the source takes a position, expresses an EXTENDED ABSTRACT argument, or provides an estimate or answer, but implicitly acknowl- Economic inequality is on the rise, negatively affecting consum- edges that other positions, arguments, or estimates exist and may be ers’ well-being around the globe (Alvaredo, Chancel, Piketty, Saez, valid. By conveying uncertainty, the source suggests that he or she & Zucman, 2018; Piketty & Saez, 2014; Saez & Zucman, 2016; Zuc- understands that his or her knowledge is not definitive and that he or man, 2019). Politicians have found it challenging to garner broad she is open to other perspectives on the topic. support for their proposals on how to deal with inequality, since Acknowledging Mistakes. Another class of acts of receptive- people often strongly disagree on how, when, and why the govern- ness involves admitting a mistake. By admitting a mistake, we mean ment should intervene by implementing income redistribution poli- explicitly or implicitly taking responsibility for an error. Explicitly cies. We argue that such disagreements arise because people differ in admitting a mistake may take the form of acknowledging the mistake their assessment of what causes an individual’s financial well-being and/or apologizing for it. For example, “I used to think X, but I was to change over time. Moreover, policy preferences are not fixed but wrong.” Implicitly admitting a mistake may take the form of chang- depend on the way a proposal is communicated. Better understanding one’s opinion to correct for a mistake. For example, if someone ing the persuasiveness of different policy proposals and messages points out mistakes in the source’s arguments, the source could up- therefore requires an appropriate model of people’s beliefs about the date his/her arguments to address the error. nature of uncertainty regarding changes in financial well-being. Acknowledging that one has committed a mistake explicitly In this article, we propose that beliefs about uncertainty in fi- or implicitly signals openness to information that is potentially ad- nancial well-being are best thought of along three conceptually in- versarial to one’s position. Rather than reverting to defensiveness dependent dimensions, respectively capturing the extent to which or blame, acknowledging a mistake signals that one is not overly changes in financial well-being are perceived to be (1) knowable and zealous or single-minded; that one is open to changing one’s mind within individuals’ control due to individual factors such as effort and updating one’s position based on the available information. Past (‘rewarding’), (2) knowable and outside of individuals’ control due research has shown that explicitly or implicitly admitting mistakes to systemic factors such as favoritism and discrimination (‘rigged’), can enhance persuasion (Reich & Maglio 2019; Kupor et al. 2018; and (3) inherently unpredictable and determined by chance events Aronson et al. 1966; John et al. 2019; Gonzales 2012; Reich & Tor- (‘random’). In four studies, we use a 9-item F-EARS scale that we mala 2013). developed to assess these three dimensions of uncertainty regarding Highlighting the Negatives. Highlighting the negatives involves changes in an individual’s financial well-being. bringing up defects, limitations, downsides, and opposing arguments Mapping beliefs about changes in financial well-being along to one’s own point of view or position. We submit that doing so sig- the conceptually independent rewarding, rigged, and random dimen- nals open-mindedness and receptiveness, and thus can offer a persua- sions combines and extends insights from previous research on be- sive advantage over doing the opposite—for instance, sharing only liefs about individual accountability (e.g., Alesina & Glaeser, 2004; supportive arguments or asserting that one’s proposal, position, or Fong, 2001; Konow, 2000) and on beliefs about the nature of uncer- offer is superior on all dimensions. Past research on two-sided mes- tainty (Fox & Ülkümen, 2011; Tannenbaum, Fox, & Ülkümen, 2017; sages (Etgar and Goodwin 1982; Golden and Alpert 1986; Kamins Ülkümen, Fox, & Malle, 2016). and Assael 1987; Settle and Golden 1974; Smith and Hunt 1978; for In Study 1, we examine the relative importance of each di- a review, see Crowley and Hoyer 1994) and on the blemishing ef- mension as a predictor of political ideology, in search for a more fect (Ein-gar et al. 2012) suggests that acknowledging negatives can complete understanding of what distinguishes conservative ideology enhance persuasion under specifiable conditions. from liberal ideology. Using a nationally representative sample (N Asking Questions. Asking questions, as opposed to making de- = 1102), we find that conservatives generally score higher on the clarative statements, is another class of acts of receptiveness. Intui- rewarding dimension (r = 0.13, p < .001), lower on the rigged dimen- tively, asking a question, even when one already knows the answer, sion (r = -0.20, p < .001), and lower on the random dimension (r = seems more open and inviting than does making a statement. A per- -0.09, p = .005), even when controlling for a number of demographic son who makes declarative statements (e.g., “this new policy will and related psychometric measures. boost productivity”), for instance, is likely to be seen as more firm We further predict that political orientation should not be the or assertive in his opinion than one who phrases the same position ultimate determinant of support for social welfare policies. Studies as a question (e.g., “won’t this new policy boost productivity?”). By 2-4 (total N=2560; pre-registered) test our prediction that we can implicitly inviting the recipients’ input, a source who asks questions increase support for social welfare policies when we describe poli- might be seen as open-minded and receptive to others’ views or per- cies in a way that is compatible with people’s beliefs about changes spectives—and past research found that asking questions can facili- in financial well-being. tate persuasion under specifiable conditions (Zillmann 1972; Petty et Study 2 asked participants to indicate how important they find al. 1981; Burnkrant & Howard 1984). each of three possible goals that a government might pursue: (1)“The government should use resources to incentivize and enable people to Advances in Consumer Research (Volume 48) / 945 pull themselves out of financial hardship and realize their full poten- How Probability Information Impacts Outcome tial” (i.e., an incentivize goal); (2) “The government should allocate Judgments resources to individuals belonging to disadvantaged groups that routinely experience financial hardship” (i.e., a redistribution goal); (3) EXTENDED ABSTRACT “The government should pool resources to support people when they Consumers’ decisions about which foods they eat and which happen to experience unforeseeable financial hardship” (i.e., a risk- health products they utilize influence their welfare and well-being. pool goal). The results suggest that higher scores on the [reward- When consumers believe these products may produce greater ben- ing/rigged/random] dimension were associated with a more positive efit, they are more likely to consume them; when they believe these change in the importance rating of the [incentivizing/redistribution/ products may cause greater harm, they are less likely to consume risk-pooling] goal, respectively, compared to the other two goals. them (Slovic 1964). When judging the size of products’ benefits and In Study 3, we presented participants with four different pub- harms, consumers are often aware of these outcomes’ probability of lic policy proposals: a more extensive disaster recovery program, a occurring. In fact, this information is often highlighted in product tuition-free higher education system, and a more extensive food-pur- advertisements, product packaging, and product websites. chasing assistance program, and a universal health coverage system. Imagine that a side effect increases a consumer’s eosinophil Each proposal was followed by three different arguments in favor of count from 397 to 435 per microliter. Of note, the side effect’s prior the policy: (1) an incentivizing argument, which highlighted how the probability of increasing eosinophils provides no objective insight policy would provide assistance to those who deserve it most; (2) into the degree to which that consumer’s eosinophil count has now a redistribution argument, which highlighted how the policy would increased, and thus this prior probability should not inform the indi- provide assistance to the groups that need it most (e.g., low-income, vidual’s decision about whether to incur a cost (e.g., buy a medical unemployed people); (3) a risk-pool argument, which highlighted treatment) to remedy that increase. Nevertheless, we theorize that how the policy would pool tax-money to collectively pay in case an such objectively irrelevant information biases magnitude judgments: individual experiences an unexpected life event. The results suggest Consumers forecast that more probable product outcomes will be that higher scores on the [rewarding/rigged/random] dimension are larger in magnitude, and also perceive them as larger in magnitude associated with a more positive change in policy support in response after those outcomes unfold. We propose that this bias emerges be- to the [incentivize/redistribution/risk-pooling] argument, respective- cause people believe larger prior probabilities emanate from more ly, as compared to the other two arguments. powerful causal antecedents that produce outcomes with larger mag- In study 4, we presented participants with statements of three nitudes. political candidates regarding welfare policies on higher education, In Study 1, undergraduates were randomly assigned to either disaster recovery, and food purchasing assistance. Each candidate a Smaller Probability or Larger Probability condition, and viewed put forward either incentivizing arguments, redistribution argu- a Claritin drug advertisement which noted that Claritin causes 6% ments, or risk-pooling arguments for all three proposals. We find that (vs. 68%) of users to experience a side effect of coughing. When par- higher scores on the [rewarding/redistribution/random] dimension ticipants learned that someone experienced coughing as a result of are associated with a more positive change in support for the [incen- taking Claritin, they forecasted that the person experienced a larger tivizing/redistribution/risk-pooling] candidate, respectively, as com- number of coughs, t(108) = 3.35, p < .001. Study 2 replicated this pared to the other two candidates, even when controlling for political phenomenon when the probability information was communicated ideology. Perhaps more importantly, scores on the three dimensions via frequencies rather than percentages, t(198) = 4.23, p < .001. are associated with greater intention to vote for candidates that made Studies 3-4 found that probability information also distorts per- belief-compatible arguments. ceptions of objective magnitude information. In Study 3, participants These findings broaden our understanding of what is driving in the Smaller (vs. Larger) Probability condition read that when a support for or opposition to redistribution policies. To truly grasp person eats an orange, the orange has a 14% (vs. 66%) chance of political and policy preferences we need a complete and accurate increasing the concentration of trypsin in the body. Next, participants model of how people think about changes in financial well-being. read that someone’s trypsin count increased from 397 to 435 per mi- Our findings confirm that a model based only on perceptions of in- croliter after eating an orange. Participants entered the percentage by dividual control is insufficient, and that distinguishing between the which the person’s trypsin increased when it increased from 397 to perceived knowability and randomness of changes in financial well- 435 per microliter. Participants were incentivized for accuracy. Par- being allows for more complete understanding and better predictive ticipants in the Larger (vs. Smaller) Probability condition perceived ability. that this increase constituted a larger percent increase, t(198) = 4.14, Despite the commonly held wisdom that support for welfare p < .001. policies depends on political attitudes, we show that belief-compati- Study 4 examined an anchoring explanation. Participants in ble messages can increase support for redistribution policies and vot- the Smaller (vs. Larger) Probability condition read that consuming ing intentions for the politicians proposing these policies, across the a banana has a .03% (vs. 22%) chance of increasing people’s eye- entire political spectrum. When people disagree about a certain dis- lash length. Participants then read that a woman’s lashes increased tribution policy, this disagreement may stem from a failure to jointly in length after she ate a banana, and viewed pictures of her lashes define what exactly the policy entails–who it helps, on what basis, before and after she ate the banana. Participants further read that and with what purpose. Whenever such disagreements arise, a slight her lashes were now .75 inches after eating the banana; beneath this reframing of the policy, guided by an understanding of beliefs about information and the before-and-after pictures, participants indicated the uncertainty in financial well-being, may bring opinions on the the length of her lashes before she ate the banana. Whereas an an- matter closer together. choring alternative predicts that participants who viewed the larger prior probability would report a larger initial magnitude, we found that participants who viewed the larger prior probability reported a smaller initial magnitude (which indicated that they perceived more 946 / Persuasive Uncertainty: Toward Understanding How Uncertainty Influences the Formulation of Beliefs growth depicted in the before-and-after pictures), t(198) = 2.56, p = 2008). Nevertheless, little is known about the influence of inciden- .011. tal uncertainty on consumers’ confidence in their consumption de- Study 5 found this bias emerges because people believe larger cisions. The present research reveals that the impact of incidental probabilities emanate from more powerful causal antecedents that uncertainty on confidence is not limited to the congruency effects produce outcomes with larger magnitudes. In addition, it found that demonstrated in other domains, but rather is context dependent. Evi- the current phenomenon impacts intended behavior. Specifically, dence from three studies shows that the negative effect of incidental participants in Study 5 read that consuming pumpkin seeds has a uncertainty on confidence vanishes, and even turns positive, when chance of increasing bone density. Participants viewed a bar graph consumers make subjective (i.e., matters of taste) consumption deci- which depicted that eating one pound of pumpkin seeds has a 6% sions. chance of increasing bone density (and “6%” was printed above the Intuitively, decision confidence is a belief in the correctness of bar). The only difference between conditions was the y-axis, which a decision. But evidence points to cues from processing—fluency, ranged from 0% to 5% (vs. 100%) in the Larger (vs. Smaller) Prob- speed, degree of deliberation and conflict—as the key drivers of de- ability condition. A pretest revealed that this manipulation caused cision confidence (Koriat 2012). We propose that the effect of inci- participants to perceive that the 6% probability was larger in the dental uncertainty on cues from processing differs depending on the Larger (vs. Smaller) Probability condition. Participants in the Larg- context of the decision. Given the self-referential nature of subjec- er (vs. Smaller) Probability condition judged that pumpkin seeds tive consumption decisions, we posit that increased processing mo- were more powerful (t(202) = 5.49, p < .001), that pumpkin seeds tivation (Tiedens and Linton 2001) and greater semantic clustering would more greatly increase bone density if it did increase a person’s (Bass, de Dreu, and Nijstad 2011, 2012) that arise under conditions bone density (t(202) = 5.04, p < .001), and were more likely to eat of uncertainty prompt the generation of thoughts more favorable to pumpkin seeds (t(202) = 3.76, p < .001). A serial mediation with the chosen alternative and lead to faster decision making. Rather bootstrapping indicated that participants in the Larger (vs. Smaller) than feeling more difficult, these cues signal greater confidence. Probability condition were more likely to eat the seeds because they In Study 1, participants (N = 405) read a research abstract that perceived them to be more powerful and thus would produce a larger manipulated incidental uncertainty vs. certainty (used in all studies; boost in bone density. Study 6 leveraged a similar design in the field, adapted from Faraji-Rad and Pham 2017). Participants decided be- and found that the current phenomenon shifts consumers’ likelihood tween renting a functionally superior or an emotionally appealing of clicking on online ads to purchase a promoted product (χ 2 = 5.64, apartment (Faraji-Rad and Pham 2017) and reported their confidence p = .018). in and reasons for their decision. Participants’ selection served as Study 5 found that this bias emerges because people believe proxy for the objectivity of the decision (supported by analysis of larger probabilities emanate from more powerful causal antecedents reported reasons; significantly greater objectivity for the function- that produce outcomes with larger magnitudes. As a result, Study 7 ally superior apartment, p < .001). The congruency effect evidenced found that probability information no longer influences magnitude in past research emerges when the functionally superior apartment judgments when people believe that the antecedent that determines is chosen (p = .018) but vanishes when the emotionally appealing an outcome’s magnitude is different than the antecedent that deter- apartment is selected (F1,402 = 0.04), demonstrating that incidental mines that outcome’s probability. A 2 (Probability: Smaller vs. Larg- uncertainty need not always undermine decision confidence. er) × 2 (Antecedent: Connected vs. Disconnected) ANOVA revealed In Study 2 we directly manipulated context to test our hypoth- a significant interaction (F(1, 396) = 7.67, p = .006): Probability esis that, unlike the judgments studied in the past, confidence in sub- information influenced magnitude perceptions when participants be- jective preferential choices is boosted by incidental uncertainty. Par- lieved that the same antecedent determined both an outcome’s prob- ticipants (N = 374) either selected a painting from four alternatives ability and its magnitude (F(1, 396) = 18.84, p < .001), but not when (subjective choice) or responded to a reasoning problem (judgment). they did not (F(1, 396) = .14, p = .708). A directional congruency effect emerges in the judgment context, These findings provide novel insight into the inferences con- whereas uncertainty boosts decision confidence in the subjective sumers draw from probabilities and the architecture of consumers’ choice, (F1,371 = 3.97, p < .047). Uncertainty increases the perceived causal schemas. Importantly, the perceived magnitude of products’ difficulty of the judgment (p = .031), but not of the subjective choice benefits is a dominant input into consumers’ decisions about whether (t195 = 0.08), suggesting decision difficulty does not explain the posi- to purchase those products (Rundmo and Nordfjjaern 2017; Slovic tive effects of uncertainty and pointing to a difference in the nature 1964). Thus, our findings have numerous strategic implications. of processing between these two contexts. Study 3 generalizes the positive effect of uncertainty to another How Uncertainty Boosts Confidence in Consumption consumption domain and probes the nature of decision processing in Decisions matters of subjective taste. Participants (N = 118) selected a video to watch from five comedy video screenshots (unknown to partici- EXTENDED ABSTRACT pants, all segments from the same video), reported decision confi- Could feeling uncertain increase consumers’ confidence in their dence, decision difficulty, and thoughts they had about the selected preferential choices? The answer to this question is not obvious. video while making the decision, then watched the video. Uncer- But it is important because decision confidence increases consum- tainty boosts decision confidence (p = .012) but has no detectable ers’ willingness to pay (Thomas and Menon 2007) and their likeli- effect on difficulty (t116 = 0.25). Participants were asked if they rec- hood to complete a purchase (Huang, Korfiatis, and Chang 2018), ommend including the video in a collection of enjoyable videos for and it promotes favorable post-consumption evaluations (Heitmann, future research. Uncertainty increases recommendation (p = .026), Lehmann, and Herrmann 2007). an effect mediated by decision confidence (b = .359, 95% CI:[.045, According to prior research, incidental feelings of uncertainty .733]). Absence of a significant interaction with scores on the causal undermine confidence in understanding (Clore and Parrott 1994), uncertainty scale (α = .90; Weary and Edwards 1994) suggests the memory (Jacobson, Weary, and Lin 2008), predictions (Tiedens and boundary condition found in prior research (effects limited to those Linton 2001), and causal judgments (Wichman, Brunner, and Weary scoring high; e.g., Jacobson et al. 2008) is not present in this context. Advances in Consumer Research (Volume 48) / 947

Critically, incidental uncertainty leads to thoughts that are margin- Chen, Serena, Kimberly Duckworth, and Shelly Chaiken ally more semantically clustered (p = .068; Bhatia 2017, 2019; Hass (1999), “MotivatedHheuristic and Systematic Processing,” 2017) and favorable to the chosen alternative (p = .077), and to faster Psychological Inquiry, 10(1), 44–49. decision making (p = .038). Clore, Gerald L. and W. Gerrod Parrott (1994), “Cognitive Feelings This paper sheds new light on how feelings of uncertainty in- and Metacognitive Judgments,” European Journal of Social fluence consumers’ beliefs about the quality of their consumption Psychology, 24(1), 101–115. decisions. In matters of subjective taste, consumers gain confidence Cojuharenco, Irina and Natalia Karelaia (2020), “When Leaders from incidental feelings of uncertainty. These findings inject nuance Ask Questions, Can Humility Premiums Buffer the Effects of into our understanding of the context dependent dynamics of deci- Competence Penalties?” Organizational Behavior and Human sion processing, offering initial support for the novel proposal that Decision Processes, 156 (January), 113–34. engaging in more systematic subjective evaluation of choice alter- Coombs, Clyde H. and David Beardslee (1954), “On Decision- natives may boost the fluency, speed, and favorability of decision Making Under Uncertainty,” In R. M. Thrall, C. H. Coombs, processing. and Davis R. L. (Eds.), Decision processes (pp. 255–285). Oxford, England: John Wiley and Sons, Inc. REFERENCES Dhar, Sanjay K., Claudia González-Vallejo, and Dilip Soman Ahluwalia, Rohini and Robert E Burnkrant (2004), “Answering (1999), “Modeling the Effects of Advertised Price Claims, Questions about Questions: A Persuasion Knowledge Tensile Versus Precise Claims?” Marketing Science, 18(2), Perspective for Understanding the Effects of Rhetorical 154–77. Questions,” Journal of Consumer Research, 31(18), 26-42. Ein-Gar, Danit, Baba Shiv, and Zakary L. Tormala (2012), “When Alesina, Alberto, Edward Glaeser, and Edward Ludwig Glaeser Blemishing Leads to Blossoming, The Positive Effect of (2004), Fighting poverty in the US and Europe: A world of Negative Information,” Journal of Consumer Research, 38(5), difference. Oxford University Press. 846–59. Alvaredo, Facundo, Lucas Chancel, Thomas Piketty, Emmanuel Etgar, Michael and Stephen A. Goodwin (1982), “One-Sided Versus Saez, and Gabriel Zucman (2018), World Inequality Report Two-Sided Comparative Message Appeals for New Brand 2018. Belknap Press. Introductions,” Journal of Consumer Research, 8(4), 460. Aronson, Elliot, Judith A Turner, and J Merrill Carlsmith (1963), Faraji-Rad, Ali and Michel Tuan Pham (2017), “Uncertainty “Communicator Credibility and Communication Discrepancy Increases the Reliance on Affect in Decisions,”Journal of Ad Determinants of Opinion Change,” Journal of Abnormal Consumer Research, 44(1), 1–21. Psychology, 67(1), 31–36. Fazio, Russell H. and Mark P. Zanna (1978), “On the Predictive Baas, Matthijs, Carsten De Dreu, and Bernard A. Nijstad (2011), Validity of Attitudes: The Roles of Direct Experience and “Creative Production by Angry People Peaks Early On, Confidence,” Journal of Personality, 46(2), 228–243. Decreases Over Time, and is Relatively Unstructured,” Fong, Christina (2001), “Social Preferences, self-Interest, and the Journal of Experimental Social Psychology, 47(6), 1107–1115. Demand for Redistribution,” Journal of Public Economics, Baas, Matthijs, Carsten De Dreu, and Bernard A. Nijstad (2012), 82(2), 225–246. “Emotions That Associate With Uncertainty Lead to Structured Fox, Craig R., and Gülden Ülkümen (2011), “Distinguishing Ideation,” Emotion, 12(5), 1004–1014. two dimensions of uncertainty,” Perspectives on Thinking, Bassili, John N (1996), “Meta-Judgmental Versus Operative Judging, and Decision Making, 21–35. Indexes of Psychological Attributes: The case of measures Golden, Linda L., and Mark I. Alpert (1987), “Comparative of attitude strength,” Journal of Personality and Social Analysis of the Relative Effectiveness of One- and Two- Psychology, 71(4), 637–653. Sided Communication for Contrasting Products.” Journal of Bettman, James R., Mary Frances Luce, and John W. Payne (1998), Advertising, 16(1), 18–68. “Constructive Consumer Choice Processes,” Journal of Goldsmith, Kelly and On Amir (2010), “Can Uncertainty Improve Consumer Research, 25(3), 187–217. Promotions?” Journal of Marketing Research 47(6), 1070–77. Bhatia, Sudeep (2017), “Associative Judgment and Vector Space Gonzales, M. H. (1992), “A Thousand Pardons, The Effectiveness Semantics,” Psychological Review, 124(1), 1–20. of Verbal Remedial Tactics During Account Episodes.” Bhatia, Sudeep (2019), “Semantic Processes in Preferential Journal of Language and Social Psychology 11(3), 133–51. Decision Making,” Journal of Experimental Psychology: Hass, Richard W. (2017), “Semantic Search During Divergent Learning, Memory, and Cognition, 45(4), 627–640. Thinking,” Cognition, 166, 344–357. Blankenship, Kevin L. and Traci Y. Craig (2007), “Language Heitmann, Mark, Donald R. Lehmann, and Andreas Herrmann and Persuasion, Tag Questions as Powerless Speech or as (2007), “Choice Goal Attainment and Decision and Interpreted in Context,” Journal of Experimental Social Consumption Satisfaction,” Journal of Marketing Research, Psychology, 43 (1), 112–18. 44(2), 234–250. Blankenship, Kevin L. and Thomas Holtgraves (2005), “The Howard, Daniel J. (1990), “Rhetorical Question Effects on Role of Different Markers of Linguistic Powerlessness in Message Processing and Persuasion, The Role of Information Persuasion,” Journal of Language and Social Psychology, Availability and the Elicitation of Judgment.” Journal of 24(1), 3–24. Experimental Social Psychology, 26(3), 217–39. Burnkrant, Robert E. and Daniel J Howard (1984), “Effects of the Howe, Lauren C., Bo MacInnis, Jon A. Krosnick, Ezra M. Use of Introductory Rhetorical Questions Versus Statements Markowitz, and Robert Socolow (2019), “Acknowledging on Information Processing,” Journal of Personality and Social Uncertainty Impacts Public Acceptance of Climate Scientists’ Psychology, 47(6), 1218–30. Predictions.” Nature Climate Change, October, 1-5. 948 / Persuasive Uncertainty: Toward Understanding How Uncertainty Influences the Formulation of Beliefs

Huang, Guei-Hua, Nikolaos Korfiatis, and Chun-Tuan Chang Reich, Taly and Zakary L. Tormala (2013), “When Contradictions (2018), “Mobile Shopping Cart Abandonment: The Roles of Foster Persuasion, An Attributional Perspective.” Journal of Conflicts, Ambivalence, and Hesitation,”Journal of Business Experimental Social Psychology, 49(3), 426–39. Research, 85, 165–174. Rundmo, Torbjørn and Trond Nordfjærn (2017), “Does Risk Huang, Karen, Michael Yeomans, Alison Wood Brooks, Julia Perception Really Exist?” Safety Science, 93, 230–240. Minson, and Francesca Gino (2017), “It Doesn’t Hurt to Ask, Saez, Emmanuel, and Gabriel Zucman (2016), “Wealth Inequality Question-Asking Increases Liking.” Journal of Personality in the United States Since 1913: Evidence from Capitalized and Social Psychology, 113(3), 430–52. Income Tax Data,” The Quarterly Journal of Economics, 131 Jacobson, Jill A., Gifford Weary, and Y. Sharon Lin (2008), “Causal (2), 519–578. Uncertainty and Metacognitive Inferences About Goal Settle, Robert B. and Golden, Linda L (1974), “Attribution Theory Attainment,” Cognition and Emotion, 22(7), 1276–1305. and Advertiser Credibility.” Journal of Marketing Research, John, Leslie K., Martha Jeong, Francesca Gino, and Laura Huang 11(2), 181-185 (2019), “The Self-Presentational Consequences of Upholding Shen, Luxi, Ayelet Fishbach, and Christopher K. Hsee (2015), One’s Stance in Spite of the Evidence.” Organizational “The Motivating-Uncertainty Effect, Uncertainty Increases Behavior and Human Decision Processes, 154 (September), Resource Investment in the Process of Reward Pursuit.” 1–14. Journal of Consumer Research, 41(5), 1301–15. Kahneman, Daniel and Amos Tversky (1982), “Variants of Shen, Luxi, Christopher K Hsee, and Joachim H Talloen (2019), Uncertainty,” Cognition, 11(2), 143–157. “The Fun and Function of Uncertainty, Uncertain Incentives Kamins, Michael A., Meribeth J. Brand, Stuart A. Hoeke, and Reinforce Repetition Decisions.” Edited by Gita V Johar, John C. Moe (1989), “Two-Sided versus One-Sided Celebrity Amna Kirmani, and Stijn M J van Osselaer. Journal of Endorsements, The Impact on Advertising Effectiveness and Consumer Research 46 (1), 69–81. Credibility.” Journal of Advertising, 18(2), 4–10. Slovic, Paul (1964), “Assessment of Risk Taking Behavior,” Karmarkar, Uma R. and Zakary L. Tormala (2010), “Believe Me, I Psychological Bulletin, 61, 220–233. Have No Idea What I’m Talking About, The Effects of Source Smith, Robert E. and Shelby D. Hunt (1978), “Attributional Certainty on Consumer Involvement and Persuasion.” Journal Processes and Effects in Promotional Situations.”Journal of of Consumer Research, 36(6), 1033–49. Consumer Research, 5(3), 149-158. Konow, James (2000), “Fair Shares: Accountability and Cognitive Tannenbaum, David, Craig R. Fox, and Gülden Ülkümen (2017), Dissonance in Allocation Decisions,” American Economic “Judgment Extremity and Accuracy Under Epistemic vs. Review, 90(4), 1072–1091. Aleatory Uncertainty” Management Science, 63 (2), 497–518. Koriat, Asher (2012), “The Self-Consistency Model of Subjective Tiedens, Larissa Z. and Susan Linton (2001), “Judgment Under Confidence,”Psychological review, 119(1), 80. Emotional Certainty and Uncertainty: The Effects of Specific Litt, Ab and Zakary L. Tormala (2010), “Fragile Enhancement Emotions on Information Processing,” Journal of Personality of Attitudes and Intentions Following Difficult Decisions,” and Social Psychology y, 81(6), 973–988. Journal of Consumer Research, 37(4), 584–598. Thomas, Manoj and Geeta Menon (2007), “When Internal Mishra, Himanshu, Arul Mishra, and Baba Shiv (2011), “In Reference Prices and Price Expectations Diverge: The role of Praise of Vagueness, Malleability of Vague Information as a confidence,”Journal of Marketing Research, 44(3), 401–409. Performance Booster.” Psychological Science, 22(6), 733–38. Ülkümen, Gülden, Craig R. Fox, and Bertram F. Malle (2016), Norton, Michael I., Jeana H. Frost, and Dan Ariely (2007), “Less “Two Dimensions of Subjective Uncertainty: Clues from Is More, The Lure of Ambiguity, or Why Familiarity Breeds Natural Language,” Journal of Experimental Psychology: Contempt.” Journal of Personality and Social Psychology, General, 145 (10), 1280. 92(1), 97. Weary, Gifford and John A. Edwards (1994), “Individual Petty, Richard E., John T. Cacioppo, and Martin Heesacker (1981), Differences in Causal Uncertainty,” Journal of Personality and “Effects of Rhetorical Questions on Persuasion, A Cognitive Social Psychology, 67(2), 308. Response Analysis.” Journal of Personality and Social Wichman, Aaron L., Ryan P. Brunner, and Gifford Weary (2008), Psychology 40(3), 432–40. “Immediate and Delayed Effects of Causal Uncertainty Piketty, Thomas, and Emmanuel Saez (2014), “Inequality in the Inductions on Uncertainty Accessibility,” Journal of Long Run,” Science, 344 (6186), 838–843. Experimental Social Psychology, 44(4), 1106–1113. Rabin, Matthew and Richard H. Thaler (2001), “Anomalies: Risk Zillmann, Dolf (1972), “Rhetorical Elicitation of Agreement in Aversion,” Journal of Economic Perspectives, 15(1), 219-232. Persuasion.” Journal of Personality and Social Psychology, Reich, Taly, Daniella M. Kupor, and Rosanna K. Smith (2017), 21(2), 159–65. “Made by Mistake, When Mistakes Increase Product Zillmann, Dolf and Joanne R. Cantor (1974), “Rhetorical Preference.” Journal of Consumer Research, August 1085- Elicitation of Concession in Persuasion.” Journal of Social 1103. Psychology, 94, 223–36. Reich, Taly and Sam J. Maglio (2020), “Featuring Mistakes, The Zucman, Gabriel (2019), “Global Wealth Inequality,” Annual Persuasive Impact of Purchase Mistakes in Online Reviews.” Review of Economics, 11. Journal of Marketing, 84(1), 52–65. Novel Effects of Context on Preference Formation Chair: Ioannis Evangelidis, ESADE Business School, Spain

Paper #1: The Contingent Impact of Context on Choice versus about—and anticipate lower feelings of regret from—choosing the Valuation high-quality high-price option in the presence (vs. absence) of the Fuad A. Shennib, Stanford Graduate School of Business, USA decoy. Ioannis Evangelidis, ESADE Business School, Universitat Finally, contrary to earlier research (Frederick et al. 2014; Yang Ramon Llull, Spain and Lynn, 2014), Alman and Urminsky provide evidence for the Minah Jung, New York University, USA attraction effect using visual stimuli. They show that the attraction effect can be observed when making size judgments between visual- Paper #2: Invoking Identity Changes How Consumers Resolve ly-presented product packages and when choosing between equally- Tradeoffs priced products. They also find that package orientation biases size Franklyn Shaddy, UCLA Anderson School of Management, judgments, moderating the attraction effect. USA Taken together, this session contributes to the understanding Ayelet Fishbach, University of Chicago Booth School of of how consumers construct their preferences as a function of the Business, USA decision context. Our work expands the understanding of drivers Itamar Simonson, Stanford Graduate School of Business, USA that influence consumers’ tendencies to rely on the context during Paper #3: Symmetric Dominance decision-making. Ioannis Evangelidis, ESADE Business School, Universitat Ramon Llull, Spain The Contingent Impact of Context on Choice versus Jonathan Levav, Stanford Graduate School of Business, Valuation Stanford University, USA Itamar Simonson, Stanford Graduate School of Business, EXTENDED ABSTRACT Stanford University, USA Consumers’ preferences are often swayed by the decision con- Paper #4: The Attraction Effect in Product Size Judgments text. Context effects, in which the addition of a new option to a bina- Sophie Alman, University of Chicago Booth School of ry set affects preferences for a (target) option, have been well estab- Business, USA lished across social-science disciplines (e.g., Huber, Payne, and Puto Oleg Urminsky, University of Chicago Booth School of 1982; Simonson 1989). Using two frequently employed elicitation Business, USA methods—choice and willingness-to-pay (WTP), we demonstrate that the occurrence of context effects critically depends on the elici- SESSION OVERVIEW tation procedures used. Decades of work in consumer research and psychology pro- We conjecture that the elicitation method impacts how prefer- vides evidence that consumers’ preferences depend on the context ences are constructed (Tversky, Sattath, and Slovic 1988). In choice, in which they are constructed. Classic work on context effects shows consumers predominantly rely on attribute weights to construct pref- that, contrary to normative theory, preference for a given product can erences. In WTP tasks, consumers rely on price perceptions rather depend on the presence (vs. absence) of other—often irrelevant—op- than attribute weights. In turn, they report higher WTP amounts for tions in the choice set. Recent research has challenged the robustness options perceived to be more expensive, reversing preferences from of these effects (Evangelidis et al., 2018; Frederick et al., 2014; Yang the same compromise set in a choice task. Further, we propose that and Lynn, 2014), while introducing important boundary conditions consumers will be less likely to prefer asymmetrically dominating of their occurrence. In this session, we collectively build on this work options in WTP than choice, because dominance should not influence by advancing novel factors that drive the occurrence of context ef- price perceptions. In two pre-registered studies, we test our predic- fects, ad demonstrate entirely new context effects. tions using classic compromise and attraction effect paradigms. Shennib, Evangelidis, and Jung investigate the occurrence of In Study 1, participants (N=400) were randomly assigned to one context effects in two common elicitation procedures: choice and of four conditions in a 2 (elicitation: Choice vs WTP) x 2 (options: willingness-to-pay. They demonstrate that attraction and compromise AB vs ABC) between-subjects design. Participants in the AB condi- effects replicate in choice, but are eliminated (attraction) or reversed tions made a choice between two options or reported the maximum (compromise) in willingness-to-pay because the latter is predomi- price they would be WTP for both options, separately. Similarly, par- nantly driven by price perceptions that are stable across contexts. ticipants in the ABC conditions made one choice among three op- Shaddy, Fishbach, and Simonson find that identity expression tions, which included the same AB options in addition to a C option, systematically attenuates variety seeking, the compromise effect, and or expressed their WTP for all options, separately. Prior literature balancing (vs. highlighting) between goals. The authors identify a has demonstrated greater choice preferences for the compromise/ mechanism: that, when it comes to identity, consumers want to avoid middle option in the ABC set relative to the same option in the AB sending mixed signals. Identify expression, therefore, influences set (Simonson, 1989). To create comparable metrics between choice consumers’ tradeoffs by reducing the appeal of mixed solutions, and WTP, we converted WTP into a binary (AB) or trinary (ABC) which partially satisfies multiple considerations or goals. preference indicator, such that, in each set, preferences were coded Evangelidis, Levav, and Simonson demonstrate a novel context as the option that participants were WTP most for. Preferences in the effect, symmetric dominance, whereby preference between a high- choice conditions were participants’ choices of one option in the set. quality high-price option and a low-quality low-price alternative can Participants in each condition responded to two product sets, be influenced by the addition of a decoy option that is dominated randomly ordered: BBQ grills and flashlights. Each product included by both alternatives (i.e., a low-quality high-price option). They two attributes (e.g., a flashlight’s brightness and battery life). One show that this effect is observed because consumers will feel better attribute increased in superiority while the other attribute decreased.

Advances in Consumer Research 949 Volume 48, ©2020 950 / Novel Effects of Context on Preference Formation Results showed that, in choice, participants selected the com- Invoking Identity Changes How Consumers Resolve promise options more so in the ABC sets (61.3%) than the AB sets Tradeoffs (58.6%). In WTP, however, participants preferred the compromise option strikingly less in the ABC sets (9.9%) than the AB sets EXTENDED ABSTRACT (82.0%) resulting in a highly significant interaction (β = -3.85, SE = .41, Z = -9.35, p < .001). That is, in the ABC WTP conditions, “We are what we repeatedly do.” participants were rarely WTP most for the middle options despite –Aristotle (allegedly) their ABC choice counterparts preferring this option more than 60% of the time. Marketers frequently invoke identity. For example, the makers Study 2 (N = 546) was identical to Study 1 except that we also of Jim Beam bourbon declare: “Guys never change. Neither do we.” tested (a) the attraction effect and (b) two mechanisms underlying The fast food restaurant Wendy’s explains: “Paul does the Paul burg- the effects in WTP and choice. In the attraction effect, options A and er, not a generic John Doe burger—don’t compromise.” Mercedes- B are the same as in compromise sets, but C is known as a “decoy” Benz claims its SLK roadster is “as extreme as you.” Is it simply that is asymmetrically dominated by—and thus increases prefer- coincidental that these references to specific identities (e.g., “guys”) ence for—one of the two options (Huber et al. 1982). Participants are often paired with appeals like “never change”? Or might invok- responded to four randomized products: two compromise sets that ing identity actually increase the likelihood that consumers will re- included BBQ grills and headphones and two attraction sets that infuse to “compromise” and instead choose to be “extreme”? cluded water bottles and flashlights. In this research, we answer this question by documenting how For the compromise sets, we replicated the interaction from identity expression systematically changes the way consumers re- Study 1 (β = 1.76, SE = .31, Z = 5.67, p < .001). Regarding the at- solve tradeoffs. Specifically, we find that it reduces the appeal of traction sets, in choice, participants selected the dominating options mixed solutions—outcomes that partially satisfy multiple consider- more so in the ABC set (80.1%) than the AB set (57.7%). In WTP, ations (e.g., tastes, attributes, goals, etc.). This is because when peo- however, participants valued the dominating option less in the ABC ple view decisions as expressions of identity, they are uncomfortable set (40.5%) than the AB set (57.6%). The interaction was highly sig- sending mixed signals—both to themselves and to others (Berger nificant (β = 1.76, SE = .30,Z = 5.87, p < .001). and Heath 2007; Rifkin and Etkin 2019; White and Dahl 2006). We then examined the mechanisms underlying the observed We explore the consequences shifting consumers’ tradeoff preference reversals. We hypothesize that WTP amounts are driven resolution strategies by examining several seemingly unrelated and by perceived prices, explaining higher WTP amounts for extreme well-known choice effects. Variety seeking, the compromise effect, options when there is one dominating attribute the set. In choice, and balancing between goals all represent outcomes that partially however, we hypothesize that individuals select options that rely satisfy multiple considerations—competing tastes, in the case of va- more so on subjective attribute weights. We tested our predictions riety seeking (Kahn 1995; Simonson 1990); competing attributes, in by having participants select the option in each set that they believed the case of the compromise effect (Simonson 1989); and competing was the most expensive and to rate how important each attribute was goals, in the case of balancing (Fishbach and Dhar 2005). Conse- in determining their responses from 1(not important at all) to 7(ex- quently, if invoking identity reduces the appeal of mixed solutions— tremely important). because people are uncomfortable sending mixed signals—then it Across the compromise and attraction sets, separately, we found should systematically attenuate all three. that, in WTP, preferences were significantly more associated with Importantly, while previous research has documented the vari- price perceptions than in choice. Put differently, those who were ous ways in which identity changes what people choose—someone WTP most for the compromise/attractive or extreme/non-dominating with a “rugged” identity, for example, might opt for a similarly option were also more likely to perceive that option to be the most “rugged” Jeep (Aaker 1997; Belk 1988; Escalas and Bettman 2005; expensive, whereas in choice, preferences were unrelated to price LaBeouf, Shafir, and Bayuk 2010; Reed 2004)—we examine the perceptions (Elicitation X Price Perception in compromise sets: β = way in which identity changes how people choose. 1.37, SE = .58, Z = 2.37, p = .018; in attraction sets: β = 1.67, SE = Seven studies (N=2,213) tested this account. In S1, a field study, .38, Z = 4.43, p < .001). we offered two free snacks to passers-by in a university student For the attribute weights mechanism, the results were, again, union: potato chips (indulgent) and pea crisps (healthy). Participants consistent with our predictions: in choice, attribute weight scores could choose one of each or two of the same snacks, and we manipu- predicted preferences for the compromise/attractive options. How- lated the contents of the signs advertising the snacks. The identity- ever, in WTP, preferences were unrelated to attribute weights. Put invoking sign read: “Who are you? I am a health-conscious snacker” simply, in choice, individuals factor in the attributes and their levels vs. “I am an indulgent fun-loving snacker.” The control sign read: in the option-sets significantly more so than in WTP (Elicitation X “Which snacks will you choose?” Research assistants surreptitiously Attribute Weights in compromise sets: β = .97, SE = .25, Z = 3.91, recorded actual choices, observing less variety seeking (i.e., choice p < .001; in attraction sets: β = -.27, SE = .12, z = -2.30, p = .021). of one of each snack) in the identity condition (26% vs. 43%; b=– Our results highlight that the extent to which consumers rely .94, SE=.33, z=–2.82, p=.005). on context is contingent on the elicitation procedure. To researchers In S2–4 (all preregistered), we modified a single basic scenario and practitioners, this suggests using precaution when selecting a (e.g., choosing a car) to test all three choice effects, predicting that method to elicit consumer preferences, as the generalizability of the prompting participants to express their identity would attenuate va- results may vary as a function of the elicitation procedure. Behav- riety seeking (S2), the compromise effect (S3), and balancing (S4). ioral researchers have studied context effects primarily using choice In S2, participants were less likely to rent different cars on procedures. Given the normative assumption of procedural invari- each of two different weekends (i.e., seek variety) when prompted ance, one would expect context effects to replicate across other pro- to consider their identity (e.g., “Which type of car best reflects your 2 cedures, yet our data pose a challenge to this assumption. identity?”; 17% vs. 46%; χ (1)=19.34, p<.001). In S3, participants were similarly less likely to choose the middle option when purchas- Advances in Consumer Research (Volume 48) / 951 ing one of three cars when prompted to consider their identity (e.g., introduction of the symmetric decoy should not influence relative “Which car…best reflects your identity?”; 47% vs. 64%; 2χ (1)=5.38, preference because the decoy is dominated by both alternatives. For p=.020). Finally, in S4, participants imagined having recently saved example, consider the following experiment. In the control condi- money. Those who reflected on their identity (e.g., “Does [saving tion, participants are asked to choose between a 1TB hard drive that money] make you feel like the kind of person who is financially re- costs $40 and a 2TB drive that costs $80. In the experimental condi- sponsible?”) were less likely to subsequently spend extra money up- tion, participants are presented with the same two options, as well grading a rental car (i.e., balance between saving and spending goals; as a symmetrically dominated option: a 1TB drive priced at $80. M=2.32 vs. M=3.35; F(1, 398)=12.39, p<.001). Here, the decoy is dominated by both of the original alternatives. It We next designed S5–7 to test for process. Specifically, we mea- has the same capacity as the 1TB drive but is more expensive, while sured participants’ tradeoff resolution strategies directly with a scale it has the same price but less capacity compared to the 2TB drive. that we developed. Participants indicated agreement or disagreement Although the symmetric decoy does not seem to favor one of the (e.g., “To what extent does each statement below describe how you dominating alternatives over the other, we posit that it can actually made your choice?”) with statements like: “try to get a little bit of exert a systematic impact on preference. In the absence of the decoy everything,” “set priorities and go with the top priority,” and “find option, individuals may covet the 2TB drive, but may find it difficult the middle ground.” We predicted that the reduced appeal of mixed to justify paying $40 more (vs. the 1TB drive). Consequently, indi- solutions (as measured by our scale) would mediate the attenuating viduals may defer choice. However, the introduction of the decoy effect of invoking identity on variety seeking, the compromise effect, provides a compelling reason for choosing the 2TB drive: compared and balancing. to the decoy, choice of the 2TB implies getting more capacity with- In S5, participants were less likely to vacation in different coun- out incurring additional costs. Thus, we argue that consumers, who tries each of two different years (i.e., seek variety) when prompted have an innate preference for the 2TB drive may find it easier to to consider their identity (e.g., “These trips will help define and rein- justify its purchase—and thus select that option more frequently— force your identity”; 58% vs. 78%; χ2(1)=19.31, p<.001). In S6, par- when choosing from the trinary compared to the binary set. Symmet- ticipants were similarly less likely to choose the middle option when ric dominance may seem rather unintuitive because, in theory, the renting one of three apartments when prompted to consider their presence of the decoy provides a reason for choosing either one of identity (e.g., “If you were trying to communicate [your] identity… the two dominating options. To allow for symmetric dominance, we which apartment would you choose?”; 42% vs. 55%; χ2(1)=6.40, posit that consumers will be more likely to rely on the comparison p=.011). Finally, in S7, participants imagined having recently helped between the high-quality high-price option (e.g., the 2TB drive) and clean a park (i.e., making progress toward an environmental goal). the decoy compared to the comparison between the low-quality low- Those who viewed environmental consciousness as part of their price option (e.g., the 1TB drive) and the decoy when choosing from identity were less likely to subsequently purchase an environmen- the trinary set. In turn, the symmetrically dominated decoy option tally unfriendly, but cheaper and more effective cleaning spray (i.e., will provide an asymmetric advantage to the high-quality high-price balance between hedonic and utilitarian goals; M=2.93 vs. M=3.44; option over the low-quality low-price competitor. Consumers will t(397)=2.57, p=.011). Moreover, in all three of S5–7, the reduced feel better about—and anticipate lower regret from—choosing the appeal of mixed solutions mediated the attenuating effect of identity high-quality high-price option in the presence (vs. absence) of the expression on variety seeking, the compromise effect, and balancing. decoy. This work (1) contributes a new insight to the literature expli- We provide evidence for symmetric dominance in five well- cating the role of identity in consumer decision making, (2) provides powered preregistered experimental studies (total N = 5,888). Study a unique theoretical lens with which to reconcile potentially compet- 1 shows robust evidence for the effects using five different sets of ing predictions for when opposite choice effects will arise (e.g., vari- stimuli: backpacks, Bluetooth speakers, external hard drives, hotels, ety seeking vs. consistency seeking), and (3) illuminates the psycho- and TVs. Participants were more likely to select the high-quality logical processes governing tradeoff resolution, more broadly. And high-price options when the symmetrically dominated decoys were with respect to marketing practice, we offer that managers would be added to the choice sets compared to the control two-option condi- wise to use these findings to strategically attenuate or amplify these tions (p < .001). Further, participants were less likely to defer choice choice effects in the marketplace—especially given that our results (i.e., search for other options) when the symmetrically dominated suggest many well-known decision phenomena may not be as uni- options were introduced to the set compared to the two-option condi- versal (or unique) as previously assumed. tions (p < .001). Both effects were robust across stimuli. Study 2 provides further evidence of the effect using actual Symmetric Dominance purchase decisions of laptop bags. Participants were more likely to select a high-quality high-price laptop bag when a symmetrically EXTENDED ABSTRACT dominated decoy bag was added to the choice set compared to the Decades of research in marketing and psychology corroborate control two-option condition (p = .048). Further, participants were that—contrary to normative theory—preference for a given option less likely to defer choice (i.e., not purchase a bag) when the decoy is contingent on the decision context. Perhaps the most prominent option was introduced to the set compared to the two-option condi- empirical finding from this research is the asymmetric dominance tion (p = .014). effect (Huber, Payne, & Puto, 1982). Asymmetric dominance de- Study 3 provides evidence for the process underlying the ef- scribes the finding that adding a decoy option that is dominated by fects. First, we replicate our basic results (p < .006). Further, we only one of two alternatives in a binary set increases the choice share show that our effect is mediated by feelings of regret (ab = .053, of the dominating (target) option. In this paper, we demonstrate an- 91% LLCI = .001, ULCI = .124) rather than changes in the impor- other context effect—symmetric dominance—that has surprisingly tance that participants afford to the different attributes (ab = -.029, eluded earlier research. Symmetric dominance pertains to the case 91% LLCI = -.224, ULCI = .165). In the presence (vs. absence) of of adding a symmetrically dominated option to a binary set char- the decoy, participants anticipate weaker feelings of regret from the acterized by a tradeoff between price and quality. Intuitively, the choice of the high-quality high-price option relative to the choice of 952 / Novel Effects of Context on Preference Formation the low-quality low-price alternative. In turn, the decrease in antici- each brand, participants were shown binary trials with identically pated regret leads to a higher incidence of choice of the high-quality sized vertical and horizontal packages and ternary trials in which a high-price option. Our process evidence suggests that consumers de- decoy option was added, asymmetrically dominated by one of the sire, but are reluctant to choose the high-quality high-price option packages. The order of options from left to right, their orientations, in the absence of the decoy. However, in the presence of the decoy, and colors were all randomized. The attraction effect would be ob- consumers tend to use that option in order to support choice of the served if the probability of selecting the target option in a ternary desired high-quality high-price alternative, leading to lower feelings trial was significantly greater than the probability of selecting the of anticipated regret. same option in a binary trial. Further studies provide evidence for two moderators of the ef- Results from Experiment 1 (N = 376) revealed an overall fect. Study 4 shows that the symmetric dominance effect is larger in attraction effect for size judgments among visually presented magnitude when the decoy is located next to the high-quality high- consumer goods packages (χ²(1, N = 2107) = 20.20, p < 0.001, d price option (p < .001) compared to when it is located next to the = 0.11). low-price low-quality alternative (p < .001; interaction p = .013). This was generally consistent across brands except for the dif- Presumably, it is easier for people to rely on the comparison between ferently shaped Chips Ahoy packages, where the addition of a decoy the high-quality high-price option and the decoy when the two op- decreased the choice share of the associated target option (χ²(1, N = tions are adjacent to each other. Finally, Study 5 shows that sym- 266) = 6.33, p = 0.012, d = -0.19). Importantly, analyses revealed that metric dominance replicates when there is a trade-off between price the attraction effect, across all brands aside from Chips Ahoy, was and quality (p < .001), but is eliminated when the two focal options only observed when the target and decoy options were vertical (χ²(1, present a trade-off on two attributes, both of which are associated N = 923) = 76.77, p < 0.001, d = 0.49) but not horizontal (χ²(1, N = with product quality (p = .187; interaction p = .055). Thus, symmet- 918) = 1.29, p = 0.26, d = -0.05). We also found a strong horizontal ric dominance is more likely to manifest in consumer decisions that bias, such that people viewed the horizontal package as larger than involve price-quality tradeoffs. the equally-sized vertical package the majority (77%) of the time. Overall, our work has important implications for decision-mak- Thus, the attraction effect being moderated by target package orien- ing theory because no extant theoretical framework or model can tation is consistent with Evangelidis, Levav, and Simonson (2018), accommodate these effects. which finds stronger attraction effects when the decoy is dominated by the less preferred option. The Attraction Effect in Product Size Judgments In experiment 2 (N = 278), we replicated the visual package-size attraction effect in an in-person, paper and pencil survey in which EXTENDED ABSTRACT participants selected the largest Cheez-Its box from either a binary Marketers attempt to present products so that they appear more set or from a ternary set with a decoy, which increased selection of attractive than competing products. Context effects have been widely the target option as larger (t(274.89) = 3.49, p < 0.001, d = 0.42). studied as influencing the relative appeal of consumer products. In In Experiment 3 (N = 352), we systematically varied the particular, the attraction effect occurs when a decoy option that is package dimensions and orientation as potential moderators of the asymmetrically dominated by a target option increases the choice attraction effect. Participants judged the largest in binary or ternary share of that dominating target option when the decoy is added to a sets of packages, with eight different dimensions, ranging from binary choice set (Huber, Payne, and Puto, 1982; Huber and Puto, near square to highly elongated (e.g., as in the Chips Ahoy boxes in 1983). While the attraction effect has been observed in a large range Experiment 1). In each binary pair, the two packages (one vertical, of contexts when stimuli are depicted as stylized or numeric (Tver- one horizontal) had the same surface area. In the ternary sets, a decoy sky, 1972; Pan, O’Curry, and Pitts, 1995; Simonson, 1989; Huber package was added. We again found an overall attraction effect (χ²(1, et al.,1982), the robustness of the attraction effect with perceptual N = 1994) = 34.47, p < 0.001, d = 0.15), significant when the target stimuli has recently been called into question (Frederick et al., 2014; was vertical (χ²(1, N = 979) = 49.31, p < 0.001, d = 0.28) but not Yang and Lynn, 2014). Across four pre-registered studies, including when the target was horizontal (χ²(1, N = 1015) = 0.65, p = 0.422, paper and pencil and online studies, we demonstrate that the attrac- d = 0.03), as in Experiment 1. However, the degree of elongation of tion effect is observed when making size judgments between visu- the packages did not moderate the attraction effect, either as a main ally presented product packages and when choosing between equally effect (β = -0.07, p = 0.800) or in a non-linear relationship (β = 0.04, priced products. We also identify important factors influencing the p = 0.290). Results were robust to different types of decoys (e.g., robustness of the effect, particularly the importance of vertical vs. range vs. frequency). horizontal package orientation. In Experiment 4 (N = 446), we extended our findings from size In our first experiment, building on recent research judgments to product choices in which size was the primary basis demonstrating context effects with geometric shapes (Trueblood et for decision. We asked participants to choose which hypothetical al., 2013), we tested whether the attraction effect is observed when package of crackers to purchase (from either the binary set or ternary making size judgments between product packages. Front-facing set), with all options priced equally at $3.29. Participants made mock product packages were created for eight different brands; Ba- multiple choices, varying the elongation of the packages across rilla penne pasta, Ritz crackers, Cheezits crackers, Cheerios cereal, choice sets as in Experiment 3. We observed an overall attraction Chips Ahoy cookies, Jell-O mix, Domino sugar, and Tootsie Pops. effect (χ²(1, N = 2166) = 20.49, p = < 0.001, d = 0.10) which was For each brand, equal-area horizontally and vertically oriented boxes significant for both vertical trials (χ²(1, N = 1082) = 16.50, p = < were created, which had the same dimensions with width and height 0.001, d = 0.13) and horizontal trials (χ²(1, N = 1084) = 5.10, p = reversed (except for Chips Ahoy). 0.02, d = 0.07), albeit with small effect sizes when the target was A range-decoy image was then created for each package. Verti- horizontal. cal range decoys were less wide than vertical targets, and horizontal Overall, these findings demonstrate that the attraction effect oc- decoys were shorter than horizontal targets. Each image was also curs for visual stimuli, affecting both prompted size judgments and produced in an alternative color scheme, not used in the market. For choices among equally priced options. These findings provide robust Advances in Consumer Research (Volume 48) / 953 support for the practical application of the attraction effect in con- Simonson, Itamar (1989), “Choice Based on Reasons: The Case sumer settings. of Attraction and Compromise Effects,” Journal of Consumer Research, 16 (2), 158–174. REFERENCES Simonson, Itamar (1990), “The Effect of Purchase Quantity and Timing on Variety-Seeking Behavior,” Journal of Marketing The Contingent Impact of Context on Choice versus Research, 27 (2), 150–162. Valuation White, Katherine, and Darren W. Dahl (2006), “To Be or Not Be? Huber, J., Payne, J. W., & Puto, C. (1982). Adding asymmetrically The Influence of Dissociative Reference Groups on Consumer dominated alternatives: Violations of regularity and the Preferences,” Journal of Consumer Psychology, 16 (4), similarity hypothesis. Journal of Consumer Research, 9(1), 404–414. 90-98. Simonson, I. (1989). Choice Based on Reasons: The Case of Symmetric Dominance Attraction and Compromise Effects.Journal of Consumer Huber, Joel, John W. Payne, and Christopher Puto (1982), Research, 16(2), 158-174. “Adding Asymmetrically Dominated Alternatives: Violations Tversky, A., Sattath, S., and Slavic, P. (1988). Contingent weighting of Regularity and the Similarity Hypothesis,” Journal of in judgment and choice. Psychological Review, 95, 371–384. Consumer Research, 9 (June), 90-8.

The Motivating Role of Disassociative Out-groups in The Attraction Effect in Product Size Judgments Encouraging Sustainable Consumer Behavior Evangelidis, I., Levav, J., & Simonson, I. (2018). “The asymmetric Aaker, Jennifer L. (1999), “The Malleable Self: The Role of Self- impact of context on advantaged versus disadvantaged Expression in Persuasion,” Journal of Marketing Research, 36 options.” Journal of Marketing Research, 55 (2), 239-253. (1), 45–57. Frederick, S., Lee, L., & Baskin, E. (2014) “The Limits of Belk, Russell W. (1988), “Possessions and the Extended Self,” Attraction,” Journal of Marketing Research, 51 (4), 487-507. Journal of Consumer Research, 15 (2), 139–168 Huber, J., Payne, J. W., & Puto, C. (1982). “Adding asymmetrically Berger, Jonah, and Chip Heath (2007), “Where Consumers Diverge dominated alternatives: Violations of regularity and the From Others: Identity Signaling and Product Domains,” similarity hypothesis.” Journal of Consumer Research, 9, Journal of Consumer Research, 34 (2), 121–134. 90–98. Escalas, Jennifer Edson, and James R. Bettman (2005), “Self- Huber, J., & Christopher, P. (1983), “Market Boundaries and Construal, Reference Groups, and Brand Meaning,” Journal of Product Choice: Illustrating Attraction and Substitution Consumer Research, 32 (3), 378–389. Effects,”Journal of Consumer Research, 10 (1), 31-44. Fishbach, Ayelet, and Ravi Dhar (2005), “Goals as Excuses or Pan, Y., O’Curry, S., & Pitts, R. (1995). “The attraction effect Guides: The Liberating Effect of Perceived Goal Progress on and political choice in two elections.” Journal of Consumer Choice,” Journal of Consumer Research, 32 (3), 370–377. Psychology, 4, 85–101. Kahn, Barbara E. (1995), “Consumer Variety-Seeking Among Simonson, I. (1989). Choice based on reasons: “The case of Goods and Services: An Integrative Review,” Journal of attraction and compromise effects.”Journal of Consumer Retailing and Consumer Services, 2 (3), 139–148. Research, 16, 158–174. LeBoeuf, Robyn A., Eldar Shafir, and Julia Belyavsky Bayuk Trueblood, J. S., Drown, S., Heathcote, A., & Busemeyer, J. (2010), “The Conflicting Choices of Alternating Selves,” R. (2013), “Not just for consumers: Context effects are Organizational Behavior and Human Decision Processes, 111 fundamental to decision-making,” Psychological Science, 24 (1), 48–61. (4), 1-8 Reed, Americus (2004), “Activating the Self-Importance of Tversky, A. (1972), “Elimination by aspects: A theory of choice,” Consumer Selves: Exploring Identity Salience Effects on Psychological Review, 79 (July): 281-99. Judgments,” Journal of Consumer Research, 31 (2), 286–295. Yang, S. and Lynn, M. (2014). “More evidence challenging the Rifkin, Jacqueline R., and Jordan Etkin (2019), “Variety in robustness and usefulness of the attraction effect,”Journal of Self-Expression Undermines Self-Continuity,” Journal of Marketing Research, 51 (4), 508-513. Consumer Research, 46 (4), 725–749. From “Me” to “Vous”: Language and Culture in Consumer Behavior Chairs: Ann Kronrod, University of Massachusetts, Lowell, USA Ruth Pogacar, University of Calgary, Canada

Paper #1: Brand Names and Country Brand Personality DNA a broad audience and to spark an interesting rendez-vous among at- Ella Karat, University of Massachusetts, Lowell, USA tendees interested in language, culture, marketing communication, Ann Kronrod, University of Massachusetts, Lowell, USA advertising, and branding. Paper #2: Is Nestlé a Lady? Brand Name Gender and Loyalty Brand Names and Country Brand Personality DNA Attitudes Ruth Pogacar, University of Calgary, Canada EXTENDED ABSTRACT Justin Angle, University of Montana, USA This work introduces “Country Personality DNA” – the unique L. J. Shrum, HEC Paris, France combination of human characteristics that are associated with a coun- Frank Kardes, University of Cincinnati, USA try – and explores the way it influences perceptions of new brands Tina M. Lowrey, HEC Paris, France whose names sound typical to a specific country. Drawing from lit- Paper #3: Precisely Unique: Semantic Precision Increases erature on branding, country of origin and country stereotypes, the Perceived Uniqueness of Cultural Products work tests how spelling a brand name in a foreign language triggers Vincent Xie, University of Massachusetts, Boston, USA associations with that country, and more specifically, with the spe- Ann Kronrod, University of Massachusetts, Lowell, USA cific set of human personality traits that belongs to that country. In this study, we treat “Country Personality DNA” as a meta- SESSION OVERVIEW phor for a country’s hereditary brand personality. Through this lens, This session embraces the rendez-vous conference theme by we examine the way countries with high brand equity, such as France, examining the role of language and culture in consumer behavior. Germany or Japan, can pass on their recognizable personality traits Each paper takes a slightly different focus—on consumer percep- to new brands with names that sound as if they originated from these tions, evaluations, or decisions—using experimental and field data countries. This is done through the process of instant activation of to provide a rich understanding of the following topics: How does stereotypical personality traits associated with these countries. Based country personality influence brand perceptions via cultural stereo- on these logical links we first theorize that countries have unique and types? Why do brand names ending in vowels trigger cultural asso- identifiable combinations of humanlike personality traits (Country ciations with gender and warmth, increasing loyalty attitudes? And Personality DNA) in people’s perceptions. Further, following the why does semantic precision increase perceived uniqueness of cul- DNA metaphor, we suggest that the country acts as a “parent” to the tural products? “child” brand, and the child brand inherits this unique set of person- Karat and Kronrod introduce Country Personality DNA—the ality traits, which we call the “core genes”. unique combination of human traits that are associated with a coun- “Core genes” are the country’s highest ranking four personality try—and explore how it influences perceptions of brands whose traits. Just like the four genes that make up the human DNA, these names sound typical of a specific country. Their results suggest that genes have the strongest association with the “parent” country, and countries possess a distinct set of four “core gene” personality traits, have the highest potential of being inherited by the “child” brand. and that new brands whose names sound like they originated from Based on this reasoning, we next predict that a brand name that a particular country “inherit” these traits. This topic is of particular sounds like it originated from a particular country should “inherit” importance given the growing global marketplace. that country’s personality DNA — its unique combination of per- Pogacar, Angle, Shrum, Kardes, and Lowrey focus on the cul- sonality traits—and consequently consumers associate similar traits tural effect of gender perceptions activated by brand names. The au- with that brand. Thus, we predict that new brand names that resemble thors demonstrate that brand name length, stress, and vowel ending a particular country-of-origin language—even if they are fictitious— influence perceptions of brand name gender. Feminine name gender carry the same perceived “genetic make-up” (DNA traits) of their in turn leads to cultural inferences of warmth, which results in en- parent country’s brand. hanced loyalty. The authors use both experimental and observational data, real and invented brands, and consequential and hypothetical Method choices. These findings have important implications for naming In our studies we focused on two of the G7 European countries products that should seem ‘warm’ (and those that shouldn’t). (France, Germany) and one East Asian country (Japan), mainly be- Lastly, Xie and Kronrod’s work reveals the novel link between cause they are well-known countries which are similar to each other semantic precision and perceptions of uniqueness of cultural prod- in terms of industrial and economic growth (Laš, 2018). In addition, ucts. The mere presence of more precise language, even about non- these countries have identifiable different languages. Study 1 identi- central product attributes, influences the perceived cognitive distance fies the ‘core genes’ that make up the unique combination of country of the words, and increases perceived product uniqueness for cultural personality DNA for France, Germany and Japan. We first selected products like international tourism. This finding is relevant given the 29 personality traits from a pool identified in studies by Aaker et demand for unique consumer experiences, and the saturation of com- al., (2001), and Ferrandi et al., (2000) and Bosnjak et al., (2007). peting products vying to meet these demands. These traits were: Sincere, Kind, Spirited, Optimistic, Independent, Together the papers in this session highlight the important role Reliable, Intelligent, Confident, Glamorous, Charming, Sophisti- of language and culture in consumer behavior. Each work addresses cated, Good-looking, Shy, Sentimental, Affectionate, Tender, Genu- a unique aspect of this topic: foreign language brand names; brand ine, Stable, Aggressive, Tough, Elegant, Romantic, Peaceful, Naïve, names and gender associations; language precision and cultural Dependent, Dignified, Funny, Happy, Energetic. Next, 305 MTurk products. Given the array of questions addressed, we expect to draw participants ranked these 29 personality traits on the extent to which

Advances in Consumer Research 954 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 955 each of them described each of the 3 countries (France, Germany masculine names tend to have fewer syllables and end in a consonant and Japan) if they were a person (1 = strongly disagree; 7 = strongly (e.g., Brad, Ford). agree). Each participants ranked one country, in a 3-cell between We examine when brand names with feminine versus masculine subjects design. Results of Factor Analyses suggest that each of linguistic characteristics are advantageous. Beyond simply matching the three countries has a unique set of four ‘core genes’ in people’s brand name gender to product category or user, weposit a process minds. We defined this set as the country’s Personality DNA. The based on the Stereotype Content Model (Fiske et al., 2002) by which identified ‘core genes’ were: feminine brand names convey warmth (tolerance, good-naturedness, France: glamorous, sophisticated, elegant, and romantic; friendliness, sincerity). Because attributes like warmth are primary Germany they are: independent, intelligent, confident, and drivers of loyalty attitudes (Chaudhury and Holbrook 2001) we pro- tough; pose a link between brand names with feminine linguistic attributes, Japan: reliable, intelligent, stable and dignified. perceived warmth, and loyalty attitudes such that feminine brand names will enhance perceived warmth, which will enhance brand Study 2 demonstrates that these ‘core genes’ are ‘inherited’ by loyalty attitudes. brands whose names sound original to the associated country. First, a We test these predictions in seven studies, varying types of pretest was conducted in order to identify fictitious brand names that brands (real, invented, Interbrand top brands), methodology (obser- are correctly identified with France, Germany and Japan. 14 brand vational, experimental), stakes (hypothetical, economic, temporal), names were tested against 17 countries (Albania, Bulgaria, China, and participant population (MTurk, college students). In Study 1 we Egypt, France, Germany, Greece, Italy, Japan, Mexico, Netherlands, investigate whether brand name femininity is positively related to Portugal, Romania, Russia, Spain, Turkey and USA). The fictitious loyalty-based brand performance by examining the linguistic charac- names, included filler names from Turkish, Greek and Russian, teristics of a real-world sample of Interbrand top brand names. Study looked and sounded like they came from a specific country of origin 2 examines whether brand name gender influences actual product and were: Doğanoğlu, Μίαπαρέα, Bleusàur, L’Noir, Duprêté, Ne- choices with consequences for time (Study 2a) and money (Study mashi, Miyoshi, Takitomo Bäumtät, Müssler, Schaumbern, Süchtig, 2b). Study 3 tests the hypothesis that brand warmth mediates the Экрейс, Клятва. Based on the pretest, six brand names, that showed relationship between brand name gender and self-reported brand strong associations with France, Germany and Japan, were chosen loyalty using a sample of real-world brands. Study 4 tests the propo- (two for each country): L’Noir (81.2%), Duprêté (42.6%), Scha- sition that brand warmth mediates the effect of brand name gender on umbern (76.2%), Süchtig (63.4%) and Miyoshi (80.2%), Takitomo brand loyalty attitudes for hypothetical brands in a controlled experi- (82.2%). These names were chosen for the main study. ment. Studies 5 and 6 test theoretically relevant boundary conditions, In the main study, 392 MTurk participants ranked the 29 human showing that the feminine brand name advantage is moderated by personality traits from study 1 on the extent to which they described product category (hedonic vs. utilitarian; Study 5) and typical user each of the six brand names. Subsequently, participants guessed the gender (men vs. women; Study 6). country of origin for each of the six names. 280 participants guessed Study 1 tested the proposed association between feminine brand the country correctly, and analyses were conducted with this sub- name linguistics and brand loyalty attitudes using a sample of Inter- sample. Overall the results showed, as predicted, that there is a sig- brand’s “Global Top Brands,” which are designated, in part, based nificant overlap between the four traits of country personality DNA on consumer loyalty. We coded 171 of Interbrand’s top brands for a for each of the three focal countries, and the human traits that partici- twenty-year period from 2000-2019 using a validated name gender pants assigned to each of the brand names. Factor analyses showed normative scale that quantifies the degree of linguistic masculinity similar factor loadings for the brand names, as we found for their or femininity from -2 (most masculine) to +2 (most feminine; Barry “parent” countries. The means were strikingly similar. and Harper, 1980). Interbrand top brands have, on average, feminine To sum, we show that when brand names are recognized as names (p < .001); feminine names gender is positively associated being from a particular country, they evoke perceptions associated with average rank on the Interbrand list (p = .01); and feminine brand with humanlike personality traits associated with that country. This names spent more years on the list than did masculine names (p = study is the first to identify inheritable country personality traits, and .008). Finally, comparing the Interbrand names with 171 randomly link them to branding. Theoretically, the findings contribute to our selected companies from Thomson Reuter’s Eikon database, and understanding of the way brand names with languages which sug- controlling for net income, net revenue, years in business, and num- gest country origin, influence perceptions of brand origin, in return ber of employees, feminine brand name gender predicted Interbrands influence brand perceptions, via country stereotyping. As a market- top brand status (p = .02). ing implication, utilizing the new Country Personality DNA model In Studies 2a and 2b we explore whether brand name gender allows marketers, brand managers and entrepreneurs to create brand influences product choices with real-world consequences. Study names that can instantly signal brand personality – which is espe- 2a used a common consumer domain—streaming videos (N = 300 cially important for new brands. mTurkers). Participants were told they would be evaluating a short Is Nestlé a Lady? Brand Name Gender and Loyalty video produced by one of two channels and then chose whether to Attitudes spend their time watching a video from a feminine-named channel (Nimilia) or a masculine-named channel (Nimeld). A significant ma- EXTENDED ABSTRACT jority chose to watching a video from the feminine-named “Nimilia There are many ways a name can convey information about Channel” (71% versus Nimeld: 29%, p < .001). Study 2b replicates a brand, including semantics (Klink, 2001) and individual sounds these results with economic consequences by offering participants (Lowrey & Shrum 2007). Brand names can also suggest gender. (N = 150 students) either monetary compensation or a choice of Names sound masculine or feminine due to length, sounds, and product: Nimilia or Nimeld brand hand sanitizer. Participants were stress (Barry and Harper 1995). Specifically, feminine names tend to significantly more likely to choose the feminine-named product than be longer, with stress on the second or later syllable, and often end any other option (49% Nimilia hand sanitizer; 36% monetary com- in a vowel, especially the “a” sound (e.g., Lisa, Visa). Conversely, pensation; 14% Nimeld hand sanitizer; p < .001). 956 / From “Me” to “Vous”: Language and Culture in Consumer Behavior

Study 3 tested the hypothesized effect of brand name gender tic category to a lesser(greater) extent. Finally, Study 4 introduces a on perceived warmth and self-reported loyalty using a sample of moderator and finds the precise-unique heuristic is moderated when real brands. We elicited a sample of relevant brands and coded their a product is perceived as atypical by default. name gender using the scale from Study 1. We then asked a separate In Study 1 participants (n=162) first composed sentences from a sample from the same mTurk population (N = 517) to evaluate each list of 10 precise/non-precise words (e.g. bake/make, spotless/clean, brands’ warmth and report their loyalty toward each brand. Hayes cider/beverage). Next, participants indicated the category typicality PROCESS model 4 showed that feminine brand name gender pre- of each word to its corresponding category (“when you think of the dicts warmth (p = .03), which predicts brand loyalty (p < .001), and category, how likely are you to think of that particular word?”-1: not the mediating role of warmth is significant (95% CI = [.001, .066]). likely at all; 7: very likely). We used the mean of the ratings of the Study 4 provides causal evidence by experimentally manipulat- ten words (α=.91) as a composite measure of category typicality. Fi- ing brand name gender using a one-factor (name gender: feminine, nally, participants indicated how precise were the ten words on four masculine) between-subjects design and the same stimuli used in items (definite/exact/precise/specific, α=.91). Studies 2a and 2b. Participants (N = 250 mTurkers) evaluated the Results. The high-precision words were rated as significantly feminine brand name as warmer (p < .001); warmth increased brand more precise (M=5.33) than the low-precision words (M=4.90, loyalty attitudes (p < .001); and warmth mediated the effect of brand F(1,160)=4.06, p=.046). As predicted, the aggregated category typi- name gender on loyalty attitudes (95% CI = [.52, .91]). cality was significantly lower for the high-precision (M=5.00) than In Study 5 we tested the moderating role of product category the low precision words (M=5.32, F(1,160)=3.98, p=.048). These (hedonic vs. utilitarian) using a 2x2 between-subject experiment results provided evidence that more precise words elicit lower per- (N = 781 mTurkers) and the same names used in Studies 2a, 2b, ceived category typicality. and 4. We find that feminine brand names benefit hedonic products Study 2 (n=284) tested the effect of semantic precision on per- (chocolate) more than utilitarian products (bathroom scale; p = .02). ceived uniqueness. Participants read one of two ad texts promoting a This suggests that warmth is a more valued attribute for hedonic than t-shirt as imported from Europe/or Netherlands and rated the T-shirt utilitarian products. uniqueness (distinct/original/special/unique, α=.90) and ad precision Study 6 used a 2-factor between-subject design (N = 695 (α=.91). mTurkers) and expanded stimuli set (Woldard; Stelad; Wolda; Steda) Results. “Netherlands” was rated as significantly more precise to examine the moderating role of typical user gender. Results show (M=4.14, t(282)=6.63, p<.001) than “Europe” (M=3.01). As predict- an interaction between name gender and typical user gender (p = .03) ed, a T-shirt imported from the Netherlands was perceived as signifi- such that the feminine brand name advantage is neutralized when cantly more unique (M=3.09, t(282)=2.53, p=.01) than when it was the typical user is male (e.g., stout beer; p = .43) but intact when the from Europe (M=2.60). Study 2 supported the prediction that seman- typical user is female (e.g., strawberry daquiri; p < .001). tic precision of product descriptions increases perceived uniqueness This research has important implications for brand managers of cultural products. and best-naming practices, as well as our theoretical understanding Study 3 (n=200) examined the effect of semantic precision on of brand loyalty, warmth, and linguistics. consumer perception of product uniqueness driven by category typicality. Participants read an excerpt from a more/less precise review Precisely Unique: Semantic Precision Increases Perceived of a boat tour (e.g., “The dinner(meal) included soup(appetizer), Uniqueness of Cultural Products fish(entrée), and pudding(dessert)”). Participants rated the tour’s uniqueness, category typicality of each of the high(low) precision EXTENDED ABSTRACT words (e.g., pudding/dessert - dish), and how precise those twelve Consumers often value uniqueness of cultural products such words were as a whole (α=.91). as a folk artifact or a tour (Kasey 2019). We suggest that product Results. The review was rated as significantly more precise in uniqueness can be evoked via a “precise-unique” heuristic, whereby the high-precision condition (M=5.30) than the low-precision condi- more precise language increases perceptions of product uniqueness. tion (M=4.85, F(1,198)=6.73, p=.01). As predicted, category typical- Semantic precision is the extent to which the use of language con- ity was significantly lower in the high-precision condition (M=4.69) veys subject matters exactly as they are. We draw on the psycholin- than the low precision condition (M=5.33, F(1,198)=18.25, p<.001). guistic literature suggesting that more precise words (e.g., scarlet Perceived uniqueness of the high-precision (M=5.48) and low-preci- versus red) are less prototypical members of their category (e.g., col- sion (M=5.29) conditions did not significantly differ (F(1,198)=1.15, ors) (Rosch 1983; Rosch and Mervis 1975), and consumers are less p=.29). A mediation analysis (Hayes 2018) suggested that the effect likely to think of more precise words when they process product in- of semantic precision on perceived uniqueness was fully mediated formation within their semantic category (Hamby and Levine 2015). by perceived category typicality. Semantic precision was a signifi- We suggest that as a result, more precise words evoke perceptions cant predictor of category typicality (β=.65, p <.001), and category that the product is more unique. This effect can be salient even when typicality was a significant predictor of perceived uniqueness (β=.38, the words do not directly describe a product attribute. For example, p <.001). Importantly, semantic precision was a significant predic- a product review for a tour abroad that uses the word “robin” would tor of perceived uniqueness (β=-.43, p=.01). The mediation pathway elicit higher perceived uniqueness of the tour than if it uses the word from semantic precision to perceived uniqueness via category typi- “bird.” cality was significant (95% CI:[0.11,0.41]). Four experiments provide empirical support for our predictions. Study 3 provided evidence that semantic precision reduces cat- Study 1 shows that semantic precision evokes reduced perceived egory typicality, which in turn increases perceived product unique- category typicality. Study 2 tests the precise-unique heuristic in a ness. The insignificant direct effect of semantic precision- onper cultural product context. Study 3 demonstrates that cultural prod- ceived uniqueness can be attributed to the high inherent uniqueness ucts appear more(less) unique when their description is more(less) of the tour experience. Study 4 addressed this issue specifically. precise, and that the effect is mediated by category typicality such Study 4 (n=125) examined the moderation effect of product that the more(less) precise description resembles the same seman- inherent uniqueness on the “precise-unique” heuristic. Participants Advances in Consumer Research (Volume 48) / 957 were randomly assigned to one of four conditions representing a Ferrandi, J., Valette-Florence, P., & Fine-Falcy, S. (2000). Aaker’s 2 (more/less precise description) x2 (typical/atypical product) be- Brand Personality Scale in a French Context: A Replication tween-subjects design. They first rated perceived uniqueness of an and a Preliminary Test of Its Validity. Developments in ice-cream flavor (typical-Berry/Raspberry; atypical-Spice/Ginger). Marketing Science, 23, 7-13. Next, they rated perceived precision of the flavor word, and the ex- Fiske, Susan T., Amy J. C. Cuddy, Peter Glick, and Jun Xu (2002), tent to which the flavor was typical on three 7-point scale semantic “A Model of (Often Mixed) Stereotype Content: Competence differential items (atypical/typical, unusual/usual, uncommon/com- and Warmth Respectively Follow from Perceived Status and mon; α=.98). Competition,” Journal of Personality and Social Psychology, Results. We found two main effects: Berry flavors were rated 82 (6), 878-902. significantly more typical (M=4.32) than Spice flavors (M=2.13, Hamby, Tyler and Daniel S. Levine (2015), “Response-Scale F(1,123)=50.21, p<.001). Raspberry/Ginger was significantly Formats and Psychological Distances between Categories,” more precise (M=5.88) than Berry/Spice in describing the flavors Applied Psychological Measurement, 40(1), 73–75. (M=4.02, F(1,123)=54.93, p<.001). We also found a significant in- Kasey, Mya (2019), “Why Millennial Travelers Are Seeking More teraction between precision and typicality on perceived uniqueness Authentic Experiences,” Under30 Experiences. (available at (F(1,121)=5.32, p=.02). Within typical flavors, Raspberry (M=4.88) https://www.under30experiences.com/blog/why-millenial- was significantly more unique than Berry (M=3.48, F(1,60)=13.24, travelers-are-seeking-more-authentic-experiences, last p=.001). By contrast, for atypical flavors, Spice (M=5.40) and Gin- accessed 2/27/2020). ger (M=5.63) were not significantly different in perceived unique- Klink, R. R. (2001). Creating meaningful new brand names: ness (F(1,61)=.49, p=.49). Study 4 demonstrated further that seman- A study of semantics and sound symbolism. Journal of tic precision increases perceived product uniqueness, but also that Marketing Theory and Practice, 9(2), 27-34. the “precise-unique heuristic” is not salient for inherently atypical Laš, L. &. (2018). Japan’s Engagement in the Group of Seven. products. Journal of International Relations, 16(2), 160–182. To sum, our findings demonstrate the potential benefit of de- Lowrey, Tina M., and L. J. Shrum (2007), “Phonetic Symbolism scribing cultural products more precisely, as precise descriptions in- and Brand Name Preference,” Journal of Consumer Research, crease perceived uniqueness of the products. In a consumption world 34 (3), 406-414. valuing unique cultural experiences, the use of semantic precision Naylor, R. W., Lamberton, C. P., & Norton, D. A. (2011). Seeing can be an effective and costless tool to enhance consumer sense of ourselves in others: Reviewer ambiguity, egocentric anchoring, product uniqueness. and persuasion. Journal of Marketing Research, 48(3), 617- 631. REFERENCES Rosch, Eleanor (1983), “Prototype Classification and Logical Aaker, J. L., BenetMartínez, V., & Garolera, J. (2001). Classification: The Two Systems”, pp. 73–86 in Scholnick, Consumption symbols as carriers of culture: A study of E.K. (ed), New Trends in Conceptual Representation: Japanese and Spanish brand personality constucts. Journal of Challenges to Piaget’s Theory?, Lawrence Erlbaum Personality and Social Psychology, 81(3), 492-508. Associates, Hillsdale. Barry, Jr., Herbert, and Aylene S. Harper (1995), “Increased Choice Rosch, Eleanor and Carolyn B. Mervis (1975), “Family of Female Phonetic Attributes in First Names,” Sex Roles, 32 Resemblances: Studies in the Internal Structure of Categories”, (11/12), 809-819. Cognitive Psychology, 7(4), 573–605. Bosnjak, M. B. (2007). Dimensions of Brand Personality Attributions: A Person-Centric Approach in the German Cultural Context. Social Behavior & Personality: An International Journal, 35(3), 303-316. Consumer Decision-Making in Health Care Chairs: Sydney Scott, Washington University, USA Jonathan Zev Berman, London Business School, UK

Paper #1: The Costs of Autonomy: Decisional Autonomy Goksel, Faro and Puntoni (paper 3) examine how the cause of Undermines Advisees’ Judgments of Experts symptoms affects perceived severity of symptoms. The same physi- Samantha Kassirer, Northwestern University, USA cal symptom (e.g., a rash displayed in a photograph) is judged to Emma Levine, University of Chicago, USA be less severe and painful when people think it was caused by psy- Celia Gaertig, University of Pennsylvania, USA chological (vs. physiological) factors. Additionally, people are less likely to recommend a person seek medical help for a physical symp- Paper #2: Communication Preferences in Medical Advice tom when it was caused by a psychological (vs. physical) reason. Seeking Finally, Gershon, Lieberman and Scott (paper 4) examine one Selin Goksel, London Business School, UK reason why consumers might seek an illegal treatment. The authors Jonathan Zev Berman, London Business School, UK demonstrate an illegal=effective heuristic, where consumers think Sydney Scott, Washington University, USA the same drug will be more potent, strong and effective if it is illegal Paper #3: The Effect of Causal Explanations on Judgments of (vs. legal). Consumers even infer this when the drug is not legal for Medical Symptoms reasons unrelated to the safety and efficacy (e.g., a paperwork error). Selin Goksel, London Business School, UK Collectively, these papers provide new insights into health de- David Faro, London Business School, UK cision-making. In particular, these papers examine social intuitions Stefano Puntoni, Erasmus University Rotterdam, The and lay beliefs that might lead consumers to avoid going to a doctor, Netherlands dislike an interaction with their doctor, or even seek illegal treat- Paper #4: The Illegal = Effective Heuristic ments. Such insights contribute to theories in marketing, psychol- Rachel Gershon, University of California San Diego, USA ogy, and medical decision-making, among other fields, and therefore Alicea Lieberman, University of California San Diego, USA should be of interest to a broad audience of scientists. Sydney Scott, Washington University, USA The Costs of Autonomy: Decisional Autonomy SESSION OVERVIEW Undermines Advisees’ Judgments of Experts Understanding how consumers make health decisions is theoretically and pragmatically important. Why do consumers avoid going EXTENDED ABSTRACT to their health practitioners? What causes consumers to have positive Imagine that you were recently diagnosed with a serious illness interactions with a provider and therefore become a repeat customer? and you face a very difficult medical decision. You must choose be- Why do consumers seek illegal treatments? In part because of the tween two medical treatments: Treatment A and Treatment B. Each important implications of these types of questions, a great deal of treatment could potentially halt the progress of your illness, but they research has been devoted to understanding health decision-making each are associated with different side effects. Even though your doc- in the fields of marketing, medical decision-making, and psychology, tor has provided you with all of the available information on each of among others. This series of papers contributes to these literatures by the treatments, you still remain confused about which is the better offering novel insights about the drivers of (sometimes suboptimal) choice. Thus, you ask your doctor what you should do. Your doctor health decision-making. reviews the potential costs and benefits of both treatments and re- Currently, the medical decision-making literature emphasizes minds you that there is no objectively right answer. Ultimately, your giving patients autonomy and the ability to choose their own treat- doctor says that she cannot tell you what you should do and you will ments, in a movement for “patient-centered decision making.” Kas- have to make the final decision. How would this make you feel about sirer, Levine and Gaertig (paper 1) show the downsides of decisional your doctor? How would this make you feel about the decision? autonomy. In a series of studies, they compare advisers who give Over the past several decades, the United States medical system decisional autonomy to patients (by providing information but no has increasingly prioritized patient autonomy. Physicians routinely explicit recommendations) to advisers who give paternalistic advice encourage patients to come to their own decisions about their medi- (by providing information and explicit recommendations). Advisers cal care rather than providing patients with clearer, yet more pater- who give decisional autonomy are penalized—viewed as less com- nalistic, advice. Political theorists, bioethicists, and philosophers petent, less helpful, and are less likely to be recommended. generally see this as a positive trend. In the current research, we chal- The medical decision-making literature also suggests that pa- lenge the philosophical and practical assumptions about the benefits tients want their doctors to be socially warm during doctor-patient of autonomy by exploring decision makers’ reactions to autonomy- interactions. Goksel, Berman and Scott (paper 2) suggest that this oriented versus paternalistic advice when making difficult decisions may not always be the case. In particular, when patients have em- under uncertainty. We define autonomy-oriented advice as any ad- barrassing symptoms, they prefer doctor-patient interactions that are vice that ultimately leaves the final decision to the advisee and is not socially cold or mediated by technology. Patients expect socially accompanied by an explicit recommendation. We define paternalistic warm interactions will cause more discomfort when disclosing em- advice as any advice that contains an explicit recommendation. barrassing symptoms. In a pilot study with a sample of employed physicians (N=127), Together, the first two papers look at how social intuitions we found that physicians were torn about whether autonomy or pa- and beliefs can drive consumers to dislike an interaction with their ternalism would be seen more positively by patients. Specifically, healthcare provider. The last two papers look at the heuristics and 56%, 49%, 52%, and 48% of physicians believed that patients would biases in consumers beliefs about the severity of symptoms and the find them to be more competent, more helpful, would be more likely efficacy of treatments. to return to them, and more likely to recommend them, respectively,

Advances in Consumer Research 958 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 959 for providing decisional autonomy instead of paternalistic advice Taken together, our results demonstrate that decision makers (for all judgments, doctors’ expectations regarding which commu- evaluate advisers who give paternalistic advice more favorably than nication tactic would be favored by patients did not significantly dif- those who give autonomy-oriented advice. These results deepen our fer from the null of an equal 50%/50% distribution, all ps ≥ .214). understanding of preferences for autonomy and paternalism and However, as we reveal across 6 studies (N=3,867), patients are not challenge the benefits of recently adopted practices in medical deci- torn. Patients, and advisees in general, show a strong preference for sion-making that prioritize autonomy. paternalism. That is, they have more positive impressions of advisers who offer paternalistic rather than autonomy-oriented advice. Communication Preferences in Medical Advice Seeking In Study 1, participants (N=196) read a scenario that was similar to the opening example. Participants were asked to imagine that they EXTENDED ABSTRACT were a patient who had a very difficult medical decision to make. What types of interactions do people want with healthcare pro- They had to decide between two procedures and imagined that they viders when seeking medical advice? There is a growing conven- asked their doctor what to do. Participants were randomly assigned tional wisdom that patients benefit from interacting with practitio- to one of two conditions: autonomy or paternalism. Participants in ners who display interpersonal warmth towards their patients (e.g., both conditions read that their doctor reminded them that both pro- Di Blasi et al., 2001; Howe, et al., 2019; Kraft-Todd et. al, 2017; cedures had different risks and that either could lead to a favorable Verheul, et al., 2010, etc.), and prefer to receive advice from hu- outcome. Their doctor then either gave them a clear recommendation mans rather than technology (e.g., Longoni et al., 2019; Promberger and said that they should choose procedure A (or B) (paternalism) or & Baron, 2006, etc.). did not give them a recommendation and told them that the choice In this paper, we challenge this conventional wisdom that peo- was theirs to make (autonomy). Participants who received paternal- ple always prefer to consult with interpersonally warm healthcare istic advice evaluated the doctor as more competent (F (1,194) = providers, and explore consumers’ preferences for interpersonal 5.93, p = .016) and more helpful (F (1,194) = 6.11, p = .014) than warmth when seeking care. We argue that when patients have em- participants who received full decisional autonomy. Moreover, par- barrassing symptoms, they become averse to interpersonally warm ticipants who received paternalistic advice indicated that they would interactions with healthcare providers. We demonstrate an aversion be more likely to return to and recommend the doctor than those to interpersonal warmth both as a preference for doctors who hold who received full decisional autonomy (F (1,194) = 5.66, p = .018). an interpersonally colder demeanor towards patients (Studies 1-4), That is, participants who imagined themselves as patients showed as well as a preference for technology-mediated mediums of com- a preference for paternalistic advice over full decisional autonomy. munication (Studies 5-6). In Study 2 (N=451), we largely replicated these effects for medical, Study 1 (N=446, hospital patients) establishes an aversion to financial, and workplace advice. interpersonal warmth in an externally valid, real-world context. We In Study 3 (N=802), we randomly assigned participants to a approached patients at a hospital waiting room to complete a survey. separate or joint evaluation condition (Hsee, 1996), and we exam- Patients wrote down the symptom they had, and chose which doctor ined how advice type influenced doctor choice. In both separate eval- they would like to see among two options: an interpersonally warm uation (89.1% vs. 77.9%, X2 (1) = 9.06, p = .003) and joint evaluation (facially expressive) doctor, or an interpersonally cold (facially ex- (67.5% vs. 33.5%), X2 (1) = 46.57, p < .001), we found that patients pressionless) doctor. The more patients perceived their symptoms to were more likely to choose the doctor who offered paternalistic ad- be embarrassing, the more likely they were to choose the interper- vice. Consistent with Study 1, we again found that doctors who pro- sonally cold doctor (β = .25, p < .001). vided paternalistic advice were viewed as more helpful (ps < .001) Study 2 (pre-registered: https://aspredicted.org/blind. and competent (ps < .065). php?x=8r5yg8, N=737, MTurk) replicates these results using a dif- In Studies 4 (N=807) and 5 (N=806), we furthermore showed ferent manipulation of interpersonal warmth, and investigates the re- that the results hold regardless of whether an explicit recommenda- lationship between interpersonal warmth and anticipated discomfort tion was solicited or unsolicited and whether or not the advice was in communication. Participants imagined either having symptoms accompanied by more expert guidance. that are embarrassing (e.g., diarrhea, hand warts) or non-embarrass- Finally, in Study 6 (N = 805), we extended our investigation ing (e.g., sore throat, knee pain), and calling their clinic for an ap- to an abstracted context in which participants had to choose be- pointment. Participants chose between two doctors: an interperson- tween two raffles, which entailed different risks and benefits.- Par ally cold (strictly professional) doctor and an interpersonally warm ticipants were paired with a raffle expert, who provided participants (more casual) doctor. Participants were more likely to choose the with some information about the raffles and then delivered either interpersonally cold doctor for embarrassing symptoms (51.8%) autonomy-oriented or paternalistic advice. Before participants re- than for non-embarrassing symptoms (40.4%, p < .001). Moreover, ceived the advice, they also evaluated how difficult the choice was whereas those who felt embarrassed by their symptoms anticipated for them to make. There was no main effect of advice type for any more discomfort speaking to an interpersonally warm (vs. cold) doc- of our dependent measures (all ps ≥ .416). Importantly, however, tor (p < .001), those do not feel embarrassed by their symptoms felt there were significant interactions between advice type and decision equally comfortable talking with two types of doctors (p = .29). difficulty for three of the four measures (competence, use again, and Study 3 (pre-registered: https://aspredicted.org/blind. wages assigned; all ps ≤ .049), and a marginally significant interac- php?x=n5mk4c, N=550, MTurk) demonstrates a boundary condition tion for helpfulness (p = .096). Paternalistic advisers were seen more for aversion to social warmth. It manipulates the perspective of the favorably than advisers who offered full decisional autonomy when decision maker and shows that third parties underestimate patient decision difficulty was high, but this preference was attenuated (and preferences for socially cold communicators. Participants either took sometimes reversed) when decision difficulty was low. These results the perspective of a patient and chose a doctor for themselves (Self suggest that advisees evaluate paternalistic advisers more favorably condition), or the perspective of a nurse and assigned a patient to a than advisers who provide decisional autonomy, but only when deci- doctor (Other condition). In both conditions, patients had an embar- sion makers experience decision difficulty. rassing symptom. As predicted, participants were less likely to assign 960 / Consumer Decision-Making in Health Care a patient to the interpersonally cold doctor (38.1%) than they were The Effect of Causal Explanations on Judgments of to choose the interpersonally cold doctor for themselves (49.4%, p = Medical Symptoms .007). Moreover, participants in the Self condition anticipated more discomfort with the interpersonally warm doctor, while participants EXTENDED ABSTRACT in the Other condition expected patients to feel greater discomfort Many medical conditions are complex in nature which makes it with the interpersonally cold doctor (p < .001). difficult for lay people to understand their causes and consequences. In order to disentangle the mechanism behind these pref- For instance, chronic conditions such as fibromyalgia, psoriasis, or erences, Study 4 (pre-registered: https://aspredicted.org/blind. irritable bowel syndrome diseases still lack awareness and under- php?x=j6wz4u, N=141, Behavioral Lab) examines two opposite standing in society although millions of people have their everyday intuitions about interpersonal warmth in doctor-patient interactions. lives severely affected by their symptoms (e.g., Brody, 2017; Hayes We manipulated the intuition people considered before choosing a et al., 2010; Horn et al., 2007; Stern et al., 2004; Wardle, 2017, etc.). doctor: participants either evaluated how much discomfort they an- One common factor to these illnesses is that they are believed to ticipated when talking to a doctor (Discomfort condition), or how re- involve both the psychological and the physical states of individuals assured they thought a doctor would make them feel regarding their (e.g., Kellner, 1994; Perrott et al., 2000; Stojanovich & Marisavljev- embarrassing symptoms (Reassurance condition). While partici- ich, 2008, etc.). pants anticipated greater discomfort with the interpersonally warm An extensive line of research on mind-body dualism has shown doctor (p < .001), they also expected this doctor to provide greater that people tend to think of the mind and the body as two separate reassurance (p < .001). Moreover, participants were more likely to entities (Forstman & Burgmer, 2015, 2017). While in reality the choose the interpersonally warm doctor in the Discomfort condition mind and the body are often closely linked in the demonstration of (65.7%) compared to the Reassurance condition (45.1%, p = .014). illnesses (Ray, 2004), it is unclear whether lay people appreciate this Thus, even though people expect the interpersonally warm doctor to connection or whether they fall prey to dualistic reasoning instead provide them with greater reassurance about their condition, the fact when making judgments regarding different medical symptoms. that they would feel uncomfortable when revealing their symptoms In the present research, we argue that beliefs about the underly- leads them to avoid interpersonal warmth. ing causes of medical symptoms influence lay people’s perceptions Studies 1-4 demonstrate consumer preferences for interper- of the severity of medical symptoms, and their intentions to take sonally colder doctor-patient interactions. Studies 5 and 6 explore medical action regarding these conditions. While previous research the role of technology in mediating these interactions and test an looked at judgments regarding medical symptoms in the absence implication of the previous findings: Consumers may prefer colder versus presence of medical evidence (Asbring & Narvanen, 2002; mediums of communicating with doctors when they have embar- Looper & Kirmayer, 2004; Ruddere et. al, 2012; 2013), we instead rassing symptoms. Study 5 (N=385, MTurk) investigates consumer contrast lay people’s perceptions of medical symptoms when these discomfort associated with using a variety of technology-mediated are driven by different types of causes. In particular, we investigate communication mediums when seeking medical information for whether people perceive the same observable medical symptom dif- either embarrassing or non-embarrassing symptoms. Participants ferently when they believe that it is caused by a psychological as rated the extent to which different mediums of communication were opposed to a physical reason. Across three studies, we demonstrate perceived to facilitate interpersonally warm interactions (e.g., email, that having a psychological (versus physical) cause to a medical online-chat, video call). They also reported their discomfort in us- symptom decreases (1) its perceived severity, (2) people’s tendency ing each medium to get information about either an embarrassing or to recommend medical care for it, and (3) their likelihood to examine non-embarrassing symptom. A linear regression predicting discom- its cause. fort revealed a symptom X warmth interaction (β = .57, p < .001): Study 1 (N = 488, pre-registered) investigates people’s per- warmer mediums of communication elicited more discomfort, but ceptions of a skin rash when it is caused by psychological reasons only when symptoms were embarrassing (embarrassing symptoms: or physical reasons. Participants saw a picture of a skin rash and β = .51, p < .001; non-embarrassing symptoms: β = -.06, p = .22). were told that it developed either as a reaction to temporary stress Lastly, Study 6 (N=252, Behavioral Lab) demonstrates the (psychological cause), or food (physical cause). They made a choice effect of embarrassment on the preference for using technology in about whether they would recommend a patient to go to a doctor, or order to avoid speaking to experts. Participants imagined going to wait for a few days which was our dependent measure. Participants a pharmacy to search for medicine either for an embarrassing or then completed a six-item measure of severity (e.g., how red the rash non-embarrassing symptom. They then chose between using a mo- looks, how itchy it feels, etc.). Participants were less likely to recom- bile app and speaking to a pharmacist to get more information. As mend the patient to see a doctor when the rash had a psychological predicted, participants were more likely to choose to use the app for cause as opposed to a physical cause (β = -.38, p = .05). Addition- embarrassing symptoms (37.9%) compared to non-embarrassing ally, they thought that the rash with a psychological cause was less symptoms (22.6%, p = .009). severe (p < .001). Severity ratings mediated the effect of the cause In sum, this research shows that not only are consumers averse type on their tendency to recommend medical care (IE = -.39, 95% to interpersonal warmth in doctor-patient interactions, this aversion CI = [-.64, -.18]). also encourages them to use technology in order to avoid interacting Study 2 (N = 487, pre-registered) shows the robustness of the with experts when they have embarrassing symptoms. These find- previous findings for another observable symptom. Participants lis- ings suggest that, as opposed to the conventional wisdom, setting an tened to an audio of a patient coughing and were told that the cough interpersonally colder doctor-patient relationships can mitigate the developed as a reaction to either anxiety (psychological cause) or effect of embarrassment on the avoidance of contacting experts. being exposed to chemicals in the air (physical cause). They then made the same choice of recommendation, and completed a measure of severity (e.g., how loud the cough sounds, how scratchy it feels, etc.). Participants were less likely to recommend the patient to see a doctor (β = -.57, p = .003) and perceived the cough to be less severe Advances in Consumer Research (Volume 48) / 961

(p = .005) when it had a psychological cause. As before, severity rat- we look at an important new lay belief surrounding legality. We posit ings mediated the tendency to recommend medical care (IE = -.27, that illegal drugs are viewed as more effective. 95% CI = [-.47, -.08]). As an initial test of the illegal = effective heuristic (Study 1, Study 3 (N = 898) looked at the influence of temporal primacy N=404), participants evaluated a fictional product that was either de- in judgements of medical conditions. Specifically, it explored cases scribed as legal or illegal. To test for robustness, we also manipulated when an individual suffers from multiple medical symptoms, and whether the drug was a utilitarian (sleep aid) or hedonic (intoxicant) investigated how causal belief influence the extent to which indi- product, in a 2x2 between-subjects design. Participants were asked viduals prioritize seeking care for the causally central medical con- to imagine the following: “Stava is a beverage made from Piper Me- dition among two symptoms. We predicted that individuals have an thysticum, a plant native to the western Pacific Islands.” Those in the overall tendency to treat the causally central symptom, but that this utilitarian condition further read, “Stava is typically used as a me- tendency would decrease when a psychological symptom leads to a dicinal sleep aid,” while those in the hedonic condition read, “Stava physical symptom. Participants imagined a person developing two is typically used as a social drink and makes its users intoxicated.” different medical symptoms. We had two focal conditions, where we Finally, participants read, “Stava is legal [not legal]”. Participants described an individual developing a skin rash and anxiety. In one then responded to the following items: effectiveness (3-item scale: condition, anxiety was described as the central cause leading to a strength, potency, effectiveness; alpha = .86) and safety (2-item skin rash, and in the second condition a skin rash was described as scale: safety and concern about side-effects [reverse-scored]; alpha the central cause leading to anxiety. In addition, we had two con- =.76). We found a non-significant interaction of condition (utilitar- trol conditions. In one of the conditions, a person was described ian/hedonic) and treatment (legal/illegal) on effectiveness. For both of having two psychological symptoms (i.e. anxiety and sleepless- conditions, participants believed the illegal drug was indeed less safe ness), while in the other condition a person was described of hav- (Ps<.01), but more importantly, also more effective (Ps<.01). ing two physical symptoms (i.e. sore throat and skin rash). In all Study 2 (pre-registered, N=754) was designed as an additional experimental conditions, participants were highly likely to treat the demonstration and robustness check of the illegal = effective heu- symptom that appeared first (86%-97%) except for the condition in ristic. As in study 1, our key manipulation was whether a product which a psychological symptom caused a physical symptom (67%). was described as illegal or legal, and we tested for robustness across In particular, participants were less likely to treat the causally central different products (a stimulant, an exercise supplement, or a weight- symptom when it was a psychological symptom, leading to a physi- loss drug). Therefore, this study followed a 2x3 between-subjects cal symptom, compared to all other conditions (ps < .001). design. Participants responded to the same effectiveness and safety In sum, these studies show that casual beliefs about the underly- items used in Study 1. There was a significant effect of the legal- ing cause of medical symptoms influences both perceptions of severity treatment (p<.001), but no effect of the product type condition. ity and the intentions of providing medical care. Symptoms caused That is, across all three products, participants who were told that the by psychological illness are perceived to be less severe and less product is illegal rated it as not only less safe, but also significantly likely to warrant medical care. more effective. Study 3 (N=401) was designed to test whether the illegal = ef- The Illegal = Effective Heuristic fective heuristic persists when the drug (in this case, a sleep aid) is illegal for reasons explicitly unrelated to efficacy or safety. This study EXTENDED ABSTRACT had 5 between-subjects conditions: 1) Legal, 2) Illegal (no reason When deciding whether to try a drug or treatment, whether for provided), 3) Illegal due to paperwork error, 4) Illegal due to side- recreational, medical, or other purposes, two of the most important effects, and 5) illegal even though it is as safe as the leading sleep- factors for consumers to consider are safety (i.e., how likely is the aids. All illegal conditions were perceived as more effective than the drug to cause side-effects?) and efficacy (i.e., how potent andef- legal treatment (p<.001). For an additional effectiveness measure, fective is the drug?). This work looks at consumer perceptions of we asked how many people (out of 100) would likely find relief us- illegal (vs. legal) products. When discouraging the use of illegal ing the treatment. For all treatments, participants believed that more substances, public health campaigns often highlight safety concerns. people would find relief from illegal (vs. legal) treatments. Beliefs Our research finds that people already tend to believe that illegal sub- about safety remained higher for the legal condition compared to the stances are less safe than legal substances, but surprisingly, they also illegal conditions. Study 3 shows that individuals continue to believe perceive illegal substances as more effective. Three studies establish that an illegal treatment will be more effective, even when the drug the existence of this illegal = effective heuristic; across a range of is illegal for reasons unrelated to strength, potency, or even safety. substances, illegal products are viewed as stronger, more potent, and We find that illegal products are indeed viewed as less safe than more effective than legal products. This effect holds even when the legal products, but surprisingly we also find that they are perceived product is illegal for reasons explicitly unrelated to efficacy or safety to be more effective. As with other heuristics, the illegal = effective (i.e., a paperwork error). heuristic may at times be an accurate inference. For example, the Consumers often lack complete information about their product healthy = expensive intuition finds that while healthier foods are cer- choices, either because attributes cannot be observed or because pro- tainly more expensive at times, consumers overgeneralize this belief curing attribute information is prohibitively costly. In all these situ- to contexts where it is not the case (Haws, Reczek, & Sample, 2017). ations, consumers make inferences about product attributes based on There are many possible reasons for a drug to be illegal, ranging their lay beliefs. For example, a consumer might infer that a product from safety concerns to issues related to politics, sourcing, and his- is high quality because it has a warranty, or infer that a healthy food tory. We demonstrate that the illegal = effective heuristic holds even will be less tasty (Srivasta & Mitra, 1998; Raghunathan, Naylor, & when the product in question is illegal for reasons unrelated to ef- Hoyer, 2006). Oftentimes, these lay beliefs involve looking at an ficacy or safety. This finding has implications for public health and easily observable attribute (e.g., whether there is a warranty) to make marketing. For example, the legalization of products such as CBD an inference about a difficult to observe attribute (e.g., quality). Here (a cannabis derived product) may lead consumers to infer reduced potency compared to when the product was not available legally. 962 / Consumer Decision-Making in Health Care

Additionally, individuals with lower concern about legal repercus- Kellner, R. (1994). Psychosomatic syndromes, somatization and sions and high concern about potency (e.g., people suffering from somatoform disorders. Psychotherapy and Psychosomatics, addiction or other mental illness) may assume that illegal products 61(1-2), 4-24. are more effective at treating their suffering than legal alternatives Kraft-Todd, G. T., Reinero, D. A., Kelley, J. M., Heberlein, A. S., offered by their doctor. While many public health campaigns educate Baer, L., & Riess, H. (2017). Empathic nonverbal behavior individuals about the safety concerns of illegal drugs, we hope to increases ratings of both warmth and competence in a medical further explore this effect by studying whether we can reduce inter- context. PloS One, 12(5), e0177758. est in illegal substances by reducing perceived effectiveness instead. Longoni, C., Bonezzi, A., & Morewedge, C. K. (2019). Resistance to medical artificial intelligence.Journal of Consumer REFERENCES Research, 46(4), 629-650. Åsbring, P., & Närvänen, A. L. (2002). Women’s experiences Looper, K. J., & Kirmayer, L. J. (2004). Perceived stigma in of stigma in relation to chronic fatigue syndrome and functional somatic syndromes and comparable medical fibromyalgia.Qualitative Health Research, 12(2), 148-160. conditions. Journal of Psychosomatic Research, 57(4), 373- Brody, J. (2017, November 27). New Recognition for Chronic 378. Fatigue. https://www.nytimes.com/2017/11/27/well/new- Perrott, S. B., Murray, A. H., Lowe, J., & Mathieson, C. M. (2000). recognition-for-chronic-fatigue.html The psychosocial impact of psoriasis: physical severity, quality De Ruddere, L., Goubert, L., Vervoort, T., Prkachin, K. M., & of life, and stigmatization. Physiology & Behavior, 70(5), Crombez, G. (2012). We discount the pain of others when 567-571. pain has no medical explanation. The Journal of Pain, 13(12), Promberger, M., & Baron, J. (2006). Do patients trust computers?. 1198-1205. Journal of Behavioral Decision Making, 19(5), 455-468. De Ruddere, L., Goubert, L., Stevens, M., Williams, A. C. D. C., Raghunathan, R., Naylor, R. W., & Hoyer, W. D. (2006). The & Crombez, G. (2013). Discounting pain in the absence of unhealthy = tasty intuition and its effects on taste inferences, medical evidence is explained by negative evaluation of the enjoyment, and choice of food products. Journal of patient. PAIN, 154(5), 669-676. Marketing, 70 (4), 170–84. Di Blasi, Z., Harkness, E., Ernst, E., Georgiou, A., & Kleijnen, Ray, O. (2004). How the mind hurts and heals the body. American J. (2001). Influence of context effects on health outcomes: a Psychologist, 59(1), 29. systematic review. The Lancet, 357(9258), 757-762. Srivastava, J. & Mitra, A. (1998). “Warranty as a signal of quality: Forstmann, M., & Burgmer, P. (2015). Adults are intuitive mind- The moderating effect of consumer knowledge on quality body dualists. Journal of Experimental Psychology: General, evaluations,” Marketing Letters, 9 (4), 327–36. 144(1), 222. Stern, R. S., Nijsten, T., Feldman, S. R., Margolis, D. J., & Forstmann, M., & Burgmer, P. (2017). Antecedents, manifestations, Rolstad, T. (2004, March). Psoriasis is common, carries a and consequences of belief in mind–body dualism. In The substantial burden even when not extensive, and is associated Science of Lay Theories (pp. 181-205). Springer, Cham. with widespread treatment dissatisfaction. In Journal of Haws, K. L., Reczek, R. W. & Sample, K. L. (2017). Healthy diets Investigative Dermatology Symposium Proceedings (Vol. 9, make empty wallets: The healthy=expensive intuition. Journal No. 2, pp. 136-139). Elsevier. of Consumer Research, 43 (6), 992-1007. Stojanovich, L., & Marisavljevich, D. (2008). Stress as a trigger of Hayes, S. M., Myhal, G. C., Thornton, J. F., Camerlain, M., autoimmune disease. Autoimmunity Reviews, 7(3), 209-213. Jamison, C., Cytryn, K. N., & Murray, S. (2010). Fibromyalgia Verheul, W., Sanders, A., & Bensing, J. (2010). The effects of and the therapeutic relationship: where uncertainty meets physicians’ affect-oriented communication style and raising attitude. Pain Research and Management, 15. expectations on analogue patients’ anxiety, affect and Horn, E. J., Fox, K. M., Patel, V., Chiou, C. F., Dann, F., & expectancies. Patient Education and Counseling, 80(3), 300- Lebwohl, M. (2007). Are patients with psoriasis undertreated? 306. Results of National Psoriasis Foundation survey. Journal of Wardle, A. (2017, January 20). IBS is underdiagnosed and the American Academy of Dermatology, 57(6), 957-962. undertreated in the US. https://www.univadis.co.uk/ Howe, L. C., Leibowitz, K. A., & Crum, A. J. (2019). When viewarticle/ibs-is-underdiagnosed-and-undertreated-in-the- Your Doctor “Gets It” and “Gets You”: The Critical Role of us-482379 Competence and Warmth in the Patient-Provider Interaction. Frontiers in Psychiatry, 10, 475. Hsee, C.K. (1996). Evaluability hypothesis: An explanation for preference reversals between joint and separate evaluations of alternatives. Organizational Behavior and Human Decision Processes 67: 247-257 Digging Deeper into the Self: Understanding How Identity Activation and Structure Influences Behavior Chair: Hyoseok Kim, University of Alberta, Canada

Paper #1: Look for the Signature: Personal Signatures on of a salient identity’s agentic properties. Morgan, Kettle, and Reed Marketing Stimuli Affect Consumption by Making Identities show that an identity’s agentic properties derive from the tendency Salient for individuals to let others choose for them in situations where that Keri L. Kettle, University of Manitoba, Canada identity is relevant and diagnostic. They examine that making an au- Antonia Mantonakis, Brock University, Canada tonomous (vs. heteronomous) identity salient induces self-product connections and thus reduces the likelihood of post-choice switch- Paper #2: The Logged-In Shopper: How Consumer ing. Authentication Influences Purchase Behavior Together, these four papers examine mechanisms by which con- Hyoseok Kim, University of Alberta, Canada sumers’ self-identity influences how they evaluate products, make Gerald Häubl, University of Alberta, Canada choices, and act upon those choices. The research presented in this Paper #3: Choosing Between “Me’s”: Why Greater Self-Variety session has implications for researchers interested in consumer iden- Makes It More Difficult to Choose Between Identity Goods tity, self-variety, agency, and judgment and decision-making. Sara Loughran Dommer, Georgia Tech, USA Karen Page Winterich, Pennsylvania State University, USA Look for the Signature: Personal Signatures on Paper #4: Autonomous Properties of Identity Marketing Stimuli Affect Consumption by Making Carter Morgan, University of South Florida, USA Identities Salient Keri L.Kettle, University of Manitoba, Canada Americus Reed, University of Pennsylvania, USA EXTENDED ABSTRACT Management at Niagara’s Riverview Cellars once added a fac- SESSION OVERVIEW simile of the winemaker’s personal signature to the back label of a We all have a sense of who we are in the context of our social wine sold exclusively at their retail store. Adding the signature was roles, ethnic or national affiliations, and character traits (Reed et al. associated with an (statistically significant) 80% sales increase – re- 2012). Self-identity—defined as one’s overall sense of self—guides markable given that the signature was printed on the bottle’s back our thoughts and feelings and influences our behavior (Kleine et al. label. What could explain this? 1993). Early self-identity research has primarily focused on how Marketers intuit that signatures signal quality (Smith 2012). Yet, priming specific identities or constructs influences identity-relevant people sign their name to represent their identity in writing (Hawkins attitudes through self-esteem or self-consistency (Kettle and Häubl 2011), and a personal signature is thus a meaningful symbol of a 2011; Reed et al. 2012). However, much remains unknown about person’s identity (Koo and Fishbach 2016). As an identity symbol the mechanisms responsible for the effects of self-identity on con- for the signer, we hypothesize that a personal signature serves as an sumption decisions and behaviors. The papers in this session aim to identity prime and thus makes salient identities within a consumer’s address this gap by examining two important questions: 1) how does self-concept that are relevant to the signer’s identity. This implies self-identity activated by contextual cues affect product evaluations that a personal signature on marketing stimuli (e.g., label, advertise- and consumption decisions? and 2) how does a multifaceted view of ment) should enhance (attenuate) product evaluations and preference self-identity influence identity-relevant choices? among consumers with an associative (dissociative) identity relevant The first two papers in this session examine the influence of to the signer’s identity. self-identity prompted by contextual cues. Kettle and Mantonakis For clarity, identity self-importance describes the chronic role show that adding a personal signature to marketing stimuli affects of an identity within the individual’s self-concept, identity salience consumption behavior. They find that a personal signature could be describes the current activity of an identity in one’s thoughts, and an identity prime and enhance product evaluations when the con- identity prime is an intervention that generates identity salience (Ket- sumer’s relevant identity is associative to the signer’s identity but tle 2019); Reed 2012). An associative identity is one possessed by lowers evaluations when the identity is dissociative. In the second the signer that the consumer also sees in themselves (e.g., consumer paper, Kim and Häubl argue that the timing of identity salience is and signer are both American), whereas dissociative identity is an critical when self-identity is involved in the activation of an imple- identity possessed by the signer that conflicts with the consumer’s mental mindset. They find that an identification request that taps into identity (e.g., consumer is American, signer is Canadian). self-identity at the beginning of a shopping process induces an imple- Studies 1-2 were retail field experiments. In study 1, we ma- mental mindset (e.g., planning which product to buy), which renders nipulated whether the winemaker’s personal signature was printed consumers more likely to make a purchase. on a display sign located next to one type of wine in the store. We The next two papers focus on how a multifaceted view of self- had a control condition with no display sign, a second control condi- identity influences identity-relevant actions. Dommer and Winterich tion with a display sign featuring no personal signature, and a third show that consumers’ perceptions of the similarity between two iden- condition with a display sign featuring a facsimile of the personal tity options depend on self-variety—the extent to which the self is signature. The personal signature was associated with significantly composed of many different and diverse identities. They find that in- greater sales as compared to the control conditions. dividuals with low (vs. high) self-variety perceive two identity goods We conducted study 2 during a wine tasting: we manipulated as more similar (vs. dissimilar) to one another, which makes it easier whether the winemaker’s personal signature was printed on the bot- (vs. harder) for them to choose between the options. The final paper tle’s back label. Participants rated the taste and quality of the wine. examines how consumers’ decision confidence varies as a function We operationalized relevant self-important identity as whether the

Advances in Consumer Research 963 Volume 48, ©2020 964 / Digging Deeper into the Self: Understanding How Identity Activation and Structure Influences Behavior consumer self-identified as Ontarion. The personal signature gener- consumers’ likelihood of making a purchase. This is because “on- ated more favourable product evaluations and greater product sales entry” authentication requests activate an implemental mindset (i.e., among consumers with an associative identity, but less favourable focusing on the implementation of a goal-directed behavior), which evaluations (with no effect on sales) among consumers with a dis- in turn increases consumers’ likelihood of purchasing a product. On sociative identity. the other hand, when entering a store without such authentication, Studies 3-5 were online experiments. In study 3, we crafted fic- consumers are more likely to be in a deliberative mindset (i.e., con- titious signatures for the founders of two unfamiliar brands (Blue- sidering the pros and cons of pursuing a specific goal), and therefore, bird Farms, Farm Boy) and we independently manipulated whether they tend to be less certain about whether they should make a pur- a personal signature was present on (versus absent from) the product chase. In this case, encountering a request to authenticate just before label and whether the signer’s identity was associative (versus disso- completing a purchase does not activate an implemental mindset, ciative) relative to the participant’s home state. Replicating study 2, which in turn reduces their likelihood of purchasing a product. More- the personal signature on the product label led to more (less) favor- over, we propose that salience of self-identity during the authenti- able evaluations of products in the associative (dissociative) identity cation process is critical to activate an implemental mindset. Prior condition, but only among consumers with high identity self-impor- literature implies that an implemental mindset is induced when the tance. sense of self is made salient during goal pursuit (Bayer and Gollwit- Study 4 tested the mental mechanism (identity salience) by ma- zer 2005). Consistent with this, we examine how making an identity nipulating the diagnosticity of a subsequent hypothetical donation salient (vs. CAPTCHA) during the authentication process activates scenario (ostensibly) unrelated to either the personal signature or the an implemental mindset, rendering consumers more likely to pur- marketing stimuli. Using an associative identity (home state) in all chase. Three experiments support our theorizing. conditions, we independently manipulated the presence (versus ab- In Experiment 1 (N=318), MTurk participants were randomly sence) of a fictitious personal signature on the label of an unfamiliar assigned to one of two conditions (authentication request: at the be- brand (Organic Valley), and whether a subsequent donation scenario ginning vs. near the end). Participants were instructed that they were was identity-diagnostic (versus non-diagnostic). In the identity-di- considering purchasing a new backpack in an online store and that agnostic condition, we described donation recipients as originating there would be a draw at the end of the study in which one partici- from the participant’s home state: in the non-diagnostic condition we pant would be chosen as the winner. Before beginning the shopping provided no information about recipients’ origins. As hypothesized, task, participants were asked to provide their email address so that the self-importance of the consumer’s associative identity only pre- they could be contacted if they are the winner. In the authentication- dicted donation amount in one condition: the personal signature request-at-the-beginning condition, participants were required to log present and identity-diagnostic condition. In all other conditions, in by confirming their (previously typed) email address to enter the identity self-importance did not predict donations. store, whereas participants in the authentication-request-near-the- Study 5 examined generalizability by manipulating whether end condition entered the store without the login request. If partici- the personal signature of baseball legend Derek Jeter appeared on pants chose not to enter the store, they proceeded to an unrelated Modavo watch advertisements. We asked 300 American men to rate study. Once participants entered the store, they had the opportunity the watches, indicate purchase interest, and rate the self-importance to purchase one of eight different backpacks, but were free to exit the of their baseball identity. Replicating studies 2-4, the signature en- store without making a purchase. In the authentication-request-at- hanced (lowered) product evaluations among passionate baseball the-beginning condition, participants completed their purchase with- fans (men who dislike baseball), and identity self-importance mod- out additional authentication at checkout, whereas in the authenti- erated the mediating role of product evaluations in the effect of per- cation-request-near-the-end condition, participants were required to sonal signatures on purchase interest. log in by confirming their email address in order to complete their Our research contributes to our understanding of consumer purchase. The dependent variable was whether a participant com- identity by showing that an identity within a consumer’s self-concept pleted the purchase of a backpack. We showed that when consumers can be made salient by a cue that is not specific to that particular encountered an authentication request at the beginning (vs. near the identity – the personal signature of another person who possesses a end) of a shopping process, they were more (vs. less) likely to make relevant identity. a purchase (=20% vs. =10%; (1)=4.89, p=.03). In Experiment 2 (N=150), we replicated our previous findings The Logged-In Shopper: How Consumer Authentication and tested whether an implemental mindset mediates the early au- Influences Purchase Behavior thentication effect by introducing a procedure that offsets a deliberative-implemental mindset. MTurk participants were randomly EXTENDED ABSTRACT assigned to a 2 (authentication request: at the beginning vs. near the With the rise of digital retail platforms, authentication (e.g., end) x 2 (cross mindset manipulation: present vs. absent) between- login) has become an important topic for both firms and research- participants design. In the cross-mindset-manipulation-present con- ers (Schaupp and Belanger 2005; Lee et al. 2012). Authentication dition, there was an additional procedure after the manipulation is defined as the process or action of verifying the identity of a user of authentication which was the same as in Experiment 1. In the (Basu and Muylle 2003) and it generally requires the user to enter authentication-request-at-the-beginning condition, after participants identifying credentials (e.g., email address, fingerprint, etc.). Intui- logged in and entered the store, a deliberative mindset was activated tively, authentication is irritating to consumers. Prior research shows by listing five potentially positive and five potentially negative con- that breaking the continuity of online shopping can reduce consum- sequences of switching banks. By contrast, in the authentication-reers’ purchase motivation (McDowell et al. 2016). Therefore, it is in quest-near-the-end condition, after participants entered the store, an a firm’s best interest to minimize consumers’ effort associated with implemental mindset was activated by planning the implementation authentication until time for checkout (e.g., Expedia’s login recom- of switching banks and listing five steps to execute the plan. In the mendation at checkout). In contrast to this view, we propose that en- cross-mindset-manipulation-absent condition, there was no addi- countering a request to authenticate when entering a store increases tional procedure after the manipulation of authentication. The depen- Advances in Consumer Research (Volume 48) / 965 dent variable was whether a participant completed the purchase of a Assanand, and Di Paula 2000; Gramzow et al. 2000; Greenwald et backpack. We showed that the early authentication effect was attenu- al. 2000; Reed and Forehand 2016; Reed et al. 2012). The amount ated in the cross-mindset-manipulation-present condition (=10% vs. of variety within the self essentially assesses how distinctive each =15%; (1)=7.48, p=.39), but not in the cross-mindset-manipulation- identity is from others within the global self, based on the degree of absent condition (=31% vs. =6%; (1)=4.70, p=.03). shared associations between each identity (Campbell et al. 2000). At In Experiment 3 (N=398), we tested an important boundary one end of the spectrum, low self-variety, the self can contain a few condition: identity salience. MTurk participants were randomly as- identities with many associations, creating an amorphous whole with signed to a 2 (authentication request: at the beginning vs. near the little differentiation. At the other end of the spectrum, high self-va- end) x 2 (identity salience: authentication vs. CAPTCHA) between- riety, the self can contain a greater number of identities with few as- participants design. In the authentication condition, the manipulation sociations, indicating that the self is multifaceted and unique across of authentication at the beginning versus authentication near the end identities, contexts, and situations. was the same as in Experiment 1. In the CAPTCHA condition, on Identities share more in common for individuals with low self- the other hand, participants proved that they were not a robot at the variety than individuals with high self-variety. Such identity integra- beginning (vs. near the end). The dependent variable was whether a tion can help reduce conflict across identities (Benet-Martínez et al. participant completed the purchase of a water bottle. We showed that 2002; Zou, Morris, and Benet-Martínez 2008). Thus, we believe that the early authentication effect held in the authentication condition individuals with low self-variety will perceive two identity goods as (=25% vs. =13%; (1)=4.56, p=.03), but not in the CAPTCHA condi- more similar to one another, which should make it easier for them tion (=21% vs. =19%; (1)=.25, p=.61). This experiment rules out an to choose between the options (Chernev 1997; Houston et al. 1991; alternative explanation based on the sunk cost fallacy. If the increase Nagpal et al. 2011). On the other hand, because their identities are of the purchase rate in the authentication-at-the-beginning condition distinct from one another, individuals with high self-variety should is driven by the mere effort that consumers invest in the purchase view identity goods as being more dissimilar to one another. Without process, then the purchase rate in the CAPTCHA-at-the-beginning a dominant option, it should be difficult for them to choose between condition should also have increased. However, as demonstrated, two identity goods (Dhar 1997; Shugan 1980; Tversky and Shafir this was not the case. 1992). This research examines how the timing of consumer authentica- Six studies support our theorizing. In study 1a, a panel of par- tion influences shopping behavior. It contributes to prior literature on ents made a hypotehtical choice between two pairs of socks and how contextual cues that consumers encounter while shopping acti- completed our measure of self-variety. Participants with high self- vate deliberative versus implemental mindsets and thereby influence variety took longer to make a choice when their choice set contained their purchase behaviors. Our findings have an important managerial two identity-related socks compared to individuals with low self- implication: firms can benefit from providing authentication requests variety, and relative to when their choice set contained only one pair first in a shopping environment. of identity socks. Study 1b replicated the results of study 1a in a lab experiment where students made a real choice between canvas bags. Choosing Between “Me’s”: Why Greater Self-Variety In study 2, we manipulated self-variety through a writing task. Makes It More Difficult to Choose Between Identity Participants then made a hypothetical choice between two t-shirts. Goods In addition to capturing the time they took to make the decision, we also included an explicit measure of choice difficulty. Individuals in EXTENDED ABSTRACT the high self-variety condition took longer to decide and reported Consumers are motivated to behave in identity-congruent ways greater decision difficulty when two identity options were present in (Oyserman 2009). As such, prior research has established consum- a choice set compared to those in the low self-variety condition, and ers’ preference for identity products, especially when the identity is relative to when their choice set contained only one identity option. salient (Reed et al. 2012). But these paradigms have traditionally In study 3, in addition to the two identity goods, we also in- pitted goods that express or are linked to one’s identity (“identity cluded a non-identity (i.e., neutral) good the choice set. In support of goods”) with neutral (no identity) goods, prompting the question – our theorizing that deciding between two identity goods is especially what happens when an individual must choose between two identity difficult for individuals with high self-variety, these individuals were goods? more likely to “opt-out” of the decision and choose the neutral option In the typical choice model, choosing creates conflict because the compared to individuals with low self-variety. chosen option is never perfect, and the forgone option has some desir- Study 4 included a condition where instead of choosing be- able aspects that are lost (Festinger 1964; Shultz, Léveillé, and Lepper tween goods, participants simply rated them. We find our effect of 1999). The more features shared by choice options, however, the less self-variety and the number of identity goods on the length of time it time needed to make the decision (Chernev 1997; Houston, Sherman, took participants to choose between products, but not on the length and Baker 1991; Nagpal et al. 2011). In the extreme case, choosing of time it took participants to evaluate products, suggesting that deci- between two identical goods should be very easy. On the other hand, sion time is not measuring information processing or involvement. choosing between two very distinct goods is more difficult, as consum- In Study 5 participants made a real choice between two differ- ers experience conflict when choice involves trade-offs among different ent stress balls. We measured both the time they took to make their attributes (Shugan 1980). Thus, the difficulty experienced when choos- decision as well as perceptions of similarity between the two stress ing between two identity goods should depend on the similarity be- balls. We once again replicated our effect and demonstrate that per- tween the two identity options. We believe that consumers’ perceptions ceptions of similarity mediates the effect. of the similarity between two identity options will depend on how much Finally, in study 6, we primed either a maximizing or satisficing similarity (or distinctiveness) they have within their self—a measure mindset (Ma and Roese 2014). A satisficing mind-set reduced deci- we are calling “self-variety.” sion difficulty among individuals with high self-variety, suggesting The self links various aspects of self-knowledge, including that these individuals have a tendency to engage in greater compari- roles, identities, and traits based on associative relations (Campbell, son of options. Priming a maximizing mind-set, however, encour- 966 / Digging Deeper into the Self: Understanding How Identity Activation and Structure Influences Behavior ages such comparison and thus increased decision difficulty among their gender identity (man, woman) and their family identity (son, those with low self-variety. daughter). Consistent with study 1, both male and female partici- In examining the effect of self-variety on a choice set consist- pants considered their gender identity to be more autonomous than ing of more than one identity good, we make two important contri- their family (son, daughter) identity. Together, studies 1 and 2 pro- butions. First, we contribute to work on identity-based motivation vide convincing evidence that identities have distinct autonomous (Oyserman 2009) by examining a context in which consumers must properties. choose between two actions, each of which is congruent with one Having established that autonomous properties differ across of their identities. Second, while only few papers in marketing have identities, in studies 3 and 4 we investigated the extent to which examined multiple identities (Saint Clair and Forehand 2019; Win- autonomous (vs. heteronomous) identities might influence product terich and Barone 2011; Winterich, Mittal, and Ross 2009), even choice and self-product connection. In study 3, we asked participants fewer have looked at the structure among an individual’s multiple to describe an identity they possessed. We manipulated (between- identities (Savary and Dhar 2020). By demonstrating the effect of subject) whether that identity was autonomous (“an identity with high self-variety on choosing between two identity goods, we offer which you feel you have very few restrictions on your actions and a more comprehensive understanding of how one’s self influences behavior”) or heteronomous (“… very many restrictions …”). We consumer behavior. then asked participants to think of a physical item that they associate with that identity. Participants in the autonomous identity condition Autonomous Properties of Identity indicated a stronger self-product connection to their self-generated item than participants in the heteronomous identity condition. The EXTENDED ABSTRACT relationship between the autonomous (versus heteronomous) iden- We all possess different identities that comprise our overall tity and the self-product connection was fully mediated by the au- sense of who we are (Reed et al. 2012). It is well-established that dif- tonomous properties of the identity. ferent identities, ranging from gender to athlete, musician to soldier, Study 4 used a similar task as study 3: we asked participants to predictably influence consumer decisions in identity-diagnostic con- describe an identity they possess (autonomous versus heteronomous) texts (Reed and Forehand 2016). But how might identities influence and a physical item that they associate with that identity. Replicat- decisional autonomy in contexts that are not identity-diagnostic? ing study 3, self-product connection was stronger in the autonomous And what downstream consequences might these effects have? (versus heteronomous) identity condition. We then asked partici- Though autonomy is central to an individual’s sense of self pants to imagine a scenario in which a better alternative was avail- (Deci and Ryan 2002), not all identities are equally centered on au- able to replaced their existing product. Participants were less likely tonomy. We examine how the autonomous properties of an identity to switch in the autonomous (versus heteronomous) identity condi- – the centrality of autonomous decision-making in shaping that iden- tion, and this effect was fully mediated by self-product connection. tity – vary across identities, and predictably affect consumer choice In study 5, we turned our attention to pinpointing the key and post-choice behavior. First, we establish that autonomy is more characteristic underlying the difference between autonomous - ver central to certain identities, and we pinpoint the key determinant of sus heteronomous identities. We examined six potential character- an identity’s autonomous property as the extent to which one lets istics, such as how much one considers others when that identity others make choices for them when that identity is relevant. Then, is relevant, and how often they are faced with difficult decisions. we demonstrate that triggering an autonomous (versus heterono- We examined four identities (family, gender, student, Greek mem- mous) identity in a choice context strengthens self-product connec- ber). Replicating studies 1-2, the gender and student identities had tion, and leads consumers to stick with their choice in the face of stronger autonomous properties than the family and Greek member disconfirming information. identities. Importantly, only one characteristic explained the differ- We began our investigation by developing a 4-item autonomous ence in autonomous properties across all identities: for autonomous properties (AP) scale inspired by the autonomy literature (Hong and (vs. heteronomous) identities, participants were less likely to have Faedda 1996; Moller et al. 2007). This scale examines identity- other people make decisions on their behalf: this fully mediated the specific associations with autonomous versus constrained behavior relationship. (e.g., “When in a context where my [target identity] is relevant, my Our work provides an important contribution to the identity freedom of choice is restricted.”). Each item was on a 7-point scale and decision-making literatures. Where prior findings identify in- (1 = strongly disagree, 7 = strongly disagree). Scale reliability was terpersonal differences in autonomous behavior, we demonstrate an high across all four studies (α ≥ .78). intrapersonal factor – the autonomous properties of distinct identities We conducted studies 1-2 to examine how autonomous prop- within one’s self-concept – that predictably impacts self-product con- erties vary across different identities within the same individual’s nection and willingness to forego a previously chosen option. While self-concept. In study 1 we examined peoples’ perceptions for au- prior work on consumer autonomy highlights differences in agentic tonomous properties of common identities: family identity (son, behavior across individuals, we demonstrate intra-individual differ- daughter), gender identity (man, woman), university student, and ences in autonomy that directly impact consumer decision-making. Greek member (fraternity, sorority). We were particularly interested in contrasting pairs of identities that would both reside within a sin- REFERENCES gle consumer’s self-concept: man vs. son, woman vs. daughter, and Basu, Amit, and Steve Muylle (2003), “Authentication in university student vs. Greek member. Within these pairs, participants E-Commerce,” Communications of the ACM, 46 (12), 159- perceived greater autonomy for (1) the man identity versus the son 166. identity, (2) the female identity versus the daughter identity, and (3) Benet-Martínez, Verónica, Janxin Leu, Fiona Lee, and Michael W. the university student identity versus the Greek (fraternity, sorority) Morris (2002), “Negotiating Biculturalism: Cultural Frame identities. Switching in Biculturals with Oppositional Versus Compatible In study 2, using a within-subject design, we examined the au- Cultural Identities,” Journal of Cross-Cultural Psychology, 33 tonomous properties for two of each participant’s own identities: (5), 492-516. Advances in Consumer Research (Volume 48) / 967

Campbell, Jennifer D., Sunaina Assanand, and Adam Di Paula Lee, Jong-Eun Roselyn, Shailendra Rao, Clifford Nass, Karin (2000), “Structural Features of the Self-Concept Amd Forssell, and Jae Min John (2012), “When Do Online Adjustment,” in Psychological Perspectives on Self and Shoppers Appreciate Security Enhancement Efforts? Effects of Identity, Abraham Tesser and Richard B. Felson and Jerry Financial Risk and Security Level on Evaluations of Customer M. Suls, eds. Washington, DC: American Psychological Authentication,” International Journal of Human-Computer Association, 67-87. Studies, 70 (5), 364-376. Chernev, Alex (1997), “The Effect of Common Features on Brand Ma, Jingjing and Neal J Roese (2014), “The Maximizing Mind- Choice: Moderating Role of Attribute Importance,” Journal of Set,” Journal of Consumer Research, 41 (1), 71-92. Consumer Research, 23 (4), 304-11. McDowell, William C., Rachel C. Wilson, and Charles Owen Kile Dawar, Niraj, and Philip Parker (1994). “Marketing Universals: Jr (2016), “An Examination of Retail Website Design and Consumers’ Use of Brand Name, Price, Physical Appearance, Conversion Rate,” Journal of Business Research, 69 (11), and Retailer Reputation as Signals of Product Quality,” 4837-4842. Journal of Marketing, 58 (2), 81-95. Moller, Arlen C., Richard M. Ryan, and Edward L. Deci (2007), Deci, Edward L., and Richard M. Ryan (2002), “Overview of “Self-Determination Theory and Public Policy: Improving the Self-Determination Theory: An Organismic Dialectical Quality of Consumer Decisions Without Using Coercion,” Perspective,” Handbook of Self-Determination Research, 3-33. Journal of Public Policy and Marketing, 25 (1), 104-16. Dhar, Ravi (1997), “Consumer Preference for a No-Choice Nagpal, Anish, Adwait Khare, Tilottama Chowdhury, Lauren Option,” Journal of Consumer Research, 24 (2), 215-31. Labrecque, and Ameet Pandit (2011), “The Impact of the Festinger, Leon (1964), Conflict, Decision and Dissonance. Amount of Available Information on Decision Delay: The Stanford: Stanford University Press. Role of Common Features,” Marketing Letters, 22 (4), 405-21. Forehand, M. R., & Deshpandé, R. (2001), “What We See Makes Oyserman, Daphna (2009), “Identity-Based Motivation: Us Who We Are: Priming Ethnic Self-Awareness and Implications for Action-Readiness, Procedural-Readiness, and Advertising Response,” Journal of Marketing Research, 38 Consumer Behavior,” Journal of Consumer Psychology, 19 (3), 336-348. (3), 250-60. Gramzow, Richard H., Constantine Sedikides, A. T. Panter, and Reed, Americus, II and Mark R. Forehand (2016), “The Ebb and Chester A. Insko (2000), “Aspects of Self-Regulation and Flow of Consumer Identities: The Role of Memory, Emotions Self-Structure as Predictors of Perceived Emotional Distress,” and Threats,” Current Opinion in Psychology, 10, 94-100. Personality and Social Psychology Bulletin, 26 (2), 188-205. Reed, Americus, II, Mark R. Forehand, Stefano Puntoni, and Greenwald, Anthony G., Mahzarin R. Banaji, Laurie A. Rudman, Luk Warlop (2012), “Identity-Based Consumer Behavior,” Shelly D. Farnham, Brian A. Nosek, and Marshall Rosier International Journal of Research in Marketing, 29 (4), 310- (2000), “Prologue to a Unified Theory of Attitudes, 21. Stereotypes, and Self-Concept,” in Feeling and Thinking: The Saint Clair, Julian K and Mark R Forehand (2019), “The Many- Role of Addect in Social Cognition and Behavior, Joseph P. Faced Consumer: Consumption Consequences of Balancing Forgas, ed. New York: Cambridge University Press, 308-30. Multiple Identities,” Journal of Consumer Research, 46 (6), Hawkins, Chris (2011). A History of Signatures: From Cave 1011-1030. Paintings to Robo-Signings. CreateSpace. Savary, Jennifer and Ravi Dhar (2020), “The Uncertain Self: How Hong, Sung-Mook, and Salvatore Faedda (1996), “Refinement of Self-Concept Structure Affects Subscription Choice,”Journal the Hong Psychological Reactance Scale,” Educational and of Consumer Research, 46 (5), 887-903. Psychological Measurement, 56 (1), 173-82. Schaupp, L. Christian, and France Bélanger (2005), “A Conjoint Houston, David A., Steven J. Sherman, and Sara M. Baker (1991), Analysis of Online Consumer Satisfaction,” Journal of “Feature Matching, Unique Features, and the Dynamics of Electronic Commerce Research, 6 (2), 95. the Choice Process: Predecision Conflict and Postdecision Shugan, Steven M (1980), “The Cost of Thinking,” Journal of Satisfaction,” Journal of Experimental Social Psychology, 27 consumer Research, 7 (2), 99-111. (5), 411-30. Shultz, Thomas R., Elène Léveillé, and Mark R. Lepper (1999), Kettle, Keri L (2019). “Identity Salience: Understanding When “Free Choice and Cognitive Dissonance Revisited: Choosing Identity Affects Consumption,”Handbook of Research on “Lesser Evils” Versus “Greater Goods”,” Personality and Identity Theory in Marketing, Edward Elgar Publishing. Social Psychology Bulletin, 25 (1), 40-48. Kettle, Keri L., and Gerald Häubl (2011), “The Signature Effect: Tversky, Amos and Eldar Shafir (1992), “Choice under Conflict: Signing Influences Consumption-Related Behavior by Priming The Dynamics of Deferred Decision,” Psychological Science, Self-Identity,” Journal of Consumer Research, 38 (3), 474- 3 (6), 358-61. 489. Winterich, Karen Page and Michael J. Barone (2011), “Warm Glow Kleine, Robert E., Susan S. Kleine, and Jerome B. Kernan (1993), or Cold, Hard Cash? Social Identify Effects on Consumer “Mundane Consumption and the Self: A Social-Identity Choice for Donation Versus Discount Promotions,” Journal of Perspective,” Journal of Consumer Psychology, 2 (3), 209-35. Marketing Research, 48 (5), 855-68. Koo, Minjung, and Ayelet Fishbach (2016). “Giving the Self: Winterich, Karen Page, Vikas Mittal, and William T. Ross, Jr. Increasing Commitment and Generosity Through Giving (2009), “Donation Behavior toward in-Groups and out- Something That Represents One’s Essence,” Social Groups: The Role of Gender and Moral Identity,” Journal of Psychological and Personality Science, 7 (4), 339-348. Consumer Research, 36 (2), 199-214. Zou, Xi, Michael W. Morris, and Verónica Benet-Martínez (2008), “Identity Motives and Cultural Priming: Cultural (Dis) Identification in Assimilative and Contrastive Responses,” Journal of Experimental Social Psychology, 44 (4), 1151-59. Consumer Misunderstandings Chairs: Amit Kumar, University of Texas at Austin, USA Dafna Goor, Harvard Business School, USA

Paper #1: A Little Good Goes an Unexpectedly Long Way: less enjoyable over time, but actually experience more positive out- Underestimating the Positive Impact of Kindness on Recipients comes as discussions continue. As a result, people may under-exploit Amit Kumar, University of Texas at Austin, USA very effective opportunities to enrich themselves through -interac Nicholas Epley, University of Chicago, USA tions with others. Goor, Donnelley, and Norton examine how unfamiliarity can Paper #2: The I Told You So Effect lead to feelings of self-discovery. Specifically, they demonstrate that Ovul Sezer, University of North Carolina at Chapel Hill, USA although consumers tend to like familiar products and experiences, Sal Affinito, University of North Carolina at Chapel Hill, USA unfamiliar products and experiences can foster greater consumer en- Brad Staats, University of North Carolina at Chapel Hill, USA gagement. Engagement with the unfamiliar promotes a clearer under- Paper #3: Keep Talking: (Mis)Understanding the Hedonic standing of who one really is. Importantly, they document evidence Trajectory of Conversation for a particular zone of proximal discovery: Moderately unfamiliar Ed O’Brien, University of Chicago, USA consumption experiences are most likely to facilitate self-discovery. Michael Kardas, University of Chicago, USA Notably, this work showcases several topics consistent with the Juliana Schroeder, University of California at Berkeley, USA theme of this year’s conference, including gift-giving, identity, and Paper #4: The Road Not Taken: Consumption of Unfamiliar interpersonal communication. In the City of Light, we hope to shine Products Increases Feelings of Self-Discovery and Product a light on some mistaken beliefs that appear to exist in consumers’ Engagement minds. Our field is uniquely suited to understanding causes and con- Dafna Goor, Harvard Business School, USA sequences that can suggest steps consumers might take in order to Grant E. Donnelly, The Ohio State University, USA make themselves better off. A better understanding of these misper- Michael I. Norton, Harvard Business School, USA ceptions, we believe, can suggest ways for consumers to improve the quality of their day-to-day lives. SESSION OVERVIEW Nobel Prize-winning economic theory suggests that consumer’s A Little Good Goes an Unexpectedly Long Way: choices are guided by an implicit or explicit calculation of expected Underestimating the Positive Impact of Kindness on value (Becker, 1993). However, consumers may sometimes have Recipients mistaken expectations. Such misunderstandings based on erroneous beliefs or miscalibrated perceptions can result in consumers theoreti- EXTENDED ABSTRACT cally acting rationally—that is, consistent with their expectations— Positive interpersonal contact promotes well-being (Kahne- but nevertheless behaving suboptimally. man & Deaton, 2010). For example, consumers who spend money This session integrates four papers examining these sorts of on others are happier those who spend on themselves (Dunn, Aknin, consumer misunderstandings in daily life. These advanced projects & Norton, 2008). Prosociality, then, can improve consumer welfare. (involving highly-powered field, laboratory, and online experiments) Consumers have many opportunities to act prosocially but may not span topics of broad interest to consumer behavior researchers— engage in prosocial behavior as often in everyday life as might be from prosocial behavior to advice to conversation to engagement optimal. It is sometimes said that “a little good goes a long way.” during consumption. And yet, they are highly connected: Each talk We find that it goes an unexpectedly long way: Performers of acts of reveals a tendency for consumers to miscalculate the complexities of kindness systematically underestimate the positive impact they have their everyday interactions, from the surprising depth of one’s own on recipients. These miscalibrated expectations matter because they and others’ emotional reactions to the surprising pleasures of step- create a barrier to prosocial engagement (Kumar & Epley, 2018). ping outside one’s comfort zone. Together, they elucidate the need Decisions are guided partly by the expected value of action; the ex- for better understanding how to enhance consumer welfare. pected impact of a prosocial act on others may therefore guide one’s Kumar and Epley investigate the beliefs about and reality fol- behavior. lowing prosocial acts. Participants in their experiments predict how We first tested, on a broad range of participants, whether re- recipients will respond to their prosociality and recipients report their cipients of a prosocial act feel more positive than performers of the actual experience. They find that, due to an egocentric bias, those act anticipate. In Experiment 1a, MBA students (N = 106) in a field performing an act of kindness consistently underestimate how posi- experiment performed a variety of self-chosen acts of kindness and tive their recipients will feel. Consumers not recognizing the impact predicted how their recipients would react. Participants engaged in of their prosociality on others can stand in the way of themselves and many different acts, including purchasing coffee, buying flowers, and others being better off. delivering baked goods to others. We compared performer expecta- Sezer et al . explore what they call the “I Told You So Effect” tions to recipients’ actual experience. Performers reported that their in the domain of advice. Hearing someone say “I told you so” after a act was not as “big” as recipients perceived it to be, and reported misstep has been made can hurt relationships, decrease the likelihood expending less time, money, and energy than recipients believed of following subsequent advice, and thereby serve as a barrier to ef- performers had expended (ps < .001). Performers also expected re- fective learning. Nonetheless, people often mistakenly choose to use cipients to feel positive, but recipients reported feeling even more this common phrase when giving advice. positive than performers anticipated (p < .001). Consistent with prior Kardas, Schroeder, and O’Brien study mispredictions about research, performing a prosocial act was a significantly positive ex- the hedonic trajectory of conversation. Participants in these studies perience for performers as well, in this and all experiments (ps < report believing that conversations with another person will become

.001). Performers additionally indicated that they perform prosocial Whereas those who perform an act of kindness might construe acts less often than they’d like (p < .001). their act as relatively “little,” as if they are not doing much at all, Experiment 1b replicated these findings holding the act itself recipients construe it to be significantly “bigger” and of greater relatively constant. The standardized prosocial behavior was sending value than performers expect. Misunderstanding how recipients will a card to someone “just because.” One hundred undergraduates did respond may leave people choosing to engage in prosociality less this and predicted how recipients would feel. We followed-up with often than they want to, thereby representing a barrier to prosocial recipients and measured how they actually felt. Performers again re- interactions. Undervaluing the positive impact of a prosocial action ported that their act was not as “big” as recipients perceived it to be, can therefore keep people from being prosocial enough for both their and also reported expending less time and energy than recipients be- own and others’ well-being. lieved performers had expended (ps < .001). They also significantly underestimated how positive recipients would feel (p < .05). The I Told You So Effect Experiments 2a and 2b provided tests of our hypotheses in designs enabling perfect response rates from recipients, something EXTENDED ABSTRACT we could not obtain in the field settings of Experiments 1a-1b. We From solving complex problems to making purchases, people achieved this by having participants perform a prosocial act for a often rely on advice from others. In fact, consumers rarely make stranger nearby. Experiment 2a involved members of the public (N critical decisions in isolation. Prior research has identified multiple = 84) who were visiting an attraction in a large urban field setting. factors that influence individuals’ use of advice. For instance, deci- Participants at a park gave hot chocolate to another person near an sion-makers weigh advice more heavily when the advice is costly to ice-skating rink. Givers again reported that their act was not as “big” get (Patt, Bowles, & Cash, 2006), or when the task is difficult (Gino as recipients perceived it to be, and significantly underestimated how & Moore, 2007). When people feel confident or experience anger positive it made recipients feel (ps < .001). Experiment 2b (N = 102) (Bonaccio & Dalal, 2006; Gino & Schweitzer, 2008), they are less replicated these effects in a controlled laboratory setting, with par- likely to rely on advice. Similarly, advisor characteristics also im- ticipants giving away gifts from a “lab store” as the act of kindness pact whether people follow their advice. When the advisors are more (ps again < .001). experienced (Feng & MacGeorge, 2006; Goldsmith & Fitch, 1997; We hypothesized that people undervalue the impact of proso- Harvey & Fischer, 1997; Sniezek, Schrah, & Dalal, 2004; Soll & Lar- ciality due to a perspective-based asymmetry in attention paid to rick, 2009; Yaniv, 2004; Yaniv & Milyavsky, 2007) and have greater competence versus warmth in evaluating interpersonal behavior confidence (Phillips, 1999; Sniezek & Buckley, 1995; Sniezek & (Wojciszke, 1994). People focus more on competence when evaluat- Van Swol, 2001; Soll & Larrick, 2009; Van Swol & Sniezek, 2005; ing their own behavior but more on warmth when evaluating others’ Yaniv & Foster, 1997), the advice is weighted more heavily. behavior. This could cause givers to focus inordinately on the details At the same time, however, the consequences of expressions of the act itself: what one is giving and its objective value. Recipients that advisors use depend on the psychological and emotional reac- of an act of kindness, however, care about what is given but also the tion of the targets, and critically on their appraisal of the advisor positive intention and warmth that the act signifies. Experiment 3 statements as helpful or not, because interpersonal aspects in advice tested this explanation by comparing an act of kindness condition to exchange are as important as the quality of the advice (Blunden, a control condition in which participants received the same objec- Logg, Brooks, John, & Gino, 2019). While previous research has fo- tive item but without it resulting from a prosocial act. Forty-nine cused on expressions that prompt advisor credibility and confidence participants gave a cupcake to 49 strangers in a park and predicted (Soll & Larrick, 2009), little is known about the effects of statements their reactions, while another 50 participants predicted the actual re- that backfire. In the current research, we investigate a common and sponses of 50 participants who received a cupcake from an experi- a universal statement that is of critical importance in advice-giving menter (not as an act of kindness). Recipients reported a more posi- contexts: “I told you so.” tive and valuable experience in the kindness condition than in the Building on past research on advice-giving and learning, we conceptualize “I told you so” as a statement that stems from the sat- control condition (pbig < .001; pmood = .06), but those predicting their responses anticipated similar experiences for recipients in these two isfaction of being right. We propose that individuals who hear this conditions (ps > .1). A recipient’s reaction to a prosocial act comes statement view the advice giver as more condescending, less em- from the objective value of the act itself, plus the warmth conveyed pathic, and less trustworthy. We suggest that although it is a very by it. Here, performers may have attended to how much someone common statement that exists in all languages across the world, and would like a cupcake, but not to the fact the cupcake was also a gift is a universal phenomenon, saying I told you so backfires as it harms given in kindness. trust in advice exchange and hurts learning. Experiment 4 (N = 200) examined whether underestimating a In Study 1 (N = 302), participants estimated an individual’s recipient’s emotional reaction also leads to underestimating their be- weight from a photograph for three rounds. Out of 302 participants, havioral reaction. Kindness can spread through indirect reciprocity 133 of them ignored their advisor’s estimate in Round 1, and 169 of (Gray, Ward, & Norton, 2014). Those who receive kindness are more them took the advice. In Round 2, those who didn’t follow the advice likely to behave kindly to others in the future. If people undervalue in Round 1, got another piece of advice, but half of them heard the the positive impact of prosociality on others, they should also under- new advice with “I told you so” while the other half only heard just estimate the magnitude of indirect reciprocity it produces in others. the advice. Among the ones who ignored their partner’s advice in the Participants completed a laboratory version of Experiment 3 (using first round, those who received Round 2 advice with “I told you so” gifts from Experiment 2b), and recipients were later assigned the role perceived the advice to be lower quality than those who received the of “decider” in a dictator game (Camerer, 2003). Performers of the same advice without the I told you so statement, t (131) = 2.64, p = . act predicted their recipient’s behavior. They again underestimated 009, M = 4.11 vs M = 4.75. Similarly, the group that received “I told the positive impact their prosocial act had on the recipient, both on you so” perceived their advisor to be more condescending, t (131) = the recipient’s emotional experience and on their increased kindness 7.69, p = < .001, M = 4.89 vs M = 2.53, and less empathic t (131) = towards a third party (ps < .05). 3.43, p = . 0008, M = 3.54 vs M = 4.24. 970 / Consumer Misunderstandings

In Study 2 (N=500), we employed the same design as Study they and their partner must maintain a conversation—despite the 1, and randomly assigned participants to one of two between-par- aforementioned research suggesting the opposite. ticipants conditions: “I told you so” vs “You were wrong.”. When Two other literatures support this possibility. First, predictions participants didn’t follow their partners’ advice, in Round 2, they about future experiences reflect one’s mental simulations of those either received “I told you so” or “You were wrong” as a message in experiences, but mental simulations are “mere cardboard cutouts of addition to advice for the new round. Out of 500 participants, 276 of reality” (Gilbert & Wilson, 2007, p. 1354). As a result, after people them ignored the advice from their partner in Round 1. Among the initially experience an enjoyable stimulus, they tend to underappre- group who ignored their partner’s advice, those who received “I told ciate the extent to which repeat exposures can reveal exciting new you so” perceived their partners to be more condescending t (274) = information (O’Brien, 2019). People notoriously misunderstand how 8.07, p = < 001, M = 5.01 vs M = 3.71, and less empathic t (274) = experiences build and grow over time. Second, people are especially 4.32, p = .002, M = 1.81 vs M = 2.37, than those who received “You prone to misunderstand the depth and complexity of other people, were wrong” from their partners. given the inaccessibility of others’ elaborate mental lives from an In Study 3 (N=102), participants in the lab completed the same observer perspective (Waytz et al., 2013). For example, people tend weight estimation task, and those who received “I told you so” in to underestimate others’ capacities to generate nuanced opinions and Round 2, along with the advice, perceived their partner to be more experience intense emotions (Pronin, 2008). condescending, t (100) = 9.67, p < 001, M = 5.03 vs M = 2.87, less Put succinctly, these literatures suggest people may underappre- empathic t (100) = 7.23, p < 001, M = 1.81 vs M = 3.56, and less ciate the extent to which they and a new acquaintance will continue trustworthy t (100) = 3.89, p = .004, M = 3.56 vs M = 4.35 than the to find new things to discuss beyond their initial interaction, leading participants who didn’t hear I told you so but received the same ad- people to underappreciate their enjoyment as they continue convers- vice. More importantly, those who received “I told you so” were less ing—even if they enjoyed chatting initially. likely to choose the same partner for the subsequent round, p = .03. In Experiment 1, pairs of strangers (N = 100 pairs) conversed In Study 4 (N=200), participants engaged in 10 rounds of the for one round (3 minutes). Some pairs then predicted their trajectory same task, but as an advisor this time. After each round, they found of enjoyment over four additional rounds. Others actually continued out they were right, and their partner was wrong, and they could conversing and reported their enjoyment round-by-round. We ob- either send “Bummer!” or “Bummer, I told you so!” as a message. served the hypothesized Role × Round interaction (p < .001): Where- 142 out of 200 participants (71%) chose to send the “I told you so” as Predictors expected enjoyment to decline over time (p < .001), message at least once across 10 rounds, and on average, participants Experiencers reported that enjoyment actually increased (p = .010). chose to send this message 3.235 times of the possible 10. About Experiment 2 utilized a within-participant design, with the 23% of participants chose to say I told you so after only Round 1. In same participants first making predictions and then reporting their other words, despite not liking this statement (as found in Studies 1, experiences. We also tested mechanism: Whether people misunder- 2, and 3), when given a chance, participants use the opportunity to stand the hedonic trajectory of conversation because they underes- say “I told you so.” timate how much their pair will have to discuss. Pairs of strangers We contribute to the advice literature by focusing on interper- (N = 50 pairs) completed a similar experiment to Experiment 1. We sonal effects, as opposed to decision accuracy and quality, which also measured conversation material (“How much will you [did has been the primary focus of previous research (Bonaccio & Da- you] have to discuss during this round?”). For enjoyment, we again lal, 2006). We introduce and examine the psychology underlying observed the hypothesized Phase × Round interaction (p = .017): a common advice statement “I told you so” and shed light on the Participants predicted that enjoyment would decline (p < .001), yet importance of empathy and trust in learning and building an advisor- those same participants then experienced similarly high enjoyment advisee relationship. (p = .920). This interaction also emerged for conversation material (p < .001): Participants predicted that conversation material would Keep Talking: (Mis)Understanding the Hedonic diminish (p < .001), yet then maintained ample material (p = .797). Trajectory of Conversation The effect of Phase on changes in enjoyment, via changes in conversation material, was significant b( = -0.38, 95% CI = [-0.80, -0.15]). EXTENDED ABSTRACT If people underestimate enjoyment because they underestimate People engage in conversation every day, from small-talking their pair’s ability to sustain material, they might underestimate en- with colleagues to checking in with close others. We explore one joyment more when having one long conversation with the same common but consequential conversation opportunity: talking with person (deepening one relationship) versus having multiple shorter new acquaintances. People encounter far more opportunities to inter- conversations with different people (broadening their network). Ex- act with new acquaintances than they take (Hill & Dunbar, 2003), yet periment 3 tested this possibility (N = 395 individuals). Procedures doing so forms the basis of richer relationships. Typically, the more were identical to Experiment 2, except participants were randomly people talk, the more they learn, the closer they feel, and the more assigned to either have five rounds of conversation with the same they like each other (Aron et al., 1997; Quoidbach et al., 2019; Reis, partner, or five rounds of conversation each with new partners. Again, 2012). Only by continuing the conversation do two strangers have a there were significant Phase × Round interactions for enjoyment chance of becoming two friends. and conversation material (ps < .001), with participants generally Why might people forego such opportunities? On the one hand, underestimating both. Most interesting, we observed split evidence people cannot always tell whether others are willing to engage at all. for Partner effects. There wasno 3-way interaction for enjoyment (p Thus, people often fail to initially approach others to start any con- = .584): All participants (regardless of partner) predicted declining versation (Boothby et al., 2018; Dunn et al., 2007; Epley & Schro- enjoyment (ps < .001) yet then experienced increasing enjoyment eder, 2014). We build on this idea by examining another barrier that (ps < .019). However, there was a 3-way interaction for conversa- may rise after people have officially met and enjoyed some initial tion material: Same-partner participants were especially likely to un- conversation. Specifically, even after enjoying the start of a conver- derestimate conversation material as compared to different-partner sation, people may believe the experience will quickly grow dull as Advances in Consumer Research (Volume 48) / 971 participants (p < .001). Conversation material again served as the to merely discovering more about an unfamiliar product), which in mediator (b = -0.44, 95% CI = [-0.61, -0.28]). turn, increases engagement. These experiments reveal that people misunderstand the hedon- Three studies demonstrate our effect and suggest a zone of prox- ic trajectory of conversation. Even after enjoying some initial con- imal discovery – moderate rather than extreme levels of unfamiliar- versation, people imagined that further conversation would quickly ity facilitate the process of self-discovery. Consistent with theories grow dull—yet experienced similar or greater enjoyment in reality. in developmental psychology (Vygotsky, 1978), which suggest that Our findings advance research on social forecasting errors. children have a successful and positive experience when learning Conversation provides feedback about others and thus should cali- skills that can be directly applied to previously developed knowl- brate expectations. However, our paradigm compares predictions to edge, we demonstrate that some level of familiarity is needed to in- experiences after one round of conversation and finds that people crease consumers’ self-discovery and further interest in the product. still underestimate enjoyment under these conservative conditions. In Study 1 (N=176, MTurk), participants described a recent Social forecasting errors may arise not only from lack of knowledge purchase of a product that was unfamiliar to them (vs. a new prod- about one’s initial approach but also from misunderstanding how the uct that replaced an existing one) and indicated the extent to which experience will change as they keep talking. the product made them discover and learn about themselves using a Additionally, our participants underestimated enjoyment be- 3-item index (“Using this product I discovered new parts within my- cause they underestimated their pair’s ability to sustain material, self,” “Purchasing this product felt like a discovery,” “This product but also underestimated enjoyment before speaking with multiple helped me to learn about myself”; α=.82). In both conditions, par- partners, when material should be ample. Perhaps people are also ticipants wrote about similar purchase categories (clothing, cleaning (mistakenly) worried about how fatiguing social interaction is (Zel- supplies, yard tools, and electronics). Importantly, as expected, un- enski et al., 2013). Future research should further unpack whether familiar products increased self-discovery (Munfamiliar=4.23, SD=1.47) people are concerned that they themselves (versus their partner) will compared to new products (Mnew=3.76, SD=1.61; F(1,174)=3.94, struggle to generate material. Finally, future research should exam- p=.049). ine downstream effects. People may exit conversations earlier than In Study 2 (N=274), undergraduates were randomly assigned they should for their own happiness, just as our participants underap- to drink a small sample of Watermelon (unfamiliar flavor) vs. Or- preciated how much further conversation would measure up to initial ange (familiar flavor) juice. A pre-test (N=100) confirmed that Wa- conversation. Accordingly, people may keep to themselves or pursue termelon is a less familiar juice flavor than orange (p<.001). After impoverished interactions through social media (Kross et al., 2013), tasting the juice, participants indicated how interested and engaged while returning to those we already know may prove more reward- they were by the juice using a 6-item scale (e.g., “I feel engaged ing. in the experience of drinking,” “I stopped several times to examine the juice and how it’s served,” “I’d be curious to try it again in the The Road Not Taken: Consumption of Unfamiliar future”; α=.79). They also indicated how much they like the juice in Products Increases Feelings of Self-Discovery and general (one-item), and the extent to which drinking the juice made Product Engagement them discover and learn about themselves using a 6-item scale (e.g., the juice “elicited some self-reflection,” “opened me to new experi- EXTENDED ABSTRACT ences and a new aspect within me,” “made me reevaluate my prefer- As consumers, our identity and roles in life are intertwined with ences,” and “felt like a part of a journey of self-discovery”; α=.85). our possessions and consumption behaviors (Belk, 1988). The types The results revealed that the unfamiliar juice increased interest and of products we purchase are influenced by our personality and val- engagement (Munfamiliar=3.62, SD=1.28) compared to the familiar fla- ues (Donnelly, Iyer, & Howell, 2012; Matz, Gladstone, & Stillwell, vor (Mfamiliar=3.05, SD=1.21; F(1,272)=14.70, p<.001) and this effect 2016) and consumers often turn to durable goods like clothing and was mediated by feelings of self-discovery (a×b=.3287, SE=.0953, accessories to construct and convey an actual, nostalgic, or aspira- 95% CI=[.1452,.5202]. Interestingly, the unfamiliar experience de- tional self to others (Chernev, Hamilton, & Gal, 2011; Goor, Orda- creased liking (Munfamiliar=3.75, SD=1.82) compared to the familiar bayeva, Keinan, & Crener, 2019). This research examines a different juice (Mfamiliar=4.49, SD=1.59; F(1,272)=12.66, p<.001). perspective of consumption experience and argues that products not Study 3 tested an important boundary condition: the potential only facilitate self-expression, but can also help consumers discov- zone of proximal discovery. Participants (N=303, MTurk) ranked er who they are. Specifically, we demonstrate that unfamiliar (vs. their familiarity with five music genres: country, rock, rap, jazz, and familiar) consumption experiences may increase the sense of self- classical. They were then randomly assigned to one of three condi- discovery, and thus lead to greater engagement with the unfamiliar tions: listening to a song from a familiar genre (ranked 1/5), slightly product. unfamiliar genre (ranked 3/5), or a very unfamiliar genre (ranked Prior research demonstrates that consumers seek extraordi- 5/5). After listening to a brief 60-second music clip, participants in- nary experiences to signal status (Keinan & Kivetz, 2010) or to dicated the extent to which listening to the music helped them dis- feel intensely alive (Arnould & Price, 1993). Furthermore, major cover and learn about themselves using a 5-item scale (e.g., “felt like life events, like starting a new job or starting a new relationship, a part of a journey of self-discovery,” “made me reevaluate my mu- can encourage consumers to question previous values and ideas and sic preferences,” and “advanced my self-knowledge”; α=.86). They bring about new roles in life (Schouten, 1991). Conversely, repeated also evaluated how much listening to the song helped them learn consumption can increase satiation and adaptation, and reduce con- about music using 3-item scale (listening to that song “is like tak- sumer enjoyment and product evaluation (Bhattacharjee & Mogil- ing a music lesson,” “made me feel like I know more about music,” ner, 2014; Frederick & Loewenstein, 1999; Gershon & Smith, 2019). “made me realize that music has subtle nuances”; α=.87). The re- Contributing to this literature, we find that basic albeit unfamiliar sults indicated that the slightly unfamiliar genre increased feelings of consumption experiences, such as listening to an unfamiliar song, self-discovery (M3/5=3.13, SD=1.49) compared to the familiar genre eating unfamiliar food, or using an unfamiliar product, provide con- (M1/5=2.71, SD=1.34; p=.047). The effect of listening to a very un- sumers opportunities to discover more about themselves (as opposed familiar genre was non-significant compared to the other conditions 972 / Consumer Misunderstandings

(M5/5=2.93, SD=1.63, p’s>.300). Importantly, the effect of genre fa- Dunn, E. W., Biesanz, J. C., Human, L. J., & Finn, S. (2007). miliarity on music-discovery was not significant across conditions Misunderstanding the affective consequences of everyday (p’s>.488). social interactions: The hidden benefits of putting one’s best Our work contributes to the literature on consumer psychology, face forward. Journal of Personality and Social Psychology, suggesting that consumers might be more engaged using unfamiliar 92, 990-1005. (vs. familiar) products in order to learn about themselves through Epley, N., & Schroeder, J. (2014). Mistakenly seeking solitude. interacting with the products. Preliminary results further suggest Journal of Experimental Psychology: General, 143, 1980- that older consumers are more susceptible to self-discovery using 1999. unfamiliar products. Our research has also important practical im- Feng, B., & MacGeorge, E. L. (2006). Predicting receptiveness to plications. Today, marketers often use big-data analyses to provide advice: Characteristics of the problem, the advice-giver, and consumers recommendations that are based on product familiarity the recipient. Southern Journal of Communication, 71(1), and similarity (Linden, Smith, & York, 2003), even though consum- 67–85. ers regularly express preferences for a greater variety of products Frederick, S., and Loewenstein, G. (1999). Hedonic Adaptation in (Ratner and Kahn 2002; Simonson 1990) and unique experiences Well-being: Foundations of Hedonic Psychology, 302-329. (Keinan & Kivetz, 2010). Our research demonstrates that marketers Gershon, R., and Smith, R. K. (2019). Twice-told tales: Self- could increase engagement and interest in their products by giving repetition decreases observer assessments of performer consumers the opportunity to experience new products (e.g., new authenticity. Journal of Personality and Social Psychology. collections, samples, extensions) that encourage greater self-discov- Forthcoming. ery. Gilbert, D. T., & Wilson, T. D. (2007). Prospection: Experiencing the future. Science, 317, 1351-1354. REFERENCES Gino, F., & Moore, D. A. (2007). Effects of task difficulty on use of Arnould, E. J., and Price, L. L. (1993). River magic: Extraordinary advice. Journal of Behavioral Decision Making, 20(1), 21–35. experience and the extended service encounter. Journal of Gino, F., & Schweitzer, M. (2008). Blinded by anger or feeling Consumer Research, 20(1), 24-45. the love: How emotions influence advice taking.Journal of Aron, A., Melinat, E., Aron, E. N., Vallone, R. D., & Bator, R. Applied Psychology, 93(5), 1165–1173. J. (1997). The experimental generation of interpersonal Goldsmith, D. J., & Fitch, K. (1997). The normative context of closeness: A procedure and some preliminary findings. advice as social support. Human Communication Research, Personality and Social Psychology Bulletin, 23, 363-377. 23, 454–476. Becker, G. S. (1993). Nobel lecture: The economic way of looking Goor, D., Ordabayeva, N., Keinan, A., and Crener, S. (2019). The at behavior. Journal of Political Economy, 101, 385-409. impostor syndrome from luxury consumption. Journal of Belk, R. W. (1988). Possessions and the extended self. Journal of Consumer Research. Forthcoming. Consumer Research, 15(2), 139-168. Gray, K., Ward, A.F., & Norton, M.I. (2014). Paying it forward: Bhattacharjee, A., and Mogilner, C. (2013). Happiness from Generalized reciprocity and the limits of generosity. Journal of ordinary and extraordinary experiences. Journal of Consumer Experimental Psychology: General, 143, 247-254. Research, 41(1), 1-17. Harvey, N., & Fischer, I. (1997). Taking advice: Accepting Blunden H., Logg J. M., Brooks A. W., John L. K., Gino F. (2019). help, improving judgment, and sharing responsibility. Seeker beware: the interpersonal costs of ignoring advice. Organizational Behavior and Human Decision Processes, 70, Organizational Behavior and Human Decision Processes. 150, 117–133. 83–100. Hill, R. A., & Dunbar, R. I. M. (2003). Social network size in Bonaccio, S., & Dalal, R. S. (2006). Advice taking and advice humans. Human Nature, 14, 53-72. giving in decision making: An integrative review of the Kahneman, D., & Deaton, A. (2010). High income improves literature. Organizational Behavior and Human Decision evaluation of life but not emotional well-being. Proceedings of Processes, 101, 127–151. the National Academy of Sciences, 107, 16489-16493. Boothby, E. J., Cooney, G., Sandstrom, G. M., & Clark, M. S. Keinan, A., and Kivetz, R. (2010). Productivity orientation and the (2018). The liking gap in conversations: Do people like us consumption of collectable experiences. Journal of Consumer more than we think? Psychological Science, 29, 1742-1756. Research, 37(6), 935-950. Camerer, C. (2003). Behavioral game theory: Experiments in Kross, E., Verduyn, P., Demiralp, E., Park, J., Lee, D. S., Lin, N., strategic interaction. Princeton: Princeton University Press. ... & Ybarra, O. (2013). Facebook use predicts declines in Chernev A., Hamilton R., and Gal D. (2011). Competing for subjective well-being in young adults. PloS One, 8, e69841. consumer identity: Limits to self-expression and the perils of Kumar. A., & Epley, N. (2018). Undervaluing gratitude: Expressers lifestyle branding. Journal of Marketing, 75 (3), 66-82. misunderstand the consequences of showing appreciation. Donnelly, G. E., Iyer, R., and Howell, R. T. (2012). The big five Psychological Science, 29, 1423-1435. personality traits, material values, and financial well-being Linden, G., Smith, B., and York, J. (2003). Amazon. com of self-described money managers. Journal of Economic recommendations: Item-to-item collaborative filtering.IEEE Psychology, 33(6), 1129-1142. Internet computing, (1), 76-80. Dunn, E. W., Aknin, L. B., & Norton, M. I. (2008). Spending Matz, S. C., Gladstone, J. J., and Stillwell, D. (2016). Money buys money on others promotes happiness. Science, 319, 1687- happiness when spending fitsour personality. Psychological 1688. Science, 27(5), 715-725. O’Brien, E. (2019). Enjoy it again: Repeat experiences are less repetitive than people think. Journal of Personality and Social Psychology, 116, 519-540. Advances in Consumer Research (Volume 48) / 973

Patt, A., Bowles, H. R., & Cash, D. (2006). Mechanisms for Soll, J. B., & Larrick, R. P. (2009). Strategies for revising enhancing the credibility of an adviser: Prepayment and judgment: How (and how well) people use others’ opinions. aligned incentives. Journal of Behavioral Decision Making, Journal of Experimental Psychology: Learning, Memory, and 19,347–359. Cognition, 35, 780–805. Phillips, J. M. (1999). Antecedents of leader utilization of staff Van Swol, L. M., & Sniezek, J. A. (2005). Factors affecting input in decision- making teams. Organizational Behavior and the acceptance of expert advice. British Journal of Social Human Decision Processes, 77, 215–242. Psychology, 44, 443–461. Pronin, E. (2008). How we see ourselves and how we see others. Vygotsky, L. (1978). Interaction between learning and development. Science, 320, 1177-1180. Readings on the Development of Children, 23(3), 34-41. Quoidbach, J., Taquet, M., Desseilles, M., de Montjoye, Y-A., & Waytz, A., Schroeder, J., & Epley, N. (2014). The Lesser Minds Gross, J. (2019). Happiness and social behavior. Psychological Problem. In P. Bain, J. Vaes, & J. P. Leyens (Eds.), Humanness Science, 30, 1111-1122. and dehumanization (pp. 4967). London, England: Psychology Ratner, R. K. and Kahn, B. E. (2002). The impact of private versus Press. public consumption on variety-seeking behavior. Journal of Wojciszke, B. (1994). Multiple meanings of behavior: Construing Consumer Research, 29(2), 246-257. actions in terms of competence or morality. Journal of Reis, H. T. (2012). Perceived partner responsiveness as an Personality and Social Psychology, 67, 222-232. organizing theme for the study of relationships and well- Yaniv, I. (2004). Receiving other people’s advice: Influence and being. In L. Campbell & T. J. Loving (Eds.), Interdisciplinary benefit. Organizational Behavior and Human Decision research on close relationships: The case for integration (pp. Processes, 93(1), 1–13. 27-52). Washington, DC: American Psychological Association. Yaniv, I., & Milyavsky, M. (2007). Using advice from multiple Schouten, J. W. (1991). Personal rites of passage and the sources to revise and improve judgment. Organizational reconstruction of self. ACR North American Advances. Behavior and Human Decision Processes, 103, 104–120. Simonson, I. (1990). The effect of purchase quantity and timing Yaniv, I., & Foster, D. P. (1997). Precision and accuracy on variety-seeking behavior. Journal of Marketing Research, of judgmental estimation. Journal of Behavioral Decision 27(2), 150-162. Making, 10, 21–32. Sniezek, J. A., & Buckley, T. (1995). Cueing and cognitive conflict Zelenski, J. M., Whelan, D. C., Nealis, L. J., Besner, C. M., in judge-advisor decision making. Organizational Behavior Santoro, M. S., & Wynn, J. E. (2013). Personality and affective and Human Decision Processes, 62, 159–174. forecasting: Trait introverts underpredict the hedonic benefits Sniezek, J. A., Schrah, G. E., & Dalal, R. S. (2004). Improving of acting extraverted. Journal of Personality and Social judgment with prepaid expert advice. Journal of Behavioral Psychology, 104, 1092-1108. Decision Making, 17, 173–190. Sniezek, J. A., & Van Swol, L. M. (2001). Trust, confidence, and expertise in a judge- advisor system. Organizational Behavior and Human Decision Processes, 84, 288–307. Brand Misbehavior: Integrating Multiple Perspectives Chairs: Mansur Khamitov, Nanyang Technological University, Singapore Andrea Giuffredi-Kähr, University of Zurich, Switzerland

Paper #1: Wilt Thou Forgive That Sin? – The Influence of Brand et al. reconcile prior literature on misbehaving brands by document- Heritage on Consumer Reactions to Brand Transgressions ing that depending on the transgression agency responsible, type of Amogh Kumbargeri, Indian Institute of Management transgression, brand remedial measures, and brand apology, affec- Ahmedabad, India tive commitment can predictably ‘buffer’ versus ‘amplify’ consumer Sudipta Mandal, Indian Institute of Management Indore, India backlash against the misbehaving brand. The effect is mediated by Mansur Khamitov, Nanyang Technological University, shame. In the final paper, Chang and Jain zoom in on self-brand con- Singapore nection to examine whether consumer materialism and brand type can mitigate the negative impact of brand transgressions. In particu- Paper #2: The Dirty Thirty of Brand Personality: Narcissistic, lar, materialism influences the extent to which consumers forgive Machiavellian, and Psychopathic Brands material versus experiential brands following quality-related brand Lucia Malär, University of Bern, Switzerland transgressions. Mansur Khamitov, Nanyang Technological University, This year’s conference theme is “Rendez-Vous in the City of Singapore Light” with different approaches, paradigm perspectives, and mul- Andrea Giuffredi-Kähr, University of Zurich, Switzerland tiple interests called to come together to focus on consumer research Paper #3: Does it Hurt or Protect? The Effect of Affective questions. This session embodies the spirit of integrating multiple Commitment on Unethical Behavior toward Transgressing perspectives on the topic of brand misbehavior. Specifically, the Brands four papers elucidate how consumer responses to brand misbehav- Nitika Garg, University of New South Wales, Australia ior are impacted by a series of consumer-related factors (material- Rahul Govind, University of New South Wales, Australia ism), brand-related factors (brand heritage, brand personality, brand Jatinder J. Singh, EADA Business School Barcelona, Spain type, brand transgression type and responsibility, brand remedial Wayne D. Hoyer, University of Texas, Austin, USA measures), and consumer-brand connection factors (affective com- Paper #4: Consumers’ Responses to Brand Transgressions: Role mitment, self-brand connection). To this end, the proposed special of Brand Type and Materialism session brings together researchers from nine universities in six Sylvia S. Chang, University of Washington, USA countries across all stages of an academic researcher: doctoral stu- Shailendra Pratap Jain, University of Washington, USA dent, assistant, associate, full, and chaired professor. We expect this session to appeal to a broad audience, including researchers interest- SESSION OVERVIEW ed in brand transgressions, product/service failures, consumer-brand The instances when brands violate consumers’ expectations and relationships, loyalty, marketplace ethics and morality. commit misbehavior attract ever-increasing attention and possess omnipresent damage potential which is amplified by social channels Wilt Thou Forgive That Sin? – The Influence of proliferation, consumer activism, and viral dissemination of content. Brand Heritage on Consumer Reactions to Brand This session centers around the following questions: 1) how do con- Transgressions sumers react to misbehaving brands, 2) how do these reactions vary as a function of different factors related to the consumer, brand, and EXTENDED ABSTRACT consumer-brand connection, and 3) what is the downstream impact of brand misbehavior on consumers’ brand perceptions and brand- Introduction related behaviors? Collectively, the four papers shed light on con- On April 20, 2010, a methane gas release triggered an explosion sumers’ differential responses to explicit or perceived brand trans- on the British Petroleum (BP)-run Macondo Prospect site killing 11 gressions in terms of their attitudes, forgiveness, purchase intentions, workers and causing incalculable damage to surrounding wildlife. and behaviors toward those brands. The session contributes to the BP, a British company with an origin dating back to 1909, was found emerging literature on consumer reactions to brand misbehavior guilty of this environmental catastrophe. This incident constitutes an by highlighting novel multiplex research drawing on concepts and act of transgression, defined as any violation of rules guiding con- theories from interpersonal relationships psychology, construal level sumer-brand relationships (Aaker et al. 2004). In this research, we theory, the dark triad of personality, brand relationship, commitment, ask the question: When a brand commits a transgression, are consum- and materialism theories. ers more willing to forgive heritage brands than non-heritage ones? In the first paper, Kumbargeri et al. demonstrate the impact Brand transgressions often induce negative reactions from consum- of brand heritage on consumer responses to brand transgressions. ers, which if not managed appropriately, can severely tarnish brand The authors find that consumers are more likely to forgive heritage reputation putting the very existence of a brand at risk (Bechwati and (vs. non-heritage) brands and exhibit less negative attitudes, word- Morrin 2003; Khamitov, Grégoire, and Suri 2020; Rotman, Khami- of-mouth and unethical punishment behaviors toward such brands, tov, and Connors 2018). which is mediated by differentially induced construal level. In the second paper, Malär et al. find that misbehaving brands possess dis- Theoretical Framework tinct negative brand personality traits. Transferring the dark triad Urde, Greyser, and Balmer (2007, pp. 4-5) define brand heri- of narcissism, machiavellianism, and psychopathy to the branding tage as “a dimension of a brand’s identity found in its track record, context, the paper measures dark brand personality with a novel longevity, core values, use of symbols and particularly in an organi- thirty-item scale and assesses its downstream consequences on con- zational belief that its history is important.” Brand heritage draws on sumers attitudes and behavioral intentions. In the third paper, Garg consumer’s distinct cognitive ability to think temporally in terms of a

Advances in Consumer Research 974 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 975 brand’s track record, longevity, and history as important components negative attitudes and negative word-of-mouth toward them in the of its identity (Urde et al. 2007). Here, we advance this view by fo- event of a brand transgression, thus providing evidence in support of cusing on construal level associated with heritage brands. the core effects. Experiment 2 replicated the main effect and further Construal level theory suggests that temporal distance (vs. demonstrated that the impact of brand heritage on consumer forgive- proximity) is linked to a more abstract (vs. concrete) representation ness and post-transgression attitudes is mediated by construal level of information (Kyung et al. 2010; Liberman et al. 2002; Trope and (b = .134, SE = .079, 95% CI [.004, .319]). Experiment 3 provided Liberman 2003, 2010). Applied to our context, heritage (vs. non-her- additional evidence for the proposed mechanism by showing mod- itage) brands, owing to their emphasis on temporal distance, should, eration by construal levels. Experiment 4 showed that consumers are by extension, induce a psychologically distal (vs. proximal) perspec- less likely to unethically punish the transgressing heritage brand as tive and, therefore, an abstract (vs. concrete) construal. We pose that compared to a non-heritage brand and additionally replicated the re- this is likely to help them reduce negative outcomes resulting from sults of experiments 1A and 1B. Experiment 5 showed respondents brand transgressions. Some preliminary support for this proposi- are less likely to express concern in case of a transgression by a heri- tion originates from McGraw and Warren (2010) and McGraw et tage brand as compared to a non-heritage brand along with replicat- al. (2012), who found that psychologically distant tragedies induce ing findings of the previous experiments. humor as they are perceived less threatening and benign violations. Furthermore, Gong and Medin (2012) demonstrated that abstract (vs. Contributions and Implications concrete) construal level leads to more positive judgments of moral We contribute to the academic literature on the psychology of transgressions. Taken together, we expect consumers to perceive the brand transgressions and brand heritage by highlighting brand heri- transgressions by heritage (vs. non-heritage) brands less negatively tage as one of the strategies to diminish the negative impact of brand due to the abstract (vs. concrete) construal that such brands induce. transgressions on consumer responses (Khamitov et al. 2020). We Consequently, we expect consumers to be more forgiving of heritage also contribute back to construal level theory (Trope and Liberman (vs. non-heritage) brands and exhibit less negative attitudes, reduced 2010) by showing that the influence of construal level extends to the negative word-of-mouth and unethical punishment behaviors to- evaluation of brands in the wake of a brand transgression. The cur- wards heritage (vs. non-heritage) brands in the event of a transgres- rent set of studies has substantial managerial implications as it sheds sion. And this effect of brand heritage should be mediated by the light on brand heritage as an effective strategy to buffer the nega- abstract construal induced by the heritage brand. tive consumer attitudinal and behavioral reactions in the increasingly prevalent event of a brand transgression. Methodology Six experiments across five different product categories and The Dirty Thirty of Brand Personality: Narcissistic, real as well as hypothetical brands investigate the impact of brand Machiavellian, and Psychopathic Brands heritage on consumer responses to brand transgressions. Experiment 1A was designed as a single factor (brand type: heritage vs. non- EXTENDED ABSTRACT heritage) between-subjects study with post-transgression attitudes toward a transgressing university brand and negative word-of-mouth “Boeing is a silver-haired, middle aged white man in a $3000 as the dependent variables. Experiment 1B sought to extend find- suit. He is well-connected and respected in important social ings of Experiment 1A in a different setting (health drink) and with and political circles that help him grow his business. Due to his an additional dependent variable (consumer forgiveness) to establish obsession with quarterly profits, he fails to spend time with his own the robustness of the hypothesized effects. Experiment 2 investigated family. His wife is unsatisfied and takes opiods to cope with her the mechanism underlying the hypothesized effect by establishing lost sense of self, his teenage daughter hates him, and his son is construal level mediation. This study measured the construal level spiraling out of control via drug abuse.” using the BIF scale (Vallacher and Wegner 1989) and included con- This description of the Boeing brand from a consumer inter- sumer forgiveness and post-transgression attitudes toward the brand view provides preliminary evidence that certain brands are not as- as the dependent variables and brand heritage as the independent sociated with positive personality attributes (e.g., “exciting”, “down- variable. Experiment 3 sought to further substantiate the mechanism to-earth”, “cool”; Aaker 1997), but rather with a dark personality in a by testing for moderation by construal level. If the focal effect is in- similar way that humans have dark personalities. deed mediated by construal level, then the effect should be manifest In our hyperconnected world, firms disseminate branded infor- (attenuated) when the abstract (concrete) construal is independently mation increasingly quickly and broadly (Swaminathan et al. 2020). manipulated. This study was conceived as a 2 (brand type: heritage This has led to more transparency where bombarded and frustrated vs. non-heritage) x 2 (construal: abstract vs. concrete) between-sub- consumers increasingly use their power to attack brands (e.g., con- jects design study. In Experiment 4, using the unethical punishment sumer brand sabotage; Kähr et al. 2016). These developments have matrix task adapted from Mazar, Amir, and Ariely (2008), we aimed elevated the likelihood that consumers witness brands’ undesirable to provide behavioral evidence for our hypothesized effect in the behaviors (e.g., manipulation of consumers, exploitation of employ- context of two real brands – one heritage and the other non-heritage. ees, pushy or arrogant messages) and, consequently, attribute nega- Finally, experiment 5 tests the robustness of the hypothesized effect tive personality characteristics to such brands. using a single brand positioned along the heritage vs. non-heritage Despite the sheer prevalence of dark brand personalities in the lines. In this experiment, as an additional behavioral measure, re- marketplace, prior research has primarily focused on positive and fa- spondents were asked to provide their email IDs if they wished to vorable brand personalities (e.g., Aaker 1997; Warren and Campbell express their concern about the transgression. 2014). Nevertheless, there are some initial studies that hint at exis- Results tence and magnitude of negative brand personality dimensions. For Experiments 1A and 1B demonstrated that consumers are more example, Geuens, Weijters and De Wulf (2009) include aggressive- likely to forgive heritage (vs. non-heritage) brands and exhibit less ness, Grohmann (2009) examines some negatively-valenced traits pertaining to gender dimensions of brand personality (e.g., aggres- 976 / Brand Misbehavior: Integrating Multiple Perspectives sive and dominant), and Rojas-Méndez and Papadopoulos (2012) as are unethical punishment intentions (Mhigh=1.52, SD=1.00; add brand neuroticism to their brand personality measure. Relatedly, Mlow=1.11, SD=.41; t(282)=-4.52, p=.001). Also, the higher the DT,

Awad and Youn (2018) show that narcissistic consumers love narcis- the more likely a consumer behaves dishonestly (Mhigh=1.32, SD=.47; sistic, anthropomorphized brands. However, to date, a systematic un- Mlow=1.19, SD=.39; t(282)=-2.47, p=.001). Lastly, the brand was derstanding of the dark traits of brand personality is missing as virtu- evaluated less favorably in the high DT condition than in the low DT ally all prior studies have only peripherally focused on the dark side condition: brand attitudes (Mhigh=1.60, SD=.90; Mlow=5.35, SD=1.26; of brand personality and/or treated those as rather an afterthought. t(282)= 28.82, p=.001) and purchase intentions (Mhigh=1.70, SD=.93;

Against this backdrop, the current research conducted six Mlow=5.27 SD=1.33; t(282)=26.26, p=.001). studies to systematically examine the dark side of brand personal- In sum, this research contributes to the brand personality litera- ity primarily drawing on the important concept of dark triad from ture (Aaker 1997; Geuens et al. 2009) by providing a first systematic personality psychology that consists of narcissism (i.e., admiration empirical examination of the dark traits of brand personality along seeking, feelings of superiority; Morf and Rhodewalt 2001), machia- with an effective, reliable measure to capture dark brand personal- vellianism (i.e., immoral beliefs, pursuit of self-beneficial goals and ity. Specifically, we draw on the important concept of the dark triad manipulation tactics; Rauthmann and Will 2011), and psychopathy from personality psychology (i.e., narcissism, machiavellianism, and (i.e., cold affect and antisocial behaviors; Neumann, Hare, and New- psychopathy) and bring it to the context of brands, showing that DT man 2007). impacts a series of downstream consequences. So doing, we also As a first step in the dark triad (DT) of brand personality item contribute back to the dark triad literature in personality psychology generation (Study 1a), we developed an extensive battery of 567 (Morf and Rhodewalt 2001; Neumann et al. 2007; Rauthmann and items primarily based on limited negative brand personality items in Will 2011) by demonstrating that the influence of dark triad dimen- the literature, personality psychology scales, and a series of qualita- sions meaningfully extend to the branding context. These contribu- tive consumer interviews. We eliminated items that were overlap- tions are managerially relevant: a more detailed understanding and ping or not applicable to brands. In the second stage (Study 1b), we effective measurement of the dark side of a brand’s personality may had six branding and personality psychology experts rate the items in support managers in assessing whether their brand suffers from such terms of their conceptual map on their respective trait and applicabil- a personality and provide an opportunity to intervene and counteract. ity to the branding context. Next, we initially administered and validated our remaining set Does it Hurt or Protect? The Effect of Affective of items. In Study 2a, 900 MTurk US participants were randomly Commitment on Unethical Behavior toward assigned to one brand from a sample of 13 pretested brands and in- Transgressing Brands dicated their level of agreement with the initial DT items (1 strongly disagree-7 strongly agree). In an exploratory factor analysis, a three- EXTENDED ABSTRACT factor solution emerged. We tested the three-dimensional structure Brand transgressions seem to have become more prevalent and in a confirmatory factor analysis using AMOS. Items were removed receive greater media coverage, in recent times (Dick 2019). When one at a time based on modification indices and squared multiple brands transgress, consumers can become punitive by engaging in correlations (below .6). These processes resulted in a final set of 30 negative behaviors including spreading negative word-of-mouth items. Examination of different alternative measurement models (Grégoire and Fisher 2008; Wangenheim 2005), withdrawing from (e.g., two factors) indicated that the three-factor correlated model interactions with the brand (Grégoire and Fisher 2008; Huber et al. fits the data best (χ2(402)=1460, χ2/df=3.63, NNFI=.96, CFI=.96, 2010), brand boycotting (Klein, Smith, and John 2004), switching GFI=.88, SRMR=.03, RMSEA=.06, SRMR=.03). Study 2b validat- to rival brands (Bechwati and Morrin 2003), and even engaging in ed our shortlisted set of items with a new sample and a self-selected brand sabotage (Kähr et al. 2016; Nyffenegger et al. 2018). brand approach. The final items reflect three dark-sided brand per- One potential protective mechanism against the possible nega- sonality dimensions: narcissism (e.g., BRAND pretends to be more tive responses in the face of such incidents might be affective com- important than it actually is; demonstrates arrogant behavior), ma- mitment towards the brand. Affective commitment is defined as the chiavellianism (e.g., BRAND does not care about others unless it customer’s emotional attachment to a particular brand based on con- benefits its bottom line; manipulates others to get ahead), and psy- sumers’ identification with that brand (Allen and Meyer 1990) and chopathy (e.g., BRAND tends to be unconcerned with the morality reflects their desire to maintain an on-going relationship with the of its actions; appears not to care about the well-being of others). brand (Jones et al. 2010). Recent research has suggested that one of To establish discriminant validity, Study 3 tested brand atti- the key antecedents of affective commitment is self-brand connec- tudes, Aaker’s five brand personality dimensions, and brand reputation (Turri et al. 2013) while as a consequence, it is a strong driver of tion against a second-order factor model of DT. Discriminant valid- re-purchase intentions toward the preferred brand (Fullerton, 2005; ity was supported according to the Fornell-Larcker criterion for all Jones et al. 2010) and brand loyalty (Evanschitzky et al. 2006; Kim comparisons. et al. 2008). Study 4 provides a preliminarily investigation of the DT dimen- While prior literature has extensively documented the posi- sions’ impact on downstream brand attitudes and behavioral inten- tive outcomes of affective commitment, it is less definitive on how tions. More specifically, we posit that a highly dark brand personal- it might impact consumer response when a brand transgresses. Re- ity has leads to a) greater hostile and unethical intentions as well search suggests that affective commitment could grant the trans- as dishonesty towards the brand and b) decreased brand attitudes gressing brand a certain level of immunity or ‘buffering’ as strong and purchase intentions. We manipulated DT using fictitious brand brand relationship supports the brand during a service failure (De- descriptions with two conditions: a brand high vs. low on dark traits. Witt and Brady 2003; Hess, Ganesan, and Klein 2003). In contrast, Our 30 items served as a manipulation check. literature also presents evidence of an ‘amplifying’ effect where high Results confirm the expected effects: a) the higher the DT, commitment leads to significantly lower brand evaluations after a the greater consumers’ hostile intentions towards this brand are service failure (Gregoire et al. 2009) because committed consum-

(Mhigh=2.5, SD=.1.30; Mlow=1.22, SD=.51; t(282)=-10.92, p=.001) ers consider brand transgression as a normative violation, leading to Advances in Consumer Research (Volume 48) / 977 an increased desire for retaliation (Grégoire and Fisher 2006, 2008; eral theoretically and managerially relevant moderators; both trans- Grégoire et al. 2009). In a series of four studies, the current research gression- and brand-related factors. This is an important contribution seeks to address this paradox in the novel context of unethical con- because not only do these provide a more nuanced understanding of sumer behavior toward the transgressing brand and importantly, es- affective commitment’s effect but also suggest actionable response tablishes ecologically valid moderators such as transgression agency, strategies to a brand. Finally, we establish a unique mediator of the transgression type, and brand responses, of this effect along with the relationship between affective commitment and unethical behavior mediating role of shame which arises due to high affective commit- toward the transgressing brand – shame, while ruling out alternative ment toward the transgressing brand. The studies use real brands explanations such as anger and betrayal for consumer retaliation. (Coca-Cola/ Nike), their transgressions, and measure ‘actual cheating behavior’ (Winterich, Mittal, and Morales 2014) (except Study Consumers’ Responses to Brand Transgressions: Role of 2). Brand Type and Materialism Studies 1 and 2 test the moderating effect of transgression- related factors; agency responsible for the transgression (direct – EXTENDED ABSTRACT brand vs. indirect – franchisee) and the type of transgression (per- Marketing literature on consumer-brand relationships has found formance-based (PBT) vs. value-based (VBT)), respectively. Study that a strong brand relationship is beneficial for the brand, such that 1 employed a continuous (affective commitment) x 2 (transgression consumers strongly connected to a brand are less likely to lower their agency: direct, indirect) between-subjects design (n=158). Analysis judgments of the brand when faced with negative brand informa- reveals a significant interactive effect on the amount of cheating () tion (Ahluwalia et al. 2000; Swaminathan, Page, and Gürhan‐Canli such that the amount of cheating varies across affective commitment 2007). Research has found that when a brand negatively violates consumers’ expectations, consumers process and respond to such but only in the direct transgression condition (Mhigh-direct= 5.94 vs. transgressions differently, depending on the extent to which the Mlow-direct = .66, p < .001) and not in the indirect condition (Mlow-indirect = brand is tied to their identity. In particular, the degree to which con- 3.38 vs. Mhigh-indirect = 2.33, ns). Study 2 (n=147) uses purchase intentions as the dependent variable and finds a significant interaction be- sumers view the brand as part of themselves (i.e., self-brand con- tween transgression type (VBT vs. PBT) and affective commitment nection) mitigate the impact of negative brand information (Cheng, . Importantly, it shows that affective commitment amplifies punitive White, and Chaplin 2012). When consumers view a brand as part of action for VBT (lowered purchase intentions)1 ), but has a buffering who they are, they construe negative information about the brand as effect for PBT ). Further, higher levels of shame emerge for high (vs. a self-threat, which in turn elicits efforts to defend their self-views, low) affective commitment ), and for VBT (vs. PBT) ). As predicted, including minimizing the brand transgression event. there is no evidence of mediation through shame for PBT ( but it is Building on these findings, the current research examines situ- significant for VBT (. ations when the buffering effect of self-brand connection is likely Studies 3 and 4 examine the role of brand response as modera- to occur. We identify two variables that help address this research tors of the impact of affective commitment on unethical behavior; question, one relating to the transgressing brand and one relating to brand remedial measures (vague vs. specific (concrete)) in Study 3 the consumer. Specific to the brand-related variable, research has (n=230) and brand apology (present vs. absent) in Study 4 (n=233). shown different implications such as consumer happiness and sat- In Study 3, the interaction between remedial measures and affective isfaction (Gilovich, Kumar, and Jampol 2015), depending on the commitment is significant. Specifically, we find that high affective type of purchase—whether it is material or experiential. Applying commitment leads to more unethical behavior when the remedial this distinction to brands, we posit that brands can be perceived as measures were vague (vs. specific) () Contrary to the conventional more experiential or material as well, depending on which aspect of wisdom, even with specific remedial measures, high (vs. low) af- the brand is salient. The consumer-related variable we include in our fective commitment leads to more punitive behavior (). Thus, while framework is materialism, or the level of importance a person places specific response is more effective, the backlash effect of high affec- on their possessions (Richins and Dawson 1992; Richins, Mick, and tive commitment is robust and persists. Finally, Study 4 finds that Monroe 2004). We propose that materialism influences the extent of there is a significant interaction between brand apology and affective consumer forgiveness of different type of brands (material vs. expe- commitment While high affective commitment is associated with riential) subsequent to a transgression. Consistent with prior work, more unethical behavior in the no-apology condition ), this effect we propose that the underlying mechanism for such effects is the is non-significant in the apology condition ). As predicted, there is difference in self-brand connection consumers form with material evidence of mediation through shame in the no-apology condition ( and experiential brands. but not in the apology condition (. We tested our hypotheses across six studies using different Overall, our findings make four important contributions. First, brands and varying product categories. In study 1, we examined our we extend our understanding about the effect of brand transgressions proposed interaction effect of brand type and materialism on brand in the context of unethical consumer behavior. Unethical consumer evaluations by manipulating materialism and asking participants to behavior has significant short- and long-term costs for firms and thus, recall a past brand transgression episode of a material versus expe- this is an important question. Further, existing research has focused riential brand. We find that there was a significant interaction effect of on performance-based (vs. value-based) transgressions. We contrib- brand type and materialism on brand evaluation, F(1, 182) = 4.64, p = ute by examining VBT but additionally, contrast its effects with those .03. Specifically, within the high materialism condition, material brands of PBT. Second, we resolve the affective commitment paradox in the (Mmaterial = 6.01, SD = .83) were evaluated higher than experiential brands literature for transgressing brands by identifying some of the key (Mexperiential = 5.30, SD = 1.51). Within the low materialism condition, moderating factors which might explain when commitment might there was no difference in brand evaluations between material brands have a buffering versus an amplifying effect. Third, we examine sev- (Mmaterial = 5.67, SD = .91) and experiential brands (Mexperiential = 5.66, SD = 1.11. In study 2, we replicate study 1 findings by employing a different brand type manipulation; we present a single brand and frame 1 Note: All t-tests are conducted on groups with ±1sd of affective commitment means. 978 / Brand Misbehavior: Integrating Multiple Perspectives the brand to be more material or experiential by highlighting differ- Ahluwalia, Rohini, Robert E. Burnkrant, and H. Rao Unnava ent aspects of the brand. (2000), “Consumer Response to Negative Publicity: The In studies 3 and 4, we demonstrated that the interaction effect Moderating Role of Commitment,” Journal of Marketing of brand type and materialism replicates with our proposed mediator Research, 37 (2), 203–214. of self-brand connection, the extent to which consumers incorporate Allen, Natalie J. and John P. Meyer (1990), “The Measurement brands into their self-concept. Study 3 manipulated participants’ ma- and Antecedents of Affective, Continuance and Normative terialistic mindset and then presented a single brand Nike to be seen Commitment to Organization,” Journal of Occupational as experiential or material. Results revealed a significant interaction be- Psychology, 63 (1), 1–18. tween the two factors, F(1, 173) = 3.87, p = .051. When Nike was viewed Awad, Norah and Nara Youn (2018), “You Reflect Me: Narcissistic as a material brand, participants in the high materialism condition indi- Consumers Prefer Anthropomorphized Arrogant Brands,” cated higher self-brand connection with Nike (M = 47.00, SD = 23.83) Journal of the Association for Consumer Research, 3 (4), versus those in the low materialism condition (M = 35.99, SD = 23.44; 540–554. F(1, 175) = 4.46, p = .036). On the other hand, there was no difference Bauer, Monika A., James E. B. Wilkie, Jung K. Kim, and Galen between high and low materialism conditions when Nike was viewed as V. Bodenhausen (2012), “Cuing Consumerism: Situational an experiential brand (F(1, 175) = 1.94, p = .165. Study 4 conceptually Materialism Undermines Personal and Social Well-Being,” replicates study 3 findings but using a different brand. Psychological Science, 23 (5), 517–523. Study 5 and 6 tested our full theoretical model by investigat- Bechwati, Nada N. and Maureen Morrin (2003), “Outraged ing whether the interaction effect of brand type and materialism on Consumers: Getting Even at the Expense of Getting a Good brand evaluation is mediated by self-brand connection. Specifically, Deal,” Journal of Consumer Psychology, 13 (4), 440–453. study 6 replicated study 5 by using a fictitious brand to control for Cheng, Shirley Y.Y., Tiffany B. White, and Lan N. Chaplin any prior beliefs about and attitudes toward the featured brand. Re- (2012). “The Effects of Self‐Brand Connections on sults revealed a significant moderated mediation (index = 26.65, SE Responses to Brand Failure: A New Look At the Consumer– = 19.23, 95% CI [.13, 76.78] such that the indirect effect of self- BrandRelationship,” Journal of Consumer Psychology, 22 (2), brand connection was significant at one SD above the mean level of 280–288. materialism (a x b = 21.43, 95% CI [1.09, 54.81]) but not one SD DeWitt, Tom and Michael K. Brady (2003), “Rethinking Service below the mean (a x b = -8.22, 95% CI [-37.01, 14.42]). That is, self- Recovery Strategies: The Effect of Rapport on Consumer brand connection was a significant mediator only among high (and Responses to Service Failure”, Journal of Service Research, 6 not low) materialistic individuals. (2), 193–207. Overall, our findings make several theoretical contributions. We Dick, Samantha (2019), “Most Unethical Fashion Brands extend the brand transgression literature in showing that subsequent Exposed,” Retrieved on August 15, 2019 from https:// to a brand committing a violation, consumer materialism influences thenewdaily.com.au/news/national/2019/04/10/ethical-fashion- evaluations of material versus experiential brands by increasing the brands/. self-brand connection consumers form with brands. In doing so, we Dittmar, Helga, Rod Bond, Megan Hurst, and Tim Kasser (2014), add to the body of work that studies how consumer identity influ- “The Relationship between Materialism and Personal Well- ences judgments after a transgression by identifying two novel ante- Being: A Meta-Analysis,” Journal of Personality and Social cedents of material/experiential brands and materialism. Moreover, Psychology, 107 (5), 879–924. we contribute to the purchase type literature in demonstrating that Evanschitzky, Heiner, Gopalkrishnan R. Iyer, Hilke Plassmann, the material versus experiential dichotomy can be applied to percep- Joerg Niessing, and Heribert Meffert (2006), “The Relative tion of brands. This enables us to bring together the purchase type Strength of Affective Commitment in Securing Loyalty in and the brand transgression literatures and proposes a novel dimen- Service Relationships,” Journal of Business Research, 59 (12), sion that brands that can be differentiated on, which is important in 1207–1213. understanding how consumers respond to negative brand informa- Fullerton, Gordon (2005), “How Commitment Both Enables and tion. Finally, we contribute to the materialism literature by identify- Undermines Marketing Relationships,” European Journal of ing a novel consequence of materialistic mindset. While past work Marketing, 39 (11/12), 1372–1388. suggests that a materialistic mindset generally has negative conse- George, Martin (2004), “Heritage Branding Helps in Global quences (Bauer et al. 2012; Dittmar et al. 2014), our work demon- Markets,” Marketing News, 4 (13), 1. strates that high materialism buffers against negative information for Geuens, Maggie, Bert Weijters, and Kristof De Wulf (2009), “A material (vs. experiential) brand. This finding suggests that material- New Measure of Brand Personality,” International Journal of ism does not always result in negative outcomes, particularly in the Research in Marketing, 26 (2), 97–107. context of brand transgressions. Gilovich, Thomas, Amit Kumar, and Lily Jampol (2015), “A Wonderful Life: Experiential Consumption and the Pursuit REFERENCES of Happiness,” Journal of Consumer Psychology, 25 (1), Aaker, David A. (2004), “Leveraging the Corporate Brand,” 152–165. California Management Review, 46 (3), 6–18. Grégoire, Yany and Robert J. Fisher (2006), “The Effects of Aaker, Jennifer L. (1997), “Dimensions of Brand Personality,” Relationship Quality on Customer Retaliation,” Marketing Journal of Marketing Research, 34 (3), 347–356. Letters, 17 (1), 31–46. Aaker, Jennifer L., Susan Fournier, and S. Adam Brasel (2004), Grégoire, Yany and Robert J. Fisher (2008), “Customer Betrayal “When Good Brands Do Bad,” Journal of Consumer and Retaliation: When Your Best Customers Become Your Research, 31 (1), 1–16 Worst Enemies,” Journal of the Academy of Marketing Science, 36 (2), 247–261. Advances in Consumer Research (Volume 48) / 979

Grégoire, Yany, Thomas M. Tripp, and Renaud Legoux (2009), Neumann, Craig S., Robert D. Hare, and Joseph P. Newman (2007), “When Customer Love Turns into Lasting Hate: The Effects “The Super-Ordinate Nature of the Psychopathy Checklist- of Relationship Strength and Time on Customer Revenge and Revised,” Journal of Personality Disorders, 21 (2), 102–117. Avoidance,” Journal of Marketing, 73 (6), 18–32. Nyffenegger, Bettina, Andrea Kähr, Harley Krohmer, and Wayne Gong, Han, and Douglas L. Medin (2012), “Construal Levels D. Hoyer (2018), “How Should Retailers Deal with Consumer and Moral Judgment: Some Complications,” Judgment and Sabotage of a Manufacturer Brand?” Journal of Association of Decision Making, 7 (5), 628–638. Consumer Research, 3 (3), 379–395. Grohmann, Bianca (2009), “Gender Dimensions of Brand Paulhus, Delroy L. and Kevin M. Williams (2002), “The Dark Personality,” Journal of Marketing Research, 46 (1), 105–119. Triad of Personality: Narcissism, Machiavellianism, and Hess Jr., Ronald L., Shankar Ganesan, and Noreen M. Klein (2003), Psychopathy,” Journal of Research in Personality, 36 (6), “Service Failure and Recovery: The Impact of Relationship 556–563. Factors on Customer Satisfaction,” Journal of the Academy of Rauthmann, John F. and Theresa Will (2011), “Proposing a Marketing Science, 31 (2), 127–145. Multidimensional Machiavellianism Conceptualization,” Huber Frank, Kai Vollhardt, Isabel Matthes, and Johannes Vogel Social Behavior and Personality: An International Journal, 39 (2010), “Brand Misconduct: Consequences on Consumer– (3), 391–403. Brand Relationships,” Journal of Business Research, 63 (11), Richins, Marsha L. and Scott Dawson (1992), “A Consumer Values 1113–1120. Orientation for Materialism and Its Measurement: Scale Jones, Tim, Gavin L. Fox, Shirley F. Taylor, and Leandre R. Development and Validation,” Journal of Consumer Research, Fabrigar (2010), “Service Customer Commitment and 19 (3), 303–316. Response,” Journal of Services Marketing, 24 (1), 16–28. Richins, Marsha L., David G. Mick, and Kent B. Monroe (2004), Kähr, Andrea, Bettina Nyffenegger, Harley Krohmer, and Wayne “The Material Values Scale: Measurement Properties and D. Hoyer (2016), “When Hostile Consumers Wreak Havoc on Development of a Short Form,” Journal of Consumer Your Brand: The Phenomenon of Consumer Brand Sabotage,” Research, 31 (1), 209–219. Journal of Marketing, 80 (3), 25–41. Rojas-Méndez, Jose I. and Nicolas Papadopoulos (2012), Khamitov, Mansur, Yany Grégoire, and Anshu Suri (2020), “A “Argentine Consumers’ Perceptions of the US Brand Systematic Review of Brand Transgression, Service Failure Personality,” Latin American Business Review, 13 (4), Recovery and Product-Harm Crisis: Integration and Guiding 329–345. Insights,” Journal of the Academy of Marketing Science, 48 Rotman, Jeff D., Mansur Khamitov, and Scott Connors (2018), (3), 519–542. “Lie, Cheat, and Steal: How Harmful Brands Motivate Kim, Jae Wook, Jiho Choi, William Qualls, and Kyesook Han Consumers to Act Unethically,” Journal of Consumer (2008), “It Takes a Marketplace Community to Raise Brand Psychology, 28 (2), 353–361. Commitment: The Role of Online Communities,” Journal of Swaminathan, Vanitha, Karen L. Page, and Zeynep Gürhan-Canli Marketing Management, 24 (3-4), 409–431. (2007), “‘My’ Brand or ‘Our’ Brand: The Effects of Brand Klein, Jill Gabrielle, N. Craig Smith, and Andrew John (2004), Relationship Dimensions and Self-Construal on Brand “Why We Boycott: Consumer Motivations for Boycott Evaluations,” Journal of Consumer Research, 34 (2), 248–259. Participation,” Journal of Marketing, 68 (3), 92–109. Swaminathan, Vanitha, Alina Sorescu, Jan-Benedict Steenkamp, Kyung, Ellie J., Geeta Menon, and Yaacov Trope (2010), Thomas Clayton Gibson O’Guinn, and Bernd Schmitt (2020), “Reconstruction of Things Past: Why Do Some Memories Feel “Branding in a Hyperconnected World: Refocusing Theories so Close and Others So Far Away?” Journal of Experimental and Rethinking Boundaries,” Journal of Marketing, 84 (2) Social Psychology, 46 (1), 217–220. 24–46. Liberman, Nira, Michael D. Sagristano, and Yaacov Trope (2002), Trope, Yaacov and Nira Liberman (2003), “Temporal Construal,” “The Effect of Temporal Distance on Level of Mental Psychological Review, 110 (3), 403–421. Construal,” Journal of Experimental Social Psychology, 38 Trope, Yaacov and Nira Liberman (2010), “Construal-Level Theory (6), 523–534. of Psychological Distance,” Psychological Review, 117 (2), MacInnis, Deborah J. and Valerie S. Folkes (2017), “Humanizing 440–463. Brands: When Brands Seem to Be Like Me, Part of Me, and in Turri, Anna M., Karen H. Smith, and Elyria Kemp (2013), a Relationship with Me,” Journal of Consumer Psychology, 27 “Developing Affective Brand Commitment through Social (3), 355–374. Media,” Journal of Electronic Commerce Research, 14 (3), Mazar, Nina, On Amir, and Dan Ariely (2008), “The Dishonesty 201–214. of Honest People: A Theory of Self-Concept Maintenance,” Urde, Mats, Stephen A. Greyser, and John M. T. Balmer (2007), Journal of Marketing Research, 45 (6), 633–644. “Corporate Brands with a Heritage,” Journal of Brand McGraw, A. Peter and Caleb Warren (2010), “Benign Violations: Management, 15 (1), 4–19. Making Immoral Behavior Funny,” Psychological Science, 21 Vallacher, Robin R. and Daniel M. Wegner (1989), “Levels (8), 1141–1149. of Personal Agency: Individual Variation in Action McGraw, A. Peter, Caleb Warren, Lawrence E. Williams, and Identification,”Journal of Personality and Social Psychology, Bridget Leonard (2012), “Too Close for Comfort, or Too 57 (4), 660–671. Far to Care? Finding Humor in Distant Tragedies and Close Van Boven, Leaf and Thomas Gilovich (2003), “To Do or to Have? Mishaps,” Psychological Science, 23 (10), 1215–1223. That is the Question,” Journal of Personality and Social Morf, Carolyn C. and Frederick Rhodewalt (2001), “Unraveling Psychology, 85 (6), 1193–1202. the Paradoxes of Narcissism: A Dynamic Self-Regulatory Wangenheim, Florian V. (2005), “Postswitching Negative Word of Processing Model,” Psychological Inquiry, 12 (4), 177–196. Mouth,” Journal of Service Research, 8 (1), 67–78. 980 / Brand Misbehavior: Integrating Multiple Perspectives

Warren, Caleb and Margaret C. Campbell (2014), “What Makes Winterich, Karen P., Vikas Mittal, and Andrea C. Morales (2014), Things Cool? How Autonomy Influences Perceived Coolness,” “Protect Thyself: How Affective Self-Protection Increases Journal of Consumer Research, 41 (2), 543–563. Self-Interested, Unethical Behavior,” Organizational Behavior and Human Decision Processes, 125 (2), 151–161. New Approaches in Text Analytics Chairs: Matthew D. Rocklage, College of Management, University of Massachusetts, USA Jonah Berger, The Wharton School, University of Pennsylvania, USA

Paper #1: Dynamically Solving the Self-Presenter’s Paradox: critics and lay consumers across 3 million Rotten Tomatoes reviews. When Customer Care Should be Warm vs . Competent Experiments show this numbness can be explained by cognitive Yang Li, Cheung Kong Graduate School of Business, China frameworks experts develop and then apply to the domain. Grant Packard, Schulich School of Business, York University, Hovy, Melumad, and Inman introduce a text analytics tool Canada called Wordify, which uses machine learning to identify words that Jonah Berger, The Wharton School, University of best discriminate between groups of interest. As one way of dem- Pennsylvania, USA onstrating its unique value, the authors analyze 60,000 TripAdvisor reviews that were written either on a PC or smartphone. Using their Paper #2: The Numbness of Expertise: Emotional Trajectories bottom-up approach, they find that reviewers on smartphones not in the Development of Expertise only use more emotional language than those using PCs, but emo- Matthew D. Rocklage, College of Management, University of tional language that is also more intense. Massachusetts, USA use a combination of machine Derek D. Rucker, Kellogg School of Management, Junqué de Fortuny and Lee learning, network analysis, and text analytics to identify and track Northwestern University, USA networks of influencers. Their bottom-up approach allows them to Loran F. Nordgren, Kellogg School of Management, understand how brand reference groups are dynamically created and Northwestern University, USA how the new age of online celebrities imbue brands with cultural Paper #3: Wordify: A Tool for Discovering Consumer meaning. For example, they identify a “moms” social influencer Vocabularies network that appears to be driven primarily by brand sponsorship, Dirk Hovy, Bocconi University, Italy thereby highlighting firms’ ability to successfully influence brand Shiri Melumad, The Wharton School, University of meanings and reference groups. Pennsylvania, USA With a special focus on text analytics, these papers answer the J. Jeffrey Inman, University of Pittsburgh, USA call to introduce novel methodologies and perspectives. This session Paper #4: Influencer-Generated Reference Groups should have wide appeal to researchers interested in text analysis, Enric Junqué de Fortuny, NYU Shanghai, China language, and real-world big data as well as those studying emotion, Jeffrey Lee, Kogod School of Business, American University, brands, word-of-mouth, and consumer networks. We invite others to USA join us in peering beneath the calm surface of text analytics in order to spark ideas for collaboration and gain unique consumer insights. SESSION OVERVIEW Text analytics is becoming increasingly popular. From consum- Dynamically Solving the Self-Presenter’s Paradox: When er reviews to social media posts, marketers are realizing the value Customer Care Should be Warm vs . Competent of language and beginning to collect and analyze massive amounts of text. EXTENDED ABSTRACT Despite this, many researchers use only the most basic ap- Consumer conversations are dynamic. People chat with each proaches, pairing simple word counts with t-tests or regression. What other online (discussion boards, texting, social media) and debate unique approaches are researchers developing? And how might we word of mouth opinions in the “real world.” Salespeople try to per- combine existing text analytics with novel methods to cultivate fresh suade potential buyers who sometimes push back, while call cen- insights? tre workers talk with customers to help resolve their issues. These This session draws on researchers from marketing to computer important consumer interactions are not monologues, but dynamic science to answer important questions using new approaches. How dialogues between people. do we model the flow of conversation and its consequences? How While conversations are a central feature of consumer life, they might we trace consumer trajectories over time as they gain exper- can be remarkably difficult to analyze. They entail a messy series of tise? What words best reveal distinct groups of consumers and their conversational turns with dramatic variation in content and impor- vocabularies? Can we use text analytics to capture networks of on- tance. These challenges may be why most prior language research line influencers and competing brands? examines texts or speech acts as singular, static events (e.g., Kronrod Li, Packard, and Berger show the power of functional data et al. 2011; Packard, Moore, and McFerran 2018). analysis and machine learning for parsing and understanding con- But a more granular view may be important. Clearly some parts versational dynamics. Across 200 real-world customer service calls, of conversations may matter more than others, but which parts might they find there are pivotal moments within a conversation where em- those be, and how can researchers identify them? We introduce a ployees must emphasize warmth or competence. Increasing emotion novel method allowing researchers to not only examine whether lan- at both the beginning and ending of a conversation predicts positive guage matters, but when it matters. Specifically, when different lin- customer perceptions, whereas cognitive language should be empha- guistics features in different parts of conversations play a larger role. sized in the middle. To demonstrate this approach, we explore a test case around Rocklage, Rucker, and Nordgren demonstrate how growth what are said to be the two most important dimensions of person curve modeling can be used to trace emotional trajectories as con- perception—warmth and competence (Abele and Wojciszke 2007; sumers gain expertise. Across 70,000 individuals, for each additional Fiske, Cuddy and Glick 2007). It’s difficult to seem both warm and wine or beer they taste, consumers’ language becomes more emo- competent. Trying to be more emotionally-concerned impedes per- tionally numb. These results replicate comparing professional movie ceptions of competence, while acting in a more rational, cognitively-

Advances in Consumer Research 981 Volume 48, ©2020 982 / New Approaches in Text Analytics oriented manner makes people seem less warm (the warmth / com- Discussion. This research begins to shed light on a richer theory petence trade off or “self-presenter’s paradox”; Godfrey, Jones and of conversational dynamics. While a great deal of work has looked Lord 1986; Holoien and Fiske 2013; Wang et al. 2019). As a result, at word of mouth in general, when different linguistic approaches are prior work suggests people should prioritize just one of these two most useful in conversation has received little attention. modes in a given social interaction. Our results show that customer service employees should try to However, speaking only warmly or competently may not be be warm and competent, but at different times in the conversation. ideal. For example, service agents are supposed to show they care, Further, this is just one example of how this approach can be used but also that they are thinking through and trying to solve the cus- to understand conversational dynamics. The same approach could tomer’s problems (e.g., de Ruyter and Wetzels 2000; Parasuraman et be used to understand other linguistic features thought to be benefi- al. 1985; Spiro and Weitz 1990). cial such as concreteness, asking questions, or using long sentences We suggest that when agents speak warmly vs. competently is (Castleberry et al. 1999; de Ruyter and Wetzels 2000; Schellekens et important. Specifically, more affective language at the beginning of al. 2010) in customer service, or studying word of mouth, negotia- an interaction may be important to build rapport over the customer’s tion, message recall, or other important marketing subjects. issue. That said, it is also important to turn to a more analytic, cognitive style when trying to competently solve that issue. Finally, given The Numbness of Expertise: Emotional Trajectories in the work on recency or end effects (Greene 1986), closing with af- the Development of Expertise fective language may be key to leaving the customer feeling positive about the interaction. Our approach uses dynamic quantitative mod- EXTENDED ABSTRACT eling to test this possibility. Expertise provides numerous benefits. Experts process infor- Data. We obtained 200 audio recordings of customer service mation more efficiently (Johnson and Mervis 1997), remember in- calls from a large American retailer. As a dependent measure, the formation better (LaTour and LaTour 2010), and often make better firm provided end-of-call perceptions of employee helpfulness (1 = decisions (Mitchell and Dacin 1996). not at all helpful, 4 = very helpful). A transcription company con- Yet, expertise may also carry a cost. Specifically, we propose verted the recordings to text. Each conversational turn was treated that as consumers accrue expertise in a hedonic domain, they be- as a separate record (e.g., turn 1 (agent): “How can I help you?”, come increasingly emotionally numb to their experiences. turn 2 (customer): “I can’t find (…)”), resulting in 12,410 turns for Building expertise leads to cognitive structure – i.e., knowl- analysis. The time-series of turns is standardized over all calls (time edge – for how to understand a domain (Alba and Hutchinson = 0-1). 1987). Though this knowledge is often beneficial, it may also lead For our independent measures, we captured affective versus consumers to engage in an analytical dissection of their experience cognitive language using validated linguistic dictionaries (affective and thereby detract from it. Take, for example, the movie critic who processes, cognitive processes; LIWC; Pennebaker et al. 2015) for uses their knowledge to analyze and decompose a comedy, thereby both the employee and customer. Controls included customer attri- undermining the experience. Given that people engage in consump- butes (e.g., gender, lifetime expenditures), employee attributes (e.g., tion for the very feelings it evokes (Holbrook and Hirschman 1982), tenure, quality ratings), customer language, and conversation fea- expertise may have the ironic effect of decreasing feelings for those tures (e.g., issue, severity, linguistic synchronicity). consumers most interested in a domain. Method. We use methods from functional data analysis (FDA; Study 1 provided initial evidence in a naturalistic setting. Ramsay and Silverman 2007) and machine learning (Yang and Zou We obtained all reviews from the movie website RottenTomatoes.

2015; Kong et al. 2016) to address the challenges of analyzing con- com, which includes reviews from both professional critics (ncritics = versational language. This integrated approach enables data-driven 5,780) and lay consumers (nconsumers = 642,681) across 8,627 movies identification of time-based functions (curves) of conversational fea- (nreviews = 3,009,095‬). We quantified the positivity and emotional- tures (e.g., language) significantly linked to the outcome of interest ity of reviewers’ language using the Evaluative Lexicon (EL; Rock- (e.g., perception of the other speaker). For any set of conversation lage, Rucker, and Nordgren 2018; Rocklage and Fazio 2020). Using feature(s), this method produces sensitivity curves that can deviate mixed modeling and controlling for how positive or negative a re- over time either positively or negatively from non-significance (zero view was (its extremity), critics expressed less feeling than consum- line) in relation to the outcome of interest. ers. This result held controlling for review length, movie genre, and Results. Results support our theorizing. Areas shaded in blue the normative emotion evoked by the movie. For this and all studies, reveal the importance of more cognitive language in the middle 50% results held across positive and negative emotion. of the call. That said, as shown by the negative effect of cognitive Studies 2 and 3 used advances in statistical modeling to trace language at the beginning of the call, employees who try to jump the accumulation of expertise across time via growth curve mod- straight into using cognitive language at the call’s opening-- perhaps eling (see Bolger and Laurenceau 2013). Using CellarTracker.com attempting to quickly and competently solve the customer’s prob- and BeerAdvocate.com reviews (McAuley and Leskovec 2013), we lem-- generate significantly negative social perceptions (blue shaded followed 71,610 consumers as they developed expertise in wine and area below the zero line). Instead, affective language has the most beer, respectively (nreviews = 3,175,143). Controlling for review ex- positive effect at the beginning and end of the call. tremity, consumers showed consistent decreases in feeling toward The average employee’s use of both affective and cognitive each additional wine or beer they reviewed. These results held con- language does not follow the optimal functional forms. Instead, the trolling for review length, the total reviews that user wrote, gender, agent’s use of affective language is at its lowest point at the start of age, and the normative emotion evoked by each item. They also held the call, when it is particularly important, while cognitive language for reviewers’ most emotional reaction in each review, indicating that is near its lowest point between 12.5% and 40% into the conversa- experts did not reach the same emotional heights as when they first tion, which our method reveals is when it offers its most positive began. impact on social perceptions. Study 4 examined this phenomenon in the lab and generalized it to professional photography. We showed participants (n = 102) 10 Advances in Consumer Research (Volume 48) / 983 randomly-selected photographs (5 positive and 5 negative; Kurdi, search has focused nearly exclusively on the benefits of expertise, we Lozano, and Banaji 2017). After each photograph, participants in- show that expertise can also have negative side effects: the pursuit dicated the emotionality and extremity of their reaction using an EL of expertise can detract from the very experiences experts chased in adjective checklist (Rocklage and Fazio 2015, 2016)the adjectives – the first place. individuals use can be harnessed to understand the different aspects of their attitudes. The present research introduces a novel approach Wordify: A Tool for Discovering Consumer Vocabularies to measuring attitudes that allows researchers to quantify these aspects. In Study 1, we created a list of 94 evaluative adjectives and EXTENDED ABSTRACT asked participant judges to rate the implied valence, extremity, and There is currently growing interest in how consumers use words emotionality of each adjective. This approach allowed us to quantify to express thoughts, feelings, and preferences. In this work we de- each adjective along these dimensions and thereby create the Evalu- scribe a new, easy-to-use online tool for automatic vocabulary dis- ative Lexicon (EL. They then reported their expertise in photography covery named Wordify that identifies the words most associated with (Carlson et al. 2009). This study and all the following use mixed different a priori categorizations of text. Unlike other text-analysis modeling and control for extremity. The more expertise consumers tools that may be familiar to consumer researchers (such as LIWC or had, the less feeling they had in reaction to each photograph. LDA), Wordify does not seek to provide a summary measure of senti- Study 5 tested the proposed mechanism: experts develop a cog- ment or a characterization of latent topics, but rather a list of words nitive framework that, when applied, leads to decreases in feeling. that are most predictive of different classifications that are ofa priori It also generalized the effect using a different emotion measure. We interest to researchers. The tool has multiple domains of potential used the same design as Study 4, but participants (n = 452) were application, including: randomly assigned to one of three conditions. Those in the Expertise 1. Exploratory analysis: Discovering the vocabularies that condition went through a learning module on the characteristics that consumers use when discussing a given topic (e.g., prod- experts use to understand the quality of a photograph (e.g., the mes- ucts, service experiences); sage it is attempting to convey, lighting, etc; Sethna 1992) and asked 2. Hypothesis testing: A means of testing hypotheses about to apply that learning. As a control condition, those in the Expertise- differences in word use across contexts (e.g., are the words Without-Learning condition were asked to engage with the photo- most commonly used in reviews written on phones more graphs as if they were an expert photographer, but not provided any emotional than those most commonly used in PC-generat- learning. Another control condition was not provided the framework ed reviews?); nor any specific instructions. After each photograph, individuals 3. Augmentation: When used in conjunction with tools such indicated the intensity of their emotions using the Self-Assessment as LIWC and LDA, Wordify can yield deeper insights Manikin (Bradley and Lang 1994)arousal, and dominance associ- into the words that drive classifications produced by such ated with a person’s affective reaction to a wide variety of stimuli. In methods; and this experiment, we compare reports of affective experience obtained 4. Dictionary expansion: The tool can be used to expand the using SAM, which requires only three simple judgments, to the Se- dictionaries of bag-of-words text-analysis tools (such as mantic Differential scale devised by Mehrabian and Russell (An ap- LIWC) so that they provide more accurate measure of sen- proach to environmental psychology, 1974. Those in the Expertise timent. condition experienced less feeling after each photograph versus con- One of the main advantages of Wordify is its usability. The tool trol. Expertise-Without-Learning participants showed no difference is freely available online (https://wordify.unibocconi.it/index; tem- from the Control or Expertise conditions, indicating that the cogni- porary user password 0981), and requires only the upload of a .csv tive framework was key for emotional numbness. file that includes text and a dichotomous classification variable of Study 6 was preregistered (https://bit.ly/32cZMvh) and ex- interest. The resulting output is a list of words that are most predic- amined whether mere knowledge of the cognitive framework was tive of the classification, along with a measure of their degree of enough to decrease emotion or whether individuals must actively ap- predictive validity. In this way, the tool can be seen as analogous to ply it. Participants (n = 605) were assigned to one of three conditions. discriminant analysis, but applied to a setting where the predictors Those in the Mere Knowledge condition went through the learning are potentially tens of thousands of words in a corpus. module from Study 5, but not instructed to apply their learning. In- The statistical basis of Wordify is randomized logistic regres- stead, they were told we simply wanted to understand how clear the sion (RLR, Meinshausen and Bühlmann 2010). Applied here RLR module was for future research. Those in the Application condition is a method of machine learning that repeatedly fits logistic models were instructed to apply their learning. Control participants were to random samples of words in a corpus, and then applies shrinkage not given the framework nor any specific instructions. Participants or regularization methods to correct for overfitting of these models. used EL checklists to indicate their reactions. As preregistered, those The algorithm then computes the relative frequency with which each in the Application condition felt less emotion toward each photo- word in the corpus appears as a significant predictor in each of these graph compared to Control. Those in the Mere Knowledge condition models. The final output is the set of words that are most often able to showed no difference from Control. These results indicate consum- discriminate between categories of interest. By relying on repeated ers must apply the framework and that mere knowledge is not suf- sampling and randomized regularization, the method approximates ficient. an “ideal” world where all possible models are estimated, and the Across multiple domains, as consumers develop expertise, they model with the best fit is selected (something that is computationally become increasingly emotionally numb. This was evident using both infeasible). traditional experimental approaches and growth curve modeling, To illustrate the use and output of Wordify, consider a recent which allowed us to track consumers’ emotional trajectories as they work by Melumad, Inman, and Pham (2019), who hypothesized that gained expertise. Though this effect could be the result of multiple consumer reviews written on smartphones will often be more emo- processes (e.g., hedonic satiation), we show that it can result from tional than that written on personal computers (PCs). The hypothesis the cognitive structure experts develop and apply. Whereas past reis that, because smartphones have a smaller keyboard and screen, 984 / New Approaches in Text Analytics content written on smartphones will tend to be shorter, leading users marketing research (e.g., McCracken 1989), consumer influencers to convey the gist or essential elements of an experience (in lieu of in digital environments can vary dramatically in terms of their au- specific details). Since the gist of one’s experience is usually based dience, and even consumers with less than 5,000 followers on soon the writer’s emotional evaluation of that experience, this results cial media (e.g., “micro-influencers”) can now be tapped to support in the creation of content that is more selectively emotional when brands. Additionally, these consumer influencers have ultimate con- written on smartphones versus PCs. trol in the publishing of content where they promote brands (Hughes, To test this, we applied Wordify to the same corpus of 61,642 Swaminathan and Brooks 2019). Finally, these consumer influenc- TripAdvisor reviews analyzed by Melumad et al. (2019). We esti- ers not only provide influence via sponsored engagements, but also mated two sets of Wordify models: one identifying the words most through unsolicited opinions on brands and products (i.e. via earned predictive of a review being written on a smartphone, and another of media). those most predictive of a review being written on a PC. The smart- Notably, because of the emergence of the consumer influencer, phone results yielded a vocabulary of 269 words that were signifi- the potential for brand reference group associations has never been cant predictors in more than 30% of RLR models, and the PC results greater. As multiple, similar consumer influencers mention a given yielded a vocabulary of 389 words (reflective of the greater num- brand, followers may repeatedly associate this brand with a certain ber of PC-generated reviews). The partial output from the Wordify “type” of person. Decades of research (e.g., Bearden and Etzel 1982; analysis is reported below. Escalas and Bettman 2005; Berger and Heath 2007; White, Argo, The results illustrate that, indeed, consumers use different vo- and Sengupta, 2012) have explored how reference groups can in- cabularies when writing reviews on their smartphones compared to fluence consumer behavior for brands or products with pre-existing PCs. As hypothesized, reviews written on smartphones tended to group associations. At the same time, recent commentary suggests contain a greater proportion of emotional words. Of the 269 words that our understanding of how cultural meanings are transferred to most predictive of a smartphone-generated review, 18.22% ex- brands (e.g., how brands gain reference group associations) is still pressed emotionality or an evaluation, while of the 389 words most underdeveloped (Batra 2019), particularly from a quantitative, em- predictive of a PC review, only 6.43% were emotional or evaluative pirical perspective. Because consumers now have digital “mega- (Wald Chi-Square = 20.44, p < .001). phones” (McQuarrie, Miller, and Phillips 2013) and are more em- Perhaps more interestingly, the results reveal that the emotional powered than ever to shape brand meanings, we chose to explore (1) words used in smartphone-generated reviews tended to be more in- how influencers might transfer reference group meanings to brands tense in character than the emotional words used in PC-generated at scale on social media, along with (2) the role of brand sponsorship reviews. For example, of the emotional words used, the three that and endorsement in instigating this meaning creation. were most predictive of a smartphone-generated review were “amaz- We analyzed a dataset of over 377,404 social media posts by ing”, “awesome”, and “delicious” (appearing in 59%, 56%, and 53% 745 US-based consumer influencers on Instagram, generated through of all models, respectively), while the three emotional words most a major social media influencer search engine. First, we identified predictive of a PC-generated review were “excellent”, “wonderful”, potential reference group categories by analyzing the text of all influ- and “fine” (appearing in 54%, 46%, and 40% of all models, respec- encers’ Instagram profile bios, which provide self-description of key tively). roles and identities (e.g., “fitness enthusiast”, “dog mom”). We con- While the words used in PC-generated reviews may have con- solidated the most frequent self-descriptions and used crowdsourced veyed less intense emotionality, more generally, such reviews made tagging (via an mTurk survey of 317 participants) to identify ad- greater use of words that were descriptive of the details surrounding ditional reference group categorizations for each influencer. We also the experience itself. For example, the three words that were overall processed users’ bio photos through Microsoft Azure’s deep learning most predictive of a PC-generated review were “restaurant”, “din- FACE API to generate demo-faceo descriptives for the influencers ing”, and “time” (significant in 58%, 57%, and 57% of models, rein our dataset. Finally, using a series of nested (hierarchical) logistic spectively). In contrast, the three words overall most predictive of a regressions alongside a bi-partite network graph (containing 24,518 smartphone were “recommend”, “amazing”, and “awesome” (sig- nodes and 94,807 edges), which we reduced via a Singular Value nificant in 59%, 59%, and 56% of models, respectively). Decomposition exercise (SVD), we confirmed that several of our As shown in the above illustration, Wordify fills a gap in the identified reference groups were uniquely different in terms of their growing array of text-analysis tools available to consumer research- social media tagging and mentioning behavior. ers interested in the study of language. Here, the focus is on the in- We then analyzed the brands that were uniquely associated with dividual words that consumers use in different contexts rather than specific, discernable reference groups in our dataset. As an example, overall patterns such as sentiment (e.g., LIWC) or topics (e.g., LDA). we considered the reference group category of “moms” mentioning As interest in language grows in the consumer research community, the apparel brands in our network graph. We developed a weighted there will be a growing need for similar turnkey tools that can be bi-partite graph to provide a visual representation of the connection used by researchers without advanced training in natural language between apparel brands and our mom reference group. This graph processing. Wordify exemplifies such a contribution. was weighted based on the following size of each influencer, such that macro-influencers exerted greater “pull” than micro-influencers Influencer-Generated Reference Groups on the positioning of each brand in the graph. A Fisher exact test with Bonferroni correction revealed five brands (Gap, Gap Kids, Express, EXTENDED ABSTRACT Old Navy, and J.Crew) significantly associated with the mom refer- In a new “hyperconnected world” where consumers help co- ence group. To explore the underlying mechanism of this associa- create brand meaning on digital platforms (e.g., Swaminathan et al. tion, we applied keyword detection (Anderson and Simester 2013) to 2020), an economy of consumer influencers has developed, where identify call-to-action and sponsorship signals (CTAs) which signal influencers with substantial followings are solicited by brands for intent to advertise. Adapted mediation analyses using a Poisson gen- mentions on social media (Khamis, Ang, and Welling 2017; Mar- eralized linear model suggested that these unique brand-reference wick 2015; Senft 2013). Differing from the celebrities in previous group associations appear to be driven by brand sponsorship and Advances in Consumer Research (Volume 48) / 985 endorsements of these influencers. At the same time, evidence of Fiske, Susan T., Amy J. C. Cuddy, and Peter Glick (2007), unpaid chatter (e.g., for Gap Kids by mom influencers) was also evi- “Universal Dimensions of Social Cognition: Warmth and dent based on our mediation analyses. Competence,” Trends in Cognitive Sciences, 11(2), 77-83. Our work builds on major ideas in consumer research on brand- Godfrey, Debra K, Edward E. Jones, Charles G. Lord (1986), “Self- ing and social groups (e.g., Bearden and Etzel 1982; McCracken promotion is not Ingratiating,” Journal of Personality and 1986; Solomon 1983), particularly on how brands gain cultural Social Psychology, 50(1), 106-115. meanings from celebrities (e.g., McCracken 1989) and other cultural Greene, Robert L. (1986), “Sources of recency effects in free artifacts. We extend this research to a new era of social media, in recall,” Psychological Bulletin, 99(2), 221-228. order to quantitatively explore the emerging relationship between Halmari, Helena (1993), “Intercultural Business Telephone brands and consumer influencers. Leveraging tools from Natural Conversations: A Case of Finns vs. Anglo-Americans,” Language Processing, network analysis and computer vision, we Applied Linguistics, 14(4), 408-430. find that brands are connecting to distinct reference groups in so- Holbrook, Morris B. and Elizabeth C. Hirschman (1982), “The cial media, and that these brand-reference group associations can be Experiential Aspects of Consumption: Consumer Fantasies, driven not only by organic brand chatter (e.g., brand mentions in Feelings, and Fun,” Journal of Consumer Research, 9 (2), everyday conversation), by also by brand sponsorship and endorse- 132–40. ment as part of a reference group branding strategy. Through this Holoien, Deborah S and Susan T. Fiske (2013), “Downplaying work, we hope to help build a bridge between longstanding cultural Positive Impressions: Compensation Between Warmth and sociological theories in consumer research and novel computer and Competence in Impression Management,” Journal of science methods, to provide new insights into the emerging digital Experimental Social Psychology, 49(1), 33-41. era of influencer-driven marketing. Hughes, C., Swaminathan, V., and Brooks, G. (2019), “Driving Brand Engagement Through Online Social Influencers: An REFERENCES Empirical Investigation of Sponsored Blogging Campaigns,” Abele, Andrea E. and Bogdan Wojciszke (2007), “Agency and Journal of Marketing, 83(5), 78–96. Communion from the Perspective of Self Versus Other,” Johnson, Kathy E. and Carolyn B. Mervis (1997), “Effects Journal of Personality and Social Psychology, 93(5), 751-763. of varying levels of expertise on the basic level of Alba, Joseph W. and J. Wesley Hutchinson (1987), “Dimensions of categorization,” Journal of Experimental Psychology: Consumer Expertise,” Journal of Consumer Research, 13 (4), General, 126 (3), 248–77. 411–54. Khamis, S, Ang, L. and Welling, R. (2017), “Self-Branding, Anderson, E. T., and Simester, Duncan I. (2014), “Reviews without ‘Micro-Celebrity’ and the Rise of Social Media Influencers,” a purchase: Low ratings, loyal customers, and deception,” Celebrity Studies, 8(2), 191–208. Journal of Marketing Research, 51(3), 249-269. Kong, Dehan, Kaijie Xue, Fang Yao, and Hao H. Zhang (2016), Batra, Rajeev (2019), “Creating Brand Meaning: A Review and “Partially Functional Linear Regression in High Dimensions,” Research Agenda. Journal of Consumer Psychology, 29(3), Biometrika, 103(1), 147-159. 535–46. Kronrod, Ann, Amir Grinstein, and Luc Wathieu (2011), “Enjoy! Bearden, William O. and Etzel, Michael J. (1982), “Reference Hedonic Consumption and Compliance with Assertive Group Influence on Product and Brand Purchase Decisions,” Messages,” Journal of Consumer Research, 39(1), 51-61. Journal of Consumer Research, 9(2), 183–94. Kurdi, Benedek, Shayn Lozano, and Mahzarin R. Banaji (2017), Berger, Jonah and Heath, Chip (2007), “Where Consumers Diverge “Introducing the Open Affective Standardized Image Set from Others: Identity Signaling and Product Domains,” (OASIS),” Behavior Research Methods, 49 (2), 457–70. Journal of Consumer Research, 34(2), 121–134. Latour, Kathryn A. and Michael S. Latour (2010), “Bridging Bolger, Niall and Jean-Philippe Laurenceau (2013), Intensive Aficionados’ Perceptual and Conceptual Knowledge to Longitudinal Methods: An Introduction to Diary and Enhance How They Learn from Experience,” Journal of Experience Sampling Research, New York, NY: The Guilford Consumer Research, 37 (4), 688–97. Press. Marwick, A. E. (2015), ““Instafame: Luxury Selfies in the Attention Bradley, Margaret M. and Peter J. Lang (1994), “Measuring Economy,” Public Culture, 27(1 (75)), 137–60. emotion: The self-assessment manikin and the semantic McAuley, Julian John and Jure Leskovec (2013), “From Amateurs differential,”Journal of Behavior Therapy and Experimental to Connoisseurs: Modeling the Evolution of User Expertise Psychiatry, 25 (1), 49–59. Through Online Reviews,” in Proceedings of the 22Nd Carlson, Jay P., Leslie H. Vincent, David M. Hardesty, and William International Conference on World Wide Web, WWW ’13, O. Bearden (2009), “Objective and Subjective Knowledge New York, NY, USA: ACM, 897–908. Relationships: A Quantitative Analysis of Consumer Research McCracken, G. (1986), “Culture and Consumption: A Theoretical Findings,” Journal of Consumer Research, 35 (5), 864–76. Account of the Structure and Movement of the Cultural Castleberry, Stephen B., C. David Shepherd, and Rick Ridnour Meaning of Consumer Goods,” Journal of Consumer (1999), “Effective Interpersonal Listening in the Personal Research, 13(1), 71. Selling Environment: Conceptualization, Measurement, and McCracken, Grant (1989), Who Is the Celebrity Endorser? Nomological Validity,” Journal of Marketing Theory and Cultural Foundations of the Endorsement Process,” Journal of Practice, 7(1), 30-38. Consumer Research, 16(3), 310. de Ruyter, Ko and Martin G. M. Wetzels (2000), “The Impact of McQuarrie, E. F., Miller, J. and Phillips, B. J. (2013), The Perceived Listening in Voice-to-Voice Service Encounters,” Megaphone Effect: Taste and Audience in Fashion Blogging,” Journal of Service Research, 2(3), 276-284. Journal of Consumer Research, 40(1), 136–58. 986 / New Approaches in Text Analytics

Meinshausen, Nicolai and Peter Bühlmann (2010), “Stability Rocklage, Matthew D., Derek D. Rucker, and Loran F. Nordgren Selection,” Journal of the Royal Statistical Society: Series B (2018), “The Evaluative Lexicon 2.0: The measurement of (Statistical Methodology), 72 (4), 417–73. emotionality, extremity, and valence in language,” Behavior Melumad, Shiri, J. Jeffrey Inman, and Michel Tuan Pham (2019), Research Methods, 50 (4), 1327–44. “Selectively Emotional: How Smartphone Use Changes User- Schellekens, Gaby A. C., Peeter Verlegh, and Ale Smidts (2010), Generated Content,” Journal of Marketing Research, 56 (4), “Language Abstraction in Word of Mouth,” Journal of 259–75. Consumer Research, 37(2), 207-23. Mitchell, Andrew A. and Peter A. Dacin (1996), “The Assessment Senft, T. M. (2013), “Microcelebrity and the Branded Self, in A of Alternative Measures of Consumer Expertise,” Journal of Companion to New Media Dynamics, ed. John Hartley, Jean Consumer Research, 23 (3), 219–39. Burgess, and Axel Bruns, Oxford, UK: Wiley-Blackwell, Packard, Grant, Sarah G. Moore, and Brent McFerran (2018), 346–54. “(I’m) Happy to Help (You): The Impact of Personal Pronouns Sethna, Edward (1992), “Analysis of Judging,” Royal Photographic Use in Customer-Firm Interactions,” Journal of Marketing Society Journal, 132, 424–26. Research, 55(4), 541-555. Solomon, Michael R. (1983), “The Role of Products as Social Parasuraman, A., Valarie A. Zeithaml, and Leonard L. Berry (1985), Stimuli: A Symbolic Interactionism Perspective,” Journal of “A Conceptual Model of Service Quality and Its Implications Consumer Research, 10(3), 319–329. for Future Research,” Journal of Marketing, 49(4), 41-50. Spiro, Rosann L. and Barton A. Weitz (1990), “Adaptive Selling: Pennebaker, James W., Ryan L. Boyd, Kayla Jordan, and Kate Conceptualization, Measurement, and Nomological Validity,” Blackburn (2015), “The development and psychometric Journal of Marketing Research, 27(1), 61-69. properties of LIWC2015,” Retrieved from https://repositories. Swaminathan, V., Sorescu, A., Steenkamp, J. B. E., O’Guinn, T. C. lib.utexas.edu/bitstream/handle/2152/31333/LIWC2015_ G., and Schmitt, B. (2020), “Branding in a Hyperconnected LanguageManual.pdf World: Refocusing Theories and Rethinking Boundaries,” Ramsay, James O. and Bernard W. Silverman (2007), Applied Journal of Marketing. Functional Data Analysis: Methods and Case Studies, New Wang, Ze, Huifang Mao, Yexin Jessica Li, and Fan Liu (2017), York: Springer. “Smile Big or Not? Effects of Smile Intensity on Perceptions Rocklage, Matthew D. and Russell H. Fazio (2015), “The of Warmth and Competence,” Journal of Consumer Research, Evaluative Lexicon: Adjective use as a means of assessing and 43(5), 787-805. distinguishing attitude valence, extremity, and emotionality,” White, Kate, Simpson, B. and Argo, Jennifer J. (2014), “The Journal of Experimental Social Psychology, 56, 214–27. Motivating Role of Dissociative Out-Groups in Encouraging Rocklage, Matthew D. and Russell H. Fazio (2016), “On the Positive Consumer Behaviors,” Journal of Marketing dominance of attitude emotionality,” Personality and Social Research, 51(4), 433–47. Psychology Bulletin, 42(2), 259–70. Yang, Yi and Hui Zou (2015), “A Fast Unified Algorithm for Rocklage, Matthew D. and Russell H. Fazio (2020), “The Solving Group-Lasso Penalized Learning Problems,” Statistics Enhancing Versus Backfiring Effects of Positive Emotion in and Computing, 25, 1129-1141. Consumer Reviews,” Journal of Marketing Research, 57(2), 332-52. Letting Go: What Does it Mean for Consumers to Declutter their Offline and Online Lives? Chair: Mariam Humayun, University of Ottawa, Canada

Paper #1: Consuming Order & Discipline ing a home. The study analyzes the rituals consumers engage in to Eric Li, University of British Columbia, Canada reach “implacement” – feeling “at home.” Russell Belk, York University, Canada The third paper “Freed from Desire: Consumers’ Escape from Market Ideologies Through Decluttering Practices?” highlights the Paper #2: Home Sweet Home: How Consumers Regain conflicts that consumers face when dispossessing objects andex- Implacement in Creating the Place of Home plores consumer motives for divestment processes that focus rather Annetta Grant, Bucknell University, USA on the societal consumption ideologies that may hinder the disposal Jay Handelman, Queen’s University, Canada process. Paper #3: Freed from Desire: Consumers’ Escape from Market The final paper “Offline Is The New Luxury? Consuming Ideologies Through Decluttering Practices? Digital Detoxing” examines the emptiness and entanglements of the Birte Manke, University of St. Gallen, Switzerland digital world and how consumers are engaging in digital detoxes and Johanna Gollnhofer, University of St. Gallen, Switzerland letting go of their digital devices and social media as a new form of Paper #4: Offline Is The New Luxury? Consuming Digital consumer resistance. Detoxing Mariam Humayun, University of Ottawa, Canada Consuming Order and Discipline

SESSION OVERVIEW EXTENDED ABSTRACT Embracing this year’s ACR theme of enlightenment, we would Consumerism and materialism are placed under a strict dietet- like to critically examine the complexities of human-object interac- ic regime through the institutionalized insistence that we own too tions in modern consumers’ lives (Epp and Price 2010; Hoffman and much stuff (e.g., Bennett, 2012; Czarniawska and Lofgren 2013; Novak 2018). In this special session we are exploring counter trends Herring, 2014; Löfgren, 2012). Consumers continue to buy and ac- to abundance and excess in various contexts, e.g. digital and physical cumulate material objects in order to feel secure or to compete with decluttering, home organization, and home renovation. referent others (Schor, 1998). Whybrow (2006) argues that Ameri- Consumers’ lives are stuffed with ‘things’, be it offline or online can consumers accumulate experiences and possessions at a rapid (Miller 2008; Johnson 2010). While some consumers still follow a and unhealthy pace due to the accelerating work-spend-consume consumerist logic and accumulate objects, there has been a counter cycle. Similarly, longing for something more is often our response quest for letting go, reducing, and curation of spaces and belongings to a sense of imbalance and insecurity as well as the desire for more – be they physical or mental. The need for reducing chaos in our lives passion in life (Barton, 2014; Charland, 2014; Williamson, 1985). has remained consistent, be it in the ethos of voluntary simplicity or One response is to acquire and accumulate ever more possessions going off to live in the hippie communes of the 1960s (Schor 1998). (Flanders, 2007). This need to “make sense” and to reach a “better place” has become No matter what drives us to buy more and possess more, our more pressing. increasing inventories of possessions lead to the formation of what Drawing on various theoretical lenses ranging from the themes we and others perceive as clutter and chaos in our homes and of- of pollution, conflict, implacement, to consumer resistance, the pa- fices. Consumer advocates and market agents have begun problema- pers in this session bring up manifold questions: How are consum- tizing and stigmatizing over-consumption and excessive accumula- ers in various contexts letting go of objects in their offline and on- tion, and at the same time medicalizing and destigmatizing practices line lives? Some of these consumers are focused on improving their now called hoarding, obsessive compulsive disorder, and addiction, homes to become idealized spaces. Others are engaged in reducing equating these practices with mental disorders (Belk, 1985; Frost and the clutter from their digital lives following the ethos of digital mini- Steketee 2014; Hirschman, 1992; Kasser, 2002; O’Guinn & Faber, malism. What are the differences and similarities of these practices? 1989; Richins & Dawson, 1992; Rook, 1987; Schor, 1998; Van Am- Decluttering can range from permanent dispossession such as selling eringen et al., 2014). Hoarding joined the others as a psychological off a house to live off-the-grid, to more temporal states such as turn- disorder in DSM-5 (American Psychiatric Publishing, 2013). ing off all social media for a day. What role does temporality play? Recent and ongoing criticism of consumerism and hoarding be- This de-cluttering takes place in “spaces” that are constituted as tan- haviors has revised societal definitions of order and disorder, cleanli- gible (offline) to the more ephemeral (online). What is the role of ness and dirtiness in North America. Such criticism also promotes materiality? How has the Covid19 pandemic affected some of these a new desire for order and discipline. As well it has renewed social acts of “letting go”? expectations and definitions of the ideal consumer-citizen. The first paper “Consuming Order & Discipline” derives the Based on our research with a professional organizing firm and key concepts of order and pollution to understand the rise of con- its clients, the aim of this paper is to unpack their co-constructed sumer disciplinary mechanisms in professional home organization disciplinary processes. As consumers become bound by these new services. Findings show how two contesting regimes – the polluting practices, the institutionalized regimes of order and discipline act as ingestion regime of consumption and the disciplinary dietetic regime a domestic “diet.” Our analysis of this intersection of materiality, of organization – govern consumers’ perceptions of self, home, and home dietetics, and consumerism, shows that the vicious cycle of life. discipline/order and regression/disorder binds our self-image and Next, the paper “Home Sweet Home: How Consumers Regain mental health to the way we manage our over-stuffed homes and Implacement in Creating the Place of Home” contributes to the bodies. theory of place through an examination of how consumers return to a Our collaborative study of professional organizers and their more balanced equilibrium by way of the transformation of renovat- clients, also shows that the construction of order and discipline is

Advances in Consumer Research 987 Volume 48, ©2020 988 / Letting Go: What Does it Mean for Consumers to Declutter their Offline and Online Lives? actively mediated by the interactions between these market agents Throughout this literature, the home emerges ubiquitously as and consumers. We find that transitions within a person’s life create a place in Western culture where consumers act out identities, re- a psychological burden and increase levels of uncertainty and insta- lationships, and build social distinction. However, this perspective bility in their lives. Too much consumption, too many possessions, portrays the home as an inert entity, a “receptacle for material and and too much failure in multi-tasking are other sources of domestic social objects… an arena in which objects [are] socially arranged disorganization. Like dietary restrictions, home organization in the and re-arranged” (Prior 1988, 88). For example, Bourdieu (1973, current study is viewed as a learnable skill and a “consumer-inten- 102) adopted a Newtonian worldview of space in which “the house sive” practice (Fischer et al., 2009), since consumers have to engage is organized according to a set of homologous oppositions,” meaning and commit to keeping up with prescribed new tasks and ordering that the home comes to stand as an objectified reflector of social life systems to achieve desirable outcomes. (Prior 1988). Such is the “disenchantment of place and its reconfigu- The societal obsession with home order and organization is ration as a mere subdivision of universal space, inert and homoge- similar to our obsession with slim and well-kept bodies. A cluttered neous” (Sherry 2000, 274). This prior literature tends to regard place home is seen as synonymous with an overweight body. In both cases as an inert platform upon which consumers engage in social activity, programmatic dietary restriction is promoted as one way to achieve but overlooks the human condition that is entwined with place. The societal ideals. Professionals (e.g., dietitians and professional or- philosopher Edward Casey sensitizes us to this dynamic. ganizers) act as coaches who help in structuring and implementing In his theory of place, Edward Casey (1993; 1998, 6) directs our these regimes. Disorganization becomes misconduct and therefore is attention to a predicament of the human condition in which people seen to require regulation and treatment. Our findings also show that seek a feeling of being in place, or “to achieve the assurance offered new organizational practices demand a new lifestyle for consumers by plenitude of place” as an antidote to “place panic; depression or who desire disciplined lives that consolidate their “good citizen- terror even at the idea, and still more in the experience, of an empty consumer” identities. place.” The quest to achieve a “plenitude of place” is driven by a We find that the order-construction process encourages self- fear of “the void of no-place [which] is to be avoided at almost any governance while simultaneously setting norms that are difficult or cost” (Casey 1998, 6). At the extreme, this void of no-place, or “ter- impossible to realistically achieve. The reinforcement of organiza- ror of the void” (Sherry 2000, 274), is evident when considering, tional concepts not only alters consumer practices but also intro- for example, the trope of immigrants as an uprooted people (Bodnar duces a new and harsh regime to their everyday lives. A clutter-free 1985) who are obliged to “leave the place of [one’s] birth to seek home is first collaboratively staged by professional organizers and safety, opportunity, or even basic necessities of life” (Peñaloza 1995, clients through purging and ordering. Order and disorder as well as 83); the struggles of homeless people who fantasize about living in cleanliness and dirtiness are not purely abstract and depersonalized a permanent home while coping with their displacement (Hill 1991); constructs. Rather, they are suffused with moralization, empathy, and or the vulnerability of people who lose their homes as a result of a sorrow (Herz, 2012; Kelly, 2011). When inevitable backsliding into natural disaster (Baker, Hunt and Rittenburg 2007). Just as Roux and clutter, disorder, and chaos occurs, a medicalized rationalization of Belk (2019) note that one’s body is the ultimate place of origin and disease averts blame and self-disparagement. destination that one cannot escape, Casey’s work sensitizes us to the Being organized is now framed as a life choice. Society sets idea that having a place we call home is not an option. new standards, rules, and regulations. The market responds with sys- To build an understanding of the dynamic of the human con- tems to discipline and monitor our bodies and homes. Marketplace dition that is entwined with place, we engaged in an ethnographic organizing services are offered to administer and reinforce these exploration of the home renovation marketplace in which consumers standards. Despite their empathetic intervention, new disciplines find themselves on seemingly endless quests to modify their well- generate fear, anxiety, and discomfort. We discuss the discursive for- functioning homes. This paper examines how the quest for plenitude mation of these disciplinary discourses and show how dietetic prin- of place—implacement—which is grounded in the existential threat ciples render our homes, minds, and bodies as “disciplined spaces.” of no-place, shapes the home renovation marketplace and consumers’ behavior around the modification of their homes. In so doing, Home Sweet Home: How Consumers Regain we address three research questions: How does the fear of no-place Implacement in Creating the Place of Home manifest in a context in which the consumer does not fear literal physical displacement from one’s home? How do modifications to EXTENDED ABSTRACT the home help the consumer address this fear of no-place? What role The home is an important consumer place. Historically, con- do media and marketers play in our understanding of the place of sumer researchers have come to understand the importance of the home and the consumer quest to modify this place? home as foundational in building “the good life” (Belk and Pollay We contribute to the theory of place with an in-depth exami- 1985, 891), which at its core involves the quest for ‘homeyness’— nation of the home renovation marketplace that comprises media a culturally constructed state that consumers create through a con- observation, participant observation in home renovation shows, and stellation of material resources within the home (McCracken 1989). longitudinal depth interviews with consumers engaging in home The home is where consumers enact authentic personal and family renovations. We draw on Edward Casey’s (1993; 1998; 2009; 2017) identities and practices (Epp and Price 2008, 2010; Tian and Belk theory of place as the lens through which to understand our data. We 2005; Wallendorf and Arnould 1991), such as the identity of ‘father’ find that through media portrayals of the home, consumers are con- (Coskuner-Balli and Thompson 2012) and other gendered identities fronted with an anonymous gaze that admonishes them for portrayals (Moisio, Arnould, and Gentry 2013). Within the home, consumers of uniqueness within the home. The cosmogenic origins to which also enact taste regimes as they build social distinction and high cul- the home is thought to align is dysplaced—socially re-engineered— tural capital through the display of fashion, wealth, and elite tastes through this anonymous gaze. This invisible presence leaves consum- (Arsel and Bean 2013; Holt 1998; Üstüner and Holt 2009) while ers with an uneasy feeling within one’s home, setting the consumer communicating and negotiating cultural norms and moral beliefs on a path of renovation and subsequent displacement—physical (Dion, Sabri, and Guillard 2014). dislocation—during the renovation. As such, the home emerges as Advances in Consumer Research (Volume 48) / 989 a place where the consumer’s quest for implacement is constantly books, and articles in blogs, and magazines on the decluttering phe- impeded by forces of dysplacement and displacement. nomenon. The interviews were systematically coded and analyzed. From our research, we first distinguished two types of disposal Freed from Desire – Consumers’ Escape from Market processes. The first type of disposal process has already been stud- Ideologies Through Decluttering Practices? ied: Consumers interact with the object and assess whether the individual meaning is still relevant to them. The second type of disposal EXTENDED ABSTRACT process involves a reflection, not only on personal meanings but also Consumers in western countries have come to a place of mate- on market ideologies, which can be materialized in the ownership of rial abundance. In 2019 Americans spent, on average, $ 1.497 per certain products (e.g. Zanette & Scaraboto, 2019). So far, we identi- month or nearly $ 18,000 per year on non-essential goods (Backman, fied four consumer ideologies that consumers may let go of during 2019). This material abundance has left many consumers living in a decluttering process: ownership, process professionalization, the cluttered environments (Arnold, Graesch, Ragazzini, & Ochs, 2019). pursuit of ideals, fear of scarcity. The first market ideology of own- One estimate suggests that the average American home is filled with ership has already been discussed by Bardhi et al (2017; 2012) and about 300,000 things (MacVean, 2014). Clutter is defined as an “… also Belk (2010). During decluttering process consumers reflect on overabundance of material possessions that collectively create disor- whether they need a physical object. Sometimes they take photos derly and chaotic home environments” (Roster, Ferrari, & Peter Jur- as digital representations or switch to access based services to let kat, 2016, 32). As a solution to the problem of abundance, consumers go of their physical objects. When consumers let go of process pro- start the disposal of objects, which has grown into a consumption fessionalization, they may dispose of the desire to own specialized trend (Cwerner & Metcalfe, 2003; Löfgren, 2017). equipment for certain tasks or events and overall simplify their con- Prior research on disposal focused on the role of consumer’s sumption practices and the belongings involved. Consumers, who let identity (Lastovicka & Fernandez, 2005; Price, Arnould, & Folkman go of the pursuit of ideals, let go of societal expectations regarding Curasi, 2000) and the transfer of meaning trough divestment pro- trendiness or looking fashionable. Instead they opt for more durable, cesses (McCracken, 1986). Building on the theory of the extended timeless classic objects after their disposal experience. Consumers self, Belk (1988) it is theorized that consumers let go of objects that may also let go of the expectation to lose weight and adhere to a no longer serve the purpose of maintaining or extending the self- societal standard of beauty by keeping clothes size to small or very concept (Kleine, Kleine, & Allen, 1995; Lastovicka & Fernandez, uncomfortable clothes that ‘look nice’. Despite living in a world of 2005; Trudel, Arg, & Meng, 2016). abundance, consumers may still hold onto a fear of scarcity. Con- Thompson and Haytko (1997) see personal identity as variable sumers may be prone to respond to deals, special offers, and freebies and cultural meanings as appropriated in a dynamic field of social on the one side and/ or keep consumption objects ‘just in case’. They relations through dialogue. Often consumers make an effort to live start to realize in the process of decluttering that there is no need to up to culturally prescribed standards of living and being (Schouten, be prepared for every instance in advance. 1991). However, consumers may also engage in modes of resistance This study contributes to consumer research by providing in- against prevailing market ideologies and consumption practices sights on conflicting personal and market ideologies. Further, we (Gollnhofer, Weijo, & Schouten, 2019; Martin & Schouten, 2014). introduce the concept of acculturation into the debate around dis- Resisting certain cultural meanings may cause a greater sense of per- posal. Lastly, we contribute to identity theory by highlighting the sonal identity than unconsciously embraced meanings (Thompson & role of disposal for the identity process as prior research only saw Haytko, 1997). Research has shown that disposal processes can be acquisition as relevant for perpetuating and creating a consistent- accompanied by internal identity conflicts, feelings of guilt, and am- self-concept. biguity (Phillips & Sego, 2011). Overall, consumers strive to resolve conflicts around their identity (Ahuvia, 2005). Therefore, we explore “OfflineIs The New Luxury? Consuming Digital the conflicts encountered during the decluttering process. Detoxing” We draw on the concept of acculturation to explain why consumers move away from market ideologies about object ownership EXTENDED ABSTRACT during a decluttering process. The concept of acculturation has previously been applied to the process of risk acculturation for skydiv- “Take the phone off the hook and disappear for a while” – Vienna ers (Celsi, Rose, & Leigh, 1993), but more often in the context of by Billy Joel. migration and minorities (Askegaard, Arnould, & Kjeldgaard, 2005; Luedicke, 2015; Üstüner & Holt, 2007). In the context of changed Back in the day, it was possible to escape the daily grind of life consumer-object-relations and the rise of more liquid consumption, by simply disconnecting one’s phone. There was a time in the early Bardhi et al (2012) started to critically reflect on the role of accul- days of the internet and the world wide web when there was a dif- turation and we would like to build on this discussion. The phenom- ference between offline and online lives. However, these boundaries enological experience of consumer’s object ownership (e.g. being have become ever more porous especially in the developed world. overwhelmed by abundance and clutter) and dominant Western mar- When once people used terms such as “BRB” (Be Right Back) or ket ideologies are in conflict, and these conflicts are materialized in “IRL” (In Real Life), today the infiltration of the digital into our a decluttering process. As these conflicts are resolved during the de- lives is ubiquitous where we are constantly connected and rarely – if cluttering process, consumers adapt their ideas around consumption, ever – exit the system. consumer identity, and object ownership. While social media was once hailed as a utopian technology To understand the conflicts consumers face during the declutter- connecting many, today there has been growing concern about the ing process, we conducted 8 in-depth interviews with recent declut- polarization and tribalism it has created (Lanier 2014; O’Neil 2016; ters and decluttering coaches, resulting in 177 pages of interview Bauman 2017). Democracies and individual freedom have been transcripts (1.5 spaced). For contextualization and triangulation, threatened by the same technologies that once held the promise of we examined series, documentaries, and YouTube videos, popular consumer empowerment and liberation (Zuckerman 2019). Social media was designed to hijack our attention by inviting us down vari- 990 / Letting Go: What Does it Mean for Consumers to Declutter their Offline and Online Lives? ous rabbit holes led by algorithmic systems (Sandvig et al. 2014; Some consumers brand themselves as modern day luddites Harris 2016). Social media and our phones were in fact built to be shunning technologies to find life at its core. Others are inspired addictive devices that keep us coming back for more content. We as by the principles of digital minimalism and mindfulness discourses consumers have taken to using free products offered by the likes of (Newport 2019). Many take to digital detoxing simply because their Google, Facebook, and TikTok, but there has always been an invis- lives have reached burnout points, especially gig-economy workers ible hefty price to pay. As the saying goes, “if you are not paying for or ‘influencers’ (consumers-turned-entrepreneurs) for whom digital the product, you are the product” (Molla 2020). media is their main source of livelihood. At the other extreme are We live in an era of big data where consumer-generated content those who are considered “data fundamentalists” who are complete- has been one of the key drivers of innovation in marketing (Kozinets ly reclaiming privacy by living off-the-grid (Rosen 2010). Market- et al. 2010). While every facet of our lives is becoming ever more ers are increasingly jumping on the digital detoxing bandwagon by digitized, tracked, and monetized (Belk 2013; Kozinets et al. 2017; sponsoring events encouraging “no Wi-Fi”. For example, Lululemon DuFault and Schouten 2018), there have been growing concerns promoting mindfulness through digital cleansing, the rising popular- over our digital futures, be it ethical concerns over AI, the rise of ity of “digital wellness” apps, lifestyle brands like Goop extolling the fake news, extensive surveillance through the social credit system, benefits of being offline, or cities positioning themselves as places to facial recognition technologies, or different forms of trolling and tar- engage in digital detox. geting enabled by social media (O’Neil 2016; Eubanks 2017; Noble Being offline has become a new form of luxury and status sym- 2018). Social media and the internet by their ubiquitous nature have bol. Ironically enough, it is the tech elites who can afford to take off become a part of the background or the digital banal (Dinnen 2018). for offline retreats in the wake of the world they created. The Cov- Technology often creates a paradoxical relationship where con- id19 pandemic has in fact revealed some of these digital inequalities sumers feel connected to each other constantly and at the same time where only a few can really afford to truly disconnect. Others remain feel alienated by these interactions (Mick and Fournier 1998; Kozi- embedded in the vortex of “doomscrolling” or “zooming”. Digital nets 2008; Turkle 2011). There has also been a significant erosion detoxes raise important questions about our digital lives and the pos- of trust in digital technologies of late (Zuboff 2015; Confessore and sibly dystopian futures and the ever-changing social imaginaries that Kang 2018; Karppi 2018; Zuboff 2019). Consumers are seeking to lie ahead. escape the relentless march of the digital through postdigital consumption (Cramer 2015; Humayun and Belk 2017, 2020). Consum- REFERENCES: er resistance has found a new meaning online (Penaloza and Price Ahuvia, A. C. (2005), “Beyond the Extended Self: Loved Objects 1993; Heath and Potter 2004; Kozinets and Handelman 2004). and Consumers’ Identity Narratives,” Journal of Consumer Given that consumers’ data is the “new oil” (Lanier 2014; Research, 32(1), 171–184. Zuboff 2019) and the recognition that we have given up our privacy American Psychiatric Publishing (2013), “Obsessive Compulsive in return for free products, many consumers are seeking to resist and Related Disorders” accessed on February 21, 2020 complete surveillance and reclaim their privacy through engaging in at http://www.dsm5.org/Documents/Obsessive%20 digital detoxes or “going offline”. Escaping the digital realm through Compulsive%20Disorders%20Fact%20Sheet.pdf postdigital consumption has become one form of such mundane re- Arsel, Zeynep and Jonathan Bean (2013), “Taste Regimes and sistance. These consumers are to an extent evangelical in their oppo- Market-Mediated Practice,” Journal of Consumer Research, sition to the cannibalization of the digital realm. Digital detoxing re- 39 (5), 899-917. fers to the practice of purging of digital technologies from daily lives Arnold, J. E., Graesch, A. P., Ragazzini, E., & Ochs, E. (2019), for either permanent or temporal periods (Syvertsen and Enli 2019). “Life at Home in the Twenty-First Century,” Life at Home in This research is based on ethnographic data drawn from con- the Twenty-First Century, Cotsen Institute of Archaeology sumer interviews, participant observation, archival data, and a net- Press. nography focused on subreddits, podcasts, YouTube, social media Askegaard, S., Arnould, E. J., & Kjeldgaard, D. (2005), announcements (e.g. Instagram, Twitter, Facebook posts of “going “Postassimilationist Ethnic Consumer Research: Qualifications offline”), alongside blogposts focused on digital detoxing. While and Extensions,” Journal of Consumer Research, 32(1), these consumers are engaging in digital detoxes, an important part 160–170. of the experience, ironically enough is sharing the experience online. Backman, M. (2019), “Americans are spending $18,000 a year There are various degrees to which consumers engage in digital on nonessentials,” Retrieved February 4, 2020, from https:// detoxing. Some might give up social media for a week – or even eu.usatoday.com/story/money/2019/05/07/americans-spend- a day – while other consumers are choosing to give up using cell- thousands-on-nonessentials/39450207/ phones or go completely “off the grid” by moving away from cities Baker, Stacey Menzel, David M. Hunt, and Terri L. Rittenburg and living in ecovillages. Consumers often announce their “offline (2007), “Consumer Vulnerability as a Shared Experience: weeks” on their social media accounts and return to post about their Tornado Recovery Process in Wright Wyoming,” Journal of offline experiences – be it through blogposts or through email news- Public Policy and Marketing, 26 (1), 6-19. letters which have made a comeback. They connect these conversa- Bardhi, F., & Eckhardt, G. M. (2017), “Liquid Consumption,” tions across disparate social media accounts by using #digitaldetox Journal of Consumer Research, 44(3), 582–597. #offline or #unplugging. Others are also relying on their podcast net- Bardhi, F., Eckhardt, G. M., & Arnould, E. J. (2012), “Liquid work to engage in connected digital detoxes hinting at an entangle- Relationship to Possessions,” Journal of Consumer Research, ment (Lupton 2015). Many influencers and gig economy workers 39(3), 510–529. in fact have been some of the key figures leading the digital detox Bauman, Zygmunt (2017), Retrotopia, Cambridge, UK: Polity movement as they experience burnout from being constantly con- Press. nected and entangled with the online world. Sharing the digital detox encounter often constitutes an important part of the experience. Advances in Consumer Research (Volume 48) / 991

Barton, Adriana (2014), “Anorexia: A ‘new passion’ may help Epp, Amber M., Hope J. Schau, and Linda L. Price (2014), “The people recover,” The Globe and Mail, September 14. Role of Brands and Mediating Technologies in Assembling Retrieved from: https://beta.theglobeandmail.com/life/health- Long-Distance Family Practices,” Journal of Marketing, 78 and-fitness/health/anorexia-a-new-passion-may-help-people- (3), 81-101. recover/article20584523/?ref=http://www.theglobeandmail. Fischer, E., Otnes, C. C., Winegard, B., Li, E. P. H., & Wilner, S. com& J. S. (2009), “Are we there yet? Co-producing Success and Belk, Russell W. (1985), “Materialism: Trait Aspects of Living Failure in a ‘Consumer-intensive’ Service Content”, In J. F. in the Material World,” Journal of Consumer Research, 12 Sherry & E. Fischer (Eds.), Explorations in Consumer Culture (December), 265-280. Theory (pp. 101-113). Abington, UK: Routledge. Belk, Russell W., and Richard W. Pollay (1985), “Images of Flanders, J. (2007), Consuming Passions: Leisure and Pleasure in Ourselves: The Good Life in Twentieth Century Advertising,” Victorian Britain, London, UK: Harper Perennial. Journal of Consumer Research, 11 (4), 887-897. Frost, Randy and Gail Steketee, eds. (2014), The Oxford Handbook Belk, Russell W. (1988), “Possessions and the Extended Self,” of Hoarding and Acquiring, Oxford: Oxford University Press. Journal of Consumer Research, 15(2), 139. Gollnhofer, J. F., Weijo, H. A., & Schouten, J. W. (2019), Belk, R. (2010), “Sharing,” Journal of Consumer Research, 36(5), “Consumer Movements and Value Regimes: Fighting Food 715–734. Waste in Germany by Building Alternative Object Pathways,” Belk, Russell (2013), “Extended Self in a Digital World,” Journal Journal of Consumer Research, 46(3), 460–482. of Consumer Research, 40 (3), 477-500. doi: 10.1086/671052. Harris, Tristan (2016), “How Technology Hijacks People’s Bennett, Jane (2012), “Powers of the Hoard: Further Notes on Minds — From a Magician and Google’s Design Ethicist.” Material Agency,” in Animal, Vegetable, Mineral” Ethics Observer, June 1, 2016.https://observer.com/2016/06/ and Objects, ed. Jeffrey Jerome Cohen, Washington, D.C.: how-technology-hijacks-peoples-minds%E2%80%8A- Oliphaunt, 237-269 %E2%80%8Afrom-a-magician-and-googles-design-ethicist/ Bourdieu, Pierre (1973), “The Berber House,” in M. Douglas, Rules Heath, Joseph, and Andrew Potter (2004), The Rebel Sell: Why the and Meanings, Harmondsworth: Penguin. Culture Can’t be Jammed: HarperCollins. Charland, Louis C. (2014), “The distinction between ‘passion’ Herring, Scott (2014), The Hoarders: Material Deviance in Modern and ‘emotion’ Vincenzo Chiarugi: a case study,” History of American Culture, Chicago: University of Chicago Press. Psychiatry, 25 (December), 477-484. Herz, Rachel (2012), That’s Disgusting: Unravelling the Mysteries Casey, Edward S. (1993), “Getting Back into Place: Toward a of Repulsion, New York, NY: W. W. Norton & Company. Renewed Understanding of the Place World”, Bloomington: Hill, Ronald Paul (1991), “Homeless Women, Special Possessions, Indiana University Press. and the Meaning of “Home”: An Ethnographic Case Study,” ______(1998), The Fate of Place: A Philosophical History, Los Journal of Consumer Research, 18 (3), 298-310. Angeles, CA: University of California Press. Hirschman, Elizabeth C. (1992), “The Consciousness of Addiction: Celsi, R. L., Rose, R. L., & Leigh, T. W. (1993), “An Exploration of Toward a General Theory of Compulsive Consumption,” High-Risk Leisure Consumption Through Skydiving,” Journal Journal of Consumer Research, 19 (September), 155-179. of Consumer Research, 20(1), 1. Hoffman, Donna L., and Thomas P. Novak (2018), “Consumer and Coskuner-Balli, Gokcen, and Craig J. Thompson (2012), “The Object Experience in the Internet of Things: An Assemblage Status Costs of Subordinate Cultural Capital: At-Home Theory Approach,” Journal of Consumer Research, 44 (6), Fathers’ Collective Pursuit of Cultural Legitimacy Through 1178-1204. Capitalizing Consumption Practices,” Journal of Consumer Holt, Douglas B. (1998), “Does Cultural Capital Structure Research, 40 (1), 9-41. American Consumption?” Journal of Consumer Research, 25 Confessore, Nicholas, and Cecilia Kang (2018), “Facebook Data (1), 1-25. Scandals Stoke Criticism That a Privacy Watchdog Too Rarely Humayun, Mariam, and Russell Belk (2017), “Being Dumb in Bites.” New York Times December 30, 2018.https://www. the Age of Smart: Analogue Object Entanglements”, in nytimes.com/2018/12/30/technology/facebook-data-privacy- Association for Consumer Research, Ayelet Gneezy, Vladas ftc.html Griskevicius and Patti Williams, ed. San Diego: Advances in Cramer, Florian (2015), “What Is ‘Post-digital’?”, in Postdigital Consumer Research. 117-122 Aesthetics: Art, Computation & Design, David Berry and Humayun, Mariam, and Russell Belk (2020), “The Analogue Michael Dieter, ed. London: Palgrave Macmillan. 12-26 Diaries of Postdigital Consumption,” Journal of Marketing Cwerner, S. B., & Metcalfe, A. (2003), “Storage and Clutter: Management, 1-27. Discourses and Practices of Order in the Domestic World,” Johnson, Barbara (2010), Persons and Things, Cambridge: Harvard Journal of Design History, 16(3), 229–239. University Press. Dion, Delphine, Ouidade Sabri, and Valérie Guillard (2014), Karppi, Tero (2018), Disconnect: Facebook’s Affective Bonds, “Home Sweet Messy Home: Managing Symbolic Pollution,” Minnesota: University of Minnesota Press. Journal of Consumer Research, 41 (3), 565-89. Kasser, Tim (2002), The High Price of Materialism. Cambridge, Dinnen, Zara (2018), The Digital Banal: New Media and American MA: MIT Press. Literature and Culture, New York: Columbia University Press. Kelly, Daniel (2011), Yuck! The Nature and Moral Significance of Epp, Amber M., and Linda L. Price (2008), “Family Identity: A Disgust, Cambridge, MA: A Bradford Book. Framework of Identity Interplay in Consumption Practices,” Kleine, S. S., Kleine, R. E. K., & Allen, C. T. (1995), “How is a Journal of Consumer Research, 35 (1), 50-70. Possession “Me” or “Not Me”? Characterizing Types and an ______(2010), “The Storied Life of Singularized Objects: Antecedent of Material Possession Attachment,” Journal of Forces of Agency and Network Transformation,” Journal of Consumer Research, 22(3), 327. Consumer Research, 36 (59), 820-837. 992 / Letting Go: What Does it Mean for Consumers to Declutter their Offline and Online Lives?

Kozinets, Robert V, and Jay M. Handelman (2004), “Adversaries O’Guinn, Thomas C. and Ronald J. Faber (1989), “Compulsive of Consumption: Consumer Movements, Activism, and Buying: A Phenomenological Exploration,” Journal of Ideology,” Journal of Consumer Research, 31 (3), 691-704. Consumer Research, 16 (September), 147-157. doi: 10.1086/425104. O’Neil, Cathy (2016), Weapons of Math Destruction: How Big Kozinets, Robert V. (2008), “Technology/Ideology: How Data Increases Inequality and Threatens Democracy, New Ideological Fields Influence Consumers’ Technology York: Crown. Narratives,” Journal of Consumer Research, 34 (6), 865-881. Penaloza, Lisa, and Linda L Price (1993), “Consumer Resistance: A doi: 10.1086/523289. Conceptual Overview,” ACR North American Advances. Kozinets, Robert V., Kristine de Valck, Andrea C. Wojnicki, Phillips, B. J., & Sego, T. (2011), “The role of identity in disposal: and Sarah J. S. Wilner (2010), “Networked Narratives: Lessons from mothers’ disposal of children’s possessions,” Understanding Word-of-Mouth Marketing in Online Marketing Theory, 11(4), 435–454. Communities,” Journal of Marketing, 74 (2), 71-89. Price, L. L., Arnould, E. J., & Folkman Curasi, C. (2000), “Older Kozinets, Robert, Anthony Patterson, and Rachel Ashman (2017), Consumers’ Disposition of Special Possessions,” Journal of “Networks of Desire: How Technology Increases Our Passion Consumer Research, 27(2), 179–201. to Consume,” Journal of Consumer Research, 43 (5), 659-682. Prior, Lindsey (1988), “The Architecture of the Hospital: A Study doi: 10.1093/jcr/ucw061. of Spatial Organization and Medical Knowledge,” The British Lanier, Jaron (2014), Who Owns The Future?, New York: Simon Journal of Sociology, XXXIX (1), 86-113. and Schuster. Richins, Marsha L. and Scott Dawson (1992), “A Consumer Values Lastovicka, J. L., & Fernandez, K. V. (2005), “Three Paths to Orientation for Materialism and Its Measurement: Scale Disposition: The Movement of Meaningful Possessions to Development and Validation,” Journal of Consumer Research, Strangers” Journal of Consumer Research, 31(4), 813–823. 19 (December), 303-316. Löfgren, O. (2017), “Mess: on domestic overflows.”Consumption Rook, Dennis W. (1987), “The Buying Impulse,” Journal of Markets and Culture, 20(1), 1–6. Consumer Research, 14 (September) 189-199. Löfgren, Orvar (2012), “It’s Simply too much! Coping with Rosen, Nick (2010), Off The Grid: Inside the Movement for More Domestic Overflow,” in Managing Overflow in Affluent Space, Less Government, and True Independence in Modern Societies, eds. Barbara Czarniawska and Orvar Löfren, America: Penguin. London: Routledge, 101-124. Roster, C. A., Ferrari, J. R., & Peter Jurkat, M. (2016), “The dark Lupton, Deborah (2015), Digital Sociology, New York: Routledge. side of home: Assessing possession “clutter” on subjective Luedicke, M. K. (2015), “Indigenes’ Responses to Immigrants’ well-being,” Journal of Environmental Psychology, 46, 32–41. Consumer Acculturation: A Relational Configuration Roux, Dominique and Russell Belk (2019), “The Body as Analysis”, Journal of Consumer Research, 42(1), 109–129. (Another) Place: Producing Embodied Heterotopias Through MacVean, M. (2014), “For many people, gathering possessions Tattooing,” Journal of Consumer Research, 46 (3), 483-507. is just the stuff of life”, Retrieved September 10, 2018, from Sandvig, Christian, Kevin Hamilton, Karrie Karahalios, and Cedric http://articles.latimes.com/2014/mar/21/health/la-he-keeping- Langbort (2014), “Auditing Algorithms: Research Methods stuff-20140322 for Detecting Discrimination on Internet Platforms,” Data and Martin, D. M., & Schouten, J. W. (2014), “Consumption-Driven Discrimination: Converting Critical Concerns into Productive Market Emergence”, Journal of Consumer Research, 40(5), Inquiry, 22. 855–870. Schor, Juliet B (1998), The Overspent American: When Buying McCracken, G. (1986), “Culture and Consumption: A Theoretical Becomes You, New York: Basic Books. Account of the Structure and Movement of the Cultural Schouten, J. W. (1991), “Selves in Transition: Symbolic Meaning of Consumer Goods,” Journal of Consumer Consumption in Personal Rites of Passage and Identity Research, 13(1), 71. Reconstruction,” Journal of Consumer Research, 17(4), 412. McCracken, Grant (1988), The Long Interview, Newbury Park, CA: Sherry, John F. (2000), “Place, Technology, and Representation,” Sage. Journal of Consumer Research, 27 (2), 273-278. Mick, David Glen, and Susan Fournier (1998), “Paradoxes of Syvertsen, Trine, and Gunn Enli (2019), “Digital Detox: Media Technology: Consumer Cognizance, Emotions, and Coping Resistance and the Promise of Authenticity,” Convergence, 0 Strategies,” Journal of Consumer Research, 25 (2), 123-143. (0), 1-15. doi: 10.1177/1354856519847325. doi: 10.1086/209531. Tian, Kelly, and Russell W. Belk (2005), “Extended Self and Miller, Daniel (2008), The Comfort of Things, Cambridge, UK: Possessions in the Workplace,” Journal of Consumer Polity. Research, 32 (2), 297-310. Molla, Rani (2020), “Why Your Free Software Is Never Thompson, C. J., & Haytko, D. L. (1997), “Speaking of Fashion: Free.” Vox, January 29, 2020.https://www.vox.com/ Consumers’ Uses of Fashion Discourses and the Appropriation recode/2020/1/29/21111848/free-software-privacy-alternative- of Countervailing Cultural Meanings,” Journal of Consumer data Research, 24(1), 15–42. Moisio, Risto, Eric J. Arnould, and James W. Gentry (2013), Turkle, Sherry (2011), Alone Together: Why We Expect More From “Productive Consumption in the Class-Mediated Construction Technology and Less From Each Other, New York Basic of Domestic Masculinity: Do-It-Yourself (DIY) Home Books. Improvement in Men’s Identity Work,” Journal of Consumer Trudel, R., Arg, J. J., & Meng, M. D. (2016), “The recycled self: Research, 40 (2), 298-316. Consumers’ disposal decisions of identity-linked products,” Newport, Cal (2019), Digital Minimalism: Choosing a Focused Journal of Consumer Research, 43(2), 246–264. Life in a Noisy World, New York: Portfolio/Penguin. Advances in Consumer Research (Volume 48) / 993

Üstüner, Tuba, and Douglas B. Holt (2007), “Dominated Consumer Zanette, M. C., & Scaraboto, D. (2019), “From the corset to Spanx: Acculturation: The Social Construction of Poor Migrant shapewear as a marketplace icon,” Consumption, Markets, and Women’s Consumer Identity Projects in a Turkish Squatter,” Culture, 22(2), 183–199. Journal of Consumer Research, 34 (1), 41-56. Zuboff, Shoshana (2015), “Big Other: Surveillance Capitalism Van Ameringen, Michael, Beth Patterson, and William Simpson and the Prospects of an Information Civilization,” Journal of (2014), “DSM-5 obsessive-compulsive and related disorders: Information Technology, 30 (1), 75-89. Clinical implications of new criteria,” Depression and Anxiety, Zuboff, Shoshana (2019), The Age of Surveillance Capitalism: The 31 (June), 487-493. Fight for a Human Future at the New Frontier of Power, New Wallendorf, Melanie, and Eric J. Arnould (1991), “‘We Gather York: Public Affairs. Together’: Consumption Rituals of Thanksgiving Day,” Zuckerman, Ethan (2019), “Building a More Honest Internet: What Journal of Consumer Research, 18 (1), 13-31. would social media look like if it served the public interest?” Whybrow, Peter C. (2006), American Mania: When More is Not Columbia Journalism Review, Fall 2019.https://www.cjr.org/ Enough, New York: W. W. Norton. special_report/building-honest-internet-public-interest.php Williamson, Judith. (1985), Consuming passions: The dynamics of popular culture, London, UK: Marion Boyars. Want/Should Conflicts in Ourselves and Others Chair: Shwetha Mariadassou, Stanford University, USA

Paper #1: To Look Good or to Be Good? Holding (but Hiding) recommendation and that they are better off as a result, demonstrat- Viceful Motivations for Health Goals ing the effectiveness of personalized nudges in improving consumer Stephanie C. Lin, INSEAD, Singapore well-being. Last, Mehr, Silverman, Sharif, Barasch, and Milkman Kaitlin Woolley, Cornell University, USA take a behavioral economic approach by using an economic game to Peggy J. Liu, University of Pittsburgh, USA model the motivational power of attaining a streak. Across five studies, they find that streak-rewarding payments increase persistence Paper #2: No Diet, No Problem: Social Judgments of Consumers’ in effortful tasks, even relative to higher-paying incentives. These Indulgences results provide implications for assisting consumers in initiating and Geoffrey R. O. Durso, DePaul University, USA achieving “should” goals. Kelly L. Haws, Vanderbilt University, USA Combined, these four papers examine how want/should con- Paper #3: Customizing Algorithmic Recommendations to flicts affect both self-presentation and interpersonal perception and Actual and Ideal Preferences also how firms can motivate behaviors in line with what consumers Poruz Khambatta, Stanford University, USA feel they should want. We believe that this session is well-suited to Shwetha Mariadassou, Stanford University, USA this year’s theme of integrating multiple perspectives as these papers Joshua I. Morris, Stanford University, USA draw on a variety of different methods and literatures, including atti- S. Christian Wheeler, Stanford University, USA tude discrepancies, signaling, goal pursuit, self-control, and nudges. Paper #4: The Motivating Power of Streaks: Incentivizing We hope that this session will attract a diverse audience and provide Streaks Increases Engagement in Effortful Tasks fruitful discussion for innovative research regarding this fundamen- Katie S. Mehr, University of Pennsylvania, USA tal conflict in consumer behavior. Jackie Silverman, University of Delaware, USA Marissa Sharif, University of Pennsylvania, USA To Look Good or to Be Good? Holding (but Hiding) Alixandra Barasch, New York University, USA Viceful Motivations for Health Goals Katherine L. Milkman, University of Pennsylvania, USA EXTENDED ABSTRACT SESSION OVERVIEW Academic research treats health behaviors (e.g., dieting, exer- People often face choices between what they want to do and cise) as virtuous (Berman and Small 2018; Dhar and Wertenbroch what they think they should do. The aim of this session is to explore 2012; Khan, Dhar, and Wertenbroch 2005; Wertenbroch 1998). We the consequences of this intrapersonal conflict and to address ways challenge this healthy=virtuous assumption and propose that con- in which consumer behavior can be shifted to better align with what sumers believe that health behaviors can be vicefully motivated from consumers think they should do. In particular, we consider this dis- a moral perspective, with implications for consumers’ public (vs. pri- crepancy from two angles: first, the impact of the discrepancy on vate) consumption behaviors. interpersonal judgments and behaviors; and second, the potential We propose that appearance motivations for health are viewed nudges that can be utilized to encourage behaviors in line with the as relatively viceful, as they violate internally held moral values “should” self. (e.g., humility, concern for society, Beatty et al. 1985; Gurel-Atay The session will open with two papers that examine the inter- et al. 2010; Schwartz 1994) by reflecting excessive pride and self- personal consequences of want/should conflicts. The first paper will involvement (Cafaro 2005; Netemeyer, Burton, and Lichtenstein focus on how the want vs. should conflict influences consumers’ pub- 1995), and perpetuating unhealthy societal standards. However, lic vs. private behavior. Specifically, Lin, Woolley, and Liu discuss given the prominence of idealized societal standards for beauty self-discrepancy in health goal motivations. Across seven studies, (Richins 1991), consumers cannot avoid internalizing appearance they show that consumers hide their appearance motivations from motivations, leading to a moral conflict: although they are actually others because they feel that they should be less driven by physi- motivated to engage in diet and exercise to improve their physical cal appearance than they actually are. In the second paper, Durso appearance, this violates their internally held moral values. We docu- and Haws discuss observers’ perceptions of an agent experiencing ment this self-discrepancy and examine its implications for consum- a want/should conflict. Whereas prior literature largely deems in- ers’ public (vs. private) behavior. dulgences as normatively “bad,” the authors show that indulgences Studies 1a and 1b (pre-registered; Ns=86, 100) documented this still generally result in positive evaluations (albeit with greater am- moral conflict in Asian and Western populations. Participants viewed bivalence). Thus, they suggest that social judgments of indulgent be- multiple motivations for pursuing health goals (physical appearance, havior are less severe and more nuanced than consumers generally long-term health, mental health, and physical capabilities). Partici- perceive them to be. pants reported how much each motivation reflected their “virtuous” In line with this year’s theme of integrating multiple para- and “actual” motivations. Participants believed that appearance was digm perspectives, the final two papers in this session utilize unique less virtuous than all other motivations, ps<.02. Furthermore, par- methods to nudge consumers into pursuing actions in line with their ticipants indicated that appearance motivations reflected their actual “should” selves. The third paper, by Khambatta, Mariadassou, more than their virtuous motivations, ps<.001, revealing the predict- Morris, and Wheeler, utilizes machine learning models to gener- ed moral conflict; by contrast, all other motivations reflected their ate real-time, personalized predictions tailored to consumers’ actual virtuous motivations equally or more than their actual motivations. or ideal preferences. The authors find that consumers who receive The remaining studies test implications of this self-discrepancy recommendations in line with their ideal (vs. actual) preferences are for consumers’ public (vs. private) behavior. Because consumers are more likely to report that their time is well-spent after they follow the more likely to act in line with their virtuous “ought” self-standards

Advances in Consumer Research 994 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 995 in public than private (Baumeister 1982; Goffman 1978; Gollwitzer condition were told that the program title would be visible to others. and Wicklund 1982; Peloza, White, and Shang 2013), we predicted Those in the private conditions were told the workout would appear that consumers hide their non-virtuous appearance motivations from with a generic “Your Fitness Program” title. Participants were more others. likely to choose the appearance program in the private condition Study 2a (pre-registered; N=326) tested the basic effect. Partici- (56.48%) than either the public (41.28%), p=.026, or private-with- pants chose hashtags for two health photographs (food, gym) to post values-reminder (42.99%), p=.049, conditions. as part of a private or public review. Those who considered posting Overall, this research contributes theoretically to the vice-virtue in private chose more appearance-related hashtags (e.g., #bikini- literature, challenging the notion that health pursuits are inherently body) (M=1.35) than those posting in public (M=.98), p<.001. virtuous. This research also offers practical implications for how Study 2b (N=302) examined real choice in the lab. Participants marketers should frame health programs to encourage private atten- believed their choices would be observed live by others in the lab tion versus public sharing. session or that they would be private. They chose three articles to read from a set of six (three appearance, three non-appearance) ar- No Diet, No Problem: Social Judgments of Consumers’ ticles. Participants chose fewer appearance articles to read in public Indulgences (M=1.24) than in private (M=1.50), p = .013. In Study 3a (N=138), participants were randomly assigned to EXTENDED ABSTRACT imagine posting to a private or public Pinterest board. They indicat- How do people judge the consumer who indulges however they ed their likelihood of posting three appearance-related articles (e.g., “want”, rather than restricts as they “should”? Consumption deci- Beach Body Challenge) and three non-appearance-related articles sions are a common basis for rendering judgments about oneself and (e.g., Strategies for Running). Participants were less likely to post others (Belk 1988; Vartanian, Herman and Polivy 2007), especially appearance-related articles in the public condition (M=1.78) than when decisions involve conflict (i.e., ambivalent feelings; Durso, the private condition (M=2.44), p=.039; this did not occur for non- Brinol and Petty 2016; Keller and Siegrist 2015; Ramanathan and Williams 2007). Research on self-control conflicts and “hedonic appearance articles (Mpublic=2.91; Mprivate=2.90); tinteraction(296)=2.45, p=.014. We further assessed how much appearance motivations re- consumption,” however, has often made strong assumptions that in- flected participants’ actual and virtuous motivations (as in studies 1a dulgences (e.g., eating pizza) are normatively “bad” whereas restric- and 1b). This effect was stronger for those with more moral conflict tions (eating salad) are “good” (Haws, Davis and Dholakia 2016; (i.e., who had a greater discrepancy between how much appearance Shiv and Fedorikhin 1999; cf. Vosgerau, Scopelliti and Huh 2020). Do consumers evaluate others’ indulgent “wants” so negatively reflects their actual and virtuous motivations), tinteration(686)=2.00, p=.046. compared to restrictive “should” choices? Or, similar to many judg- Study 3b (pre-registered, N=401) examined posting behavior ments of their own decision conflicts, do consumers instead feelam - of real Pinterest users. Participants created real private and public bivalent about others’ indulgences? If so, when? This is critical for boards for health articles. They chose three articles to post from predicting consumer behavior, as social judgments of others provide six appearance or six non-appearance related articles, randomly invaluable information toward understanding personal self-control assigned. Participants indicated whether they would post these ar- decisions (Fishbach, Eyal and Finkelstein 2010). ticles to their private or public boards. Participants posted more We examine these questions here. Across three studies, partici- pins to their private board when the posts were appearance-related pants learned about a target, “Bob,” who eats indulgently or restric- (M=1.66) versus appearance-unrelated (M=1.08), p<.001. This ef- tively. We manipulated the presence versus absence of a dieting goal fect was again (directionally) stronger for those higher in moral con- (Study 1), the relevance of the behavior (eating vs. spending) to a dieting goal (Study 2), and whether an extraordinary opportunity was flict, tinteration(397)=1.64, p=.101. Study 4 (pre-registered, N=594) tested whether framing ap- present or not (Study 3) when Bob indulged. We focus on two out- pearance as secondary to an overarching health motivation increases comes: How is Bob evaluated (positively vs. negatively), and how public sharing. Participants imagined receiving a discount for post- ambivalent (e.g., conflicted) do they feel about him? ing about a healthy meal plan subscription to Facebook. The post In Study 1, we examined judgments of Bob in a 2 (dieting goal: described appearance as a primary motivation (“We help you be present vs. absent) x 2 (behavior: indulgent vs. restrictive) between- healthier, so that you can achieve your ultimate goal to look good!”), participants design. Our findings revealed significant interactions secondary motivation (“We help you look good, so that you can on evaluations, F(1,403)=52.43, and ambivalence, F(1,403)=38.77, achieve your ultimate goal of being healthier!”), or did not men- ps<.001. Absent a diet, evaluations were equally positive whether Bob tion appearance (“We help you be healthier!”). Participants decided indulged on pizza (M=7.90) or restricted by eating salad (M=8.22), whether to incur (a real) 30-second cost to privatize the post (i.e., to t(200)=1.30, p=.195. Ambivalence was marginally greater when he make the post not viewable by Facebook friends). Participants priva- indulged (M=4.35) versus restricted (M=3.73), t(200)=1.76, p=.079. tized the post more in the appearance-primary condition (62.3%) When dieting was present, however, evaluations were less than the other two conditions (collapsed 53.2%), p=.034. positive following indulgence (M=6.31) versus restriction (M=9.25), Study 5 (N=324) tested whether people privatize their appear- t(203)=11.48, p<.001. Ambivalence was greater when he indulged ance motivations because they desire to act in line with their moral (M=6.67) versus restricted (M=3.16), t(203)=11.33, p<.001. In all, value systems. Participants were given an ostensibly real choice be- participants evaluated Bob less positively (and more ambivalently) tween two fitness subscription programs that would post daily work- when he indulged (vs. restricted) when dieting goals were present. outs to their Facebook timelines: “Trim Body Program,” (appear- However, Bob was judged equally positively absent a dieting goal, ance) and “Heart Health Program” (non-appearance). Participants regardless of indulging versus restricting. were assigned to one of three conditions: private, public, or private- In Study 2, we examined whether evaluations and ambivalence with-values-reminder. Those in the private-with-values-reminder toward Bob might differ as a function of normative indulgence ver- condition first wrote about why they should or ought to pursue their sus restriction, but now depending on whether his behavior was rele- health goals, as dictated by their moral values. Those in the public vant or irrelevant to dieting in particular. Prior research suggests that 996 / Want/Should Conflicts in Ourselves and Others self-control often manifests domain-specifically (Haws et al. 2016) they ought to behave (prudently) compared to how they desire to rather than generally (cf. Tangney et al. 2004). As such, Bob was behave (indulgently). Overall, we suggest that social judgments of always presented as dieting, but was observed either as eating (goal- others indulgent behaviors are perhaps less severe and more nuanced relevant) or spending (goal-irrelevant) in a normatively indulgent or than has been assumed. restrictive manner. Results revealed that significant interactions of context (eating Customizing Algorithmic Recommendations to Actual vs. spending) by behavior (indulgent vs. restrictive) again emerged and Ideal Preferences on evaluations, F(1,403)=71.09, and ambivalence, F(1,403)=40.94, ps<.001. Spending was evaluated equally positively regardless of its EXTENDED ABSTRACT normative indulgence (buying designer clothing with next-day de- Many of people’s decisions for what to read, what to watch, livery; M=8.09) or restriction (saving it on his wish list; M=8.43), where to eat, and what to buy are made on digital platforms. Giv- t(201)=1.43, p=.155, with equal ambivalence (Ms=4.27, 4.61), en the abundance of available options, algorithms often curate the t(201)=0.98, p=.328. choice set presented to consumers by learning consumers’ prefer- When eating, however, his indulgence (ordering “loaded ences and providing them with recommendations. Such recommen- fries”) was evaluated less positively (M=6.18) and more ambiva- dations can be helpful but can sometimes increase risks to consumer lently (M=6.29) than restriction (ordering salad; Ms=9.35, 2.84), well-being, particularly in cases of biased recommendations (Banker ts(202)=13.26, 10.45, ps<.001. Thus, while participants had less and Khetani 2019; Yeomans et al. 2019). An additional challenge positive evaluations (and more ambivalence) toward Bob when he these algorithms face is that consumers often demonstrate discrepan- indulged (vs. restricted) in his food choice, this was not the case cies between what they actually want and what they ideally want. when Bob indulged or restricted himself when spending (goal-irrele- For instance, consumers might actually want to read about celebrity vant), where he was viewed equally positively and (un)ambivalently. gossip but ideally want to learn more about personal finance. Both Finally, in Study 3 we tested another boundary condition by evaluations have been shown to independently influence subsequent providing a rationalization for indulgent behavior within the domain information processing and behavior (DeMarree et al. 2017; DeMar- of dieting. In this case, Bob was always dieting and always indulging ree et al. 2014; Milkman, Rogers, and Bazerman 2009). as in Study 2—whether his indulgence was goal-relevant (eating) or In this research, we expand on the effectiveness of personal- not (spending) and whether his birthday was mentioned (vs. unmen- ized nudges (Peer et al. 2019) and show that the decisions of recom- tioned) were manipulated orthogonally. Significant interactions of mendation algorithms regarding which types of preferences to tar- context (eating vs. spending) by situation (birthday vs. unmentioned) get can have important consequences for individuals’ consumption emerged on evaluations, F(1,395)=11.86, p=.001, and ambivalence, patterns and consumer well-being. Specifically, we examine how F(1,395)=5.47, p=.020. When Bob shopped indulgently, evalua- people respond to personalized recommendations tailored to actual tions were equally positive whether his birthday was not mentioned preferences versus ideal preferences. To test this, we built an on- (M=8.45) or mentioned (M=8.47), t(197)=0.12, p=.907, with equally line recommendation agent that inferred consumer preferences using low ambivalence (Ms=3.84, 3.91), t(197)=0.21, p=.836. pre-trained machine learning models hosted on a cloud computing When Bob ate indulgently, however, evaluations were less platform. positive when his birthday was not mentioned (M=5.76) versus men- In the training study (N = 1,003), we collected the data needed tioned (M=6.98), t(198)=4.45, p<.001, and reflected greater ambiva- to generate machine learning model predictions. We collected par- lence (Ms=6.33, 5.29), t(198)=3.08, p=.002. As such, we again found ticipants’ actual and ideal preferences for stimuli from a naturalistic that when the behavior was unrelated to the stated goal, participants context: articles posted on Twitter. For 52 tweets about news articles, evaluated Bob equally positively and with low ambivalence. But, we asked participants how likely they were to read each article now when an extraordinary circumstance was provided (versus not pro- and how likely they were to read each article later on 100-point slider vided), even indulgent behavior directly in contrast to a stated goal scales. Additionally, participants reported how much they actually was more positively evaluated and with less ambivalence. and ideally wanted to read each article. Using data from these partic- In sum, contrary to common assumptions that indulgences rep- ipants, we built machine learning models (random forests) to predict resent normatively negative outcomes—via established perspectives actual and ideal ratings for the ten articles with the largest preference in self-control, hedonic consumption, and licensing (Vosgerau et al. discrepancies. To evaluate the accuracy of our models, we used 10- 2020)—these studies reveal that even indulgences result in posi- fold cross-validation, a method in which we generated predictions tive evaluations overall. More specifically, the degree of positivity for 10% of the data at a time using the remaining 90% of the data to and ambivalence were found to vary as a function of relevant goal train the models. We evaluated the accuracy by comparing the pre- information being present versus absent (Study 1) and whether the dicted and true values using the following metrics: the average corre- dieting target indulged in a relevant (eating) or not (shopping) man- lations (r), the square root of the average squared difference (RMSE), ner (Study 2). Even an extraordinary situational attribution used and the average absolute difference (MAE). The average accuracy (r) explicitly to justify the indulgence (Study 3) improved evaluations for actual ratings was 0.54 (RMSE = 29.93, MAE = 25.15), and the while reducing ambivalence, rather than being negatively evaluated average accuracy for the ideal ratings was 0.60 (RMSE = 29.52, MAE as “excuse-making” (Kivetz and Zheng 2006). = 24.36). We used these models to generate real-time predictions in Variation in ambivalence is an especially interesting contri- the main study. bution from these studies, as this is a critical and unique predictor In the main, pre-registered study (N = 6,488), participants were of decision making and social comparison compared with outright presented with 42 tweets about news articles and reported how likely negativity (Durso et al. 2016; Petty 2006; van Harreveld et al. 2009; they were to read each article now and later. Employing an online Zemborain and Johar 2007), with clear implications for self-control cloud computing platform, we used our previously validated ma- (Keller and Siegrist 2015). This represents an unappreciated discrep- chine learning models to generate a real-time personalized recom- ancy between what consumers (and researchers) have characterized mendation for each participant based on their responses. The study as a supposedly “negative” outcome of indulgence based on how employed three conditions. Participants in the actual (ideal) condi- Advances in Consumer Research (Volume 48) / 997 tion received a recommendation for the article that we predicted they tive-compatible, pre-registered experiments, we find that encourag- actually (ideally) wanted to read the most from the set of ten possible ing consumers to complete streaks via incentives (i.e., paying people articles. Participants in the control condition were recommended a for completing three or more effortful tasks consecutively) increases randomly selected article from the set of ten articles. Participants engagement in effortful tasks because doing so feels more game-like. were asked to report their reactions to the recommendations and In Study 1, participants (N=1,104) made five decisions to either were asked if they would rather read their recommended article or engage in a lengthy, effortful task for a bonus payment (looking up an alternative article. Nevertheless, to avoid self-selection effects, the definition of three words) or a shorter, fun task (watching a clip of all participants read their recommended article and subsequently re- a popular television show) without a bonus. In the streak-rewarding ported their reactions to the article. payment condition, participants were paid for choosing to complete For our initial analysis, we compared both the actual and ideal vocabulary tasks consecutively. Participants received a lower reward conditions to the control condition. We found that customized rec- for the first vocabulary task they completed, then a higher amount for ommendations (vs. random recommendations) resulted in more fa- the second and third vocabulary tasks they completed in a row, but vorable impressions of the recommendation system and the article returned to lowest amount for the next vocabulary task completed along all reported measures. This confirms that the recommendations after watching a video. In the flat payment condition, participants were customized to participants’ unique preferences. received the maximum payment amount from the streak-rewarding For our primary analysis, we compared the actual and ideal payment condition each time they completed a vocabulary task; thus, conditions to each other. As predicted, participants whose recom- holding the number of vocabulary tasks completed constant, pay- mendations aligned with their actual preferences were more likely to ment in the flat payment condition was always larger than payment choose to read the article than participants whose recommendations in the streak payment condition. Using a t-test, we found that partici- aligned with their ideal preferences (52% vs. 40%, p < .001). These pants in the streak-rewarding payment condition completed signifi- participants also found more pleasure in reading the article: they re- cantly more vocabulary tasks (M = 2.65, SD = 2.30) than those in the ported that they liked reading the article more (Mactual = 4.12, Mideal = flat payment condition (M = 2.15, SD = 2.17; d = 0.22, t(1068.1) = 3.77, p < .001) and found the article more enjoyable than participants 3.69, p < .001). in the ideal condition (Mactual = 4.02, Mideal = 3.52, p < .001). However, In Study 2 (N = 421), all participants made 10 choices between participants in the ideal condition found the recommendations to be completing CAPTCHA tasks (decoding 6 CAPTCHAs) and watch- more helpful (Mactual = 3.83, Mideal = 4.32, p < .001) than participants ing funny videos. Payment amount was varied to establish robust- in the actual condition and also felt more positively about the recom- ness. Using linear regressions, we found that participants in the mendation agent. They were more likely to report that the organi- streak-rewarding payment condition completed more CAPTCHA zation generating the recommendation had people’s best interest at tasks (M = 6.36, SD = 3.82) than those in the flat payment condition heart (Mactual = 3.43, Mideal = 4.02, p < .001). Finally, after reading the (M = 5.40, SD = 4.04; b = 1.04, t(418) = 2.77, p = .006), and there recommended articles, participants who received recommendations was no significant interaction with incentive size (b =0.93, t(417) tailored to their ideal preferences were more likely to report that their = 1.23, p = .218). Thus, regardless of magnitude, streak-rewarding time was well-spent (Mactual = 3.87, Mideal = 4.26, p < .001) and that incentives encouraged participants to complete more unappealing they were better off as a result (Mactual = 3.20, Mideal = 4.91, p < .001). tasks than larger, flat incentives. Perhaps most importantly for a recommendation service, contrary to In Studies 3 and 4, we use the same basic paradigm to test which the notion that companies must exploit people’s actual preferences features of streak-rewarding payments lead participants to complete to increase engagement, we found that those who received recom- more effortful tasks. In Study 3 (N=712), we compare the two earlier mendations in line with their ideal preferences were more inclined conditions to an increasing payment scheme (i.e., increasing pay- to use the recommendation service again (Mactual = 3.48, Mideal = 3.30, ments for completing non-consecutive CAPTCHA sets). We repli- p = .028). These results were robust to analyzing the data of only cated our finding that streak-rewarding payments (M = 6.71, SD = participants who self-selected to read the article. 3.43) outperformed flat payments (M = 5.38, SD = 4.29; b = 1.33, In conclusion, this research suggests that, rather than exploit t(709) = 3.62, p < .001). In addition, participants in the streak-re- people’s base desires and recommend content they want to consume warding payment condition completed more CAPTCHA sets than (but wish they did not), organizations can benefit from training rec- those in the increasing-payment condition (M = 5.98, SD = 3.71; ommendation algorithms to cater to people’s ideal preferences. In ; b = 0.74, χ2 = 4.76 p = .029), and there was a marginally signifi- doing so, they can also improve overall consumer well-being by as- cant difference between the flat and increasing payment conditions sisting individuals in narrowing the gap between who they are and (b = 0.60, t(709) = 1.70, p = .089). Thus, the increasing nature of who they ideally want to be. streak-rewarding payments does not explain the observed difference between streak-rewarding and flat payments. The Motivating Power of Streaks: Incentivizing Streaks In Study 4 (N=702), we examine the costly nature of skipping Increases Engagement in Effortful Tasks a CAPTCHA set by adding a third streak-without-penalty condition in which payments remained the same, instead of decreasing, after EXTENDED ABSTRACT a skipped CAPTCHA task. Again, streak-rewarding payments (M = When faced with a choice between instantly-gratifying distrac- 7.03, SD = 3.34) outperformed flat payments (M = 5.89, SD = 4.25; tions (e.g., watching television, browsing Facebook) and exerting b = 1.15, t(699) = 3.36, p < .001). Additionally, there was no differ- effort to obtain delayed rewards (e.g., studying for tests, staying fo- ence between performance in the streak-without-penalty condition cused at the office), people often favor distractions, only to later re- (M = 6.88, SD = 3.38) and our standard streak-rewarding payment gret their choices (Bazerman, Tenbrunsel, and Wade-Benzoni 1998; condition (b = 0.16, χ2 = 0.18, p = .670). This suggests that decreas- Kotabe and Hofmann 2015; Milkman, Rogers, and Bazerman 2008; ing payments after a broken streak is not necessary to harness the Thaler and Shefrin 1981). One way to increase engagement in ef- motivating power of streak-rewarding payments. fortful behaviors is to make initiation of goal-related behaviors more In Study 5 (N = 1,204), we test if streak-rewarding incentives appealing (Milkman, Minson, and Volpp 2014). Across five incen- increase persistence in effortful tasks because they make the tasks 998 / Want/Should Conflicts in Ourselves and Others more game-like. Participants were again randomly assigned to ei- Durso, Geoffrey R.O., Pablo Briñol, and Richard E. Petty (2016), ther the streak-rewarding or flat payment condition. After complet- “From Power to Inaction: Ambivalence Gives Pause to the ing the tasks, participants were asked two items regarding the extent Powerful,” Psychological Science, 27 (12), 1660-1666. to which the tasks felt game-like (e.g., how much completing the Dweck, Carol S., and Ellen L. Leggett (1988), “A Social-Cognitive survey felt like a game; r = .74). Importantly, we once again repli- Approach to Motivation and Personality,” Psychological cated the effect of payment condition on number of tasks completed Review, 95 (2), 256-273.

(Mstreak-rewarding = 5.25, SD = 4.12 vs. Mflat = 4.52, SD =4.10; b = 0.727, Festinger, Leon (1954), “A Theory of Social Comparison t(1201.90) = -3.07, p = .002). Averaging responses to both questions, Processes,” Human Relations 7 (2), 117-140. participants in the streak-rewarding payment condition indicated that Fishbach, Ayelet, Tal Eyal, and Stacey R. Finkelstein (2010), “How the study felt marginally more game-like (M = 3.98, SD = 2.12) than Positive and Negative Feedback Motivate Goal Pursuit,” those in the flat payment condition (M = 3.74, SD = 1.96; d = 0.12, Social and Personality Psychology Compass 4 (8), 517-530. t(974.75) = 1.87, p = .061). We also found that the positive effect of Goffman, Erving (1978), The Presentation of Self in Everyday Life, the streak-rewarding payments on number of tasks completed was London: Harmondsworth. mediated at the 90% confidence level by how game-like the tasks Gollwitzer, Peter M. and Robert A. Wicklund (1982), “Admission felt (indirect effect = .13, 95% CI = [-0.004, 0.14]; 90% CI = [0.01, of Failure and Symbolic Self-Completion: Extending Lewinian 0.25]). Theory,” Journal of Personality and Social Psychology, 43(2), In sum, our work suggests that incentivizing consumers to initi- 358–71. ate streaks can increase persistence in effortful tasks, even relative Gurel-Atay, Eda, Guang-Xin Xie, Johnny Chen, and Lynn Richard to higher-paying incentives. The results of our research have several Kahle (2010), “Changes in Social Values in the United States: practical implications. For example, marketers can offer incentives 1976–2007,” Journal of Advertising Research, 50(1), 57–67. that encourage streaks via loyalty programs to increase consumer en- Haws, Kelly L., Scott Davis, and Utpal Dholakia (2016), “Salad = gagement at a lower cost. More broadly, policy makers could harness Success and Fries = Failure? Conceptualizing and Assessing the power of encouraging streaks to help people initiate and meet Self‐Control Outcome Measures in Food Decision‐Making important goals, such as those related to health and savings. Research,” Journal of Consumer Behaviour 15 (2), 99-116. Keller, Carmen, and Michael Siegrist (2015), “Does Personality REFERENCES Influence Eating Styles and Food Choices? Direct and Indirect Banker, Sachin, and Salil Khetani (2019), “Algorithm effects,”Appetite , 84, 128-138. Overdependence: How the Use of Algorithmic Khan, Uzma, Ravi Dhar, and Klaus Wertenbroch (2005), “A Recommendation Systems Can Increase Risks to Consumer Behavioral Decision Theoretic Perspective on Hedonic and Well-Being,” Journal of Public Policy & Marketing, 38(4), Utilitarian Choice,” in Inside Consumption: Frontiers of 500-515. Research on Consumer Motives, Goals, and Desires, eds. S Baumeister, Roy F. (1982), “A Self-Presentational View of Social Ratneshwar and David Glen Mick, New York, NY: Routledge, Phenomena,” Psychological Bulletin, 91(1), 3–26. 144–165. Bazerman, Max H., Ann E. Tenbrunsel, and Kimberly Wade- Kivetz, Ran, and Itamar Simonson (2002), “Self-Control for Benzoni (1998), “Negotiating with Yourself and Losing: the Righteous: Toward a Theory of Precommitment to Making Decisions with Competing Internal Preferences,” The Indulgence,” Journal of Consumer Research, 29 (2), 199-217. Academy of Management Review, 23(2), 225–241. Kivetz, Ran, and Yuhuang Zheng (2006), “Determinants of Beatty, Sharon E., Lynn R. Kahle, Pamela Homer, and Shekhar Justification and Self-Control,”Journal of Experimental Misra (1985), “Alternative Measurement Approaches to Psychology: General, 135 (4), 572-587. Consumer Values: The List of Values and the Rokeach Value Kotabe, Hiroki P., and Wilhelm Hofmann (2015), “On Integrating Survey,” Psychology and Marketing, 2(3), 181–200. the Components of Self-Control,” Perspectives on Belk, Russell W. (1988), “Possessions and the Extended Self,” Psychological Science, 10(5), 618–638. Journal of Consumer Research, 15 (2), 139-168. Milkman, Katherine L., Julia A. Minson, and Kevin GM Volpp Berman, Jonathan Z. and Deborah A. Small (2018), “Discipline and (2014), “Holding the Hunger Games Hostage at the Gym: An Desire: On the Relative Importance of Willpower and Purity in Evaluation of Temptation Bundling,” Management Science, Signaling Virtue,” Journal of Experimental Social Psychology, 60(2), 283–299. 76, 220–30. Milkman, Katherine L., Todd Rogers, and Max H. Bazerman Cafaro, Philip (2005), “Gluttony, Arrogance, Greed, and Apathy: (2008), “Harnessing Our Inner Angels and Demons: What We An Exploration of Environmental Vice,” in Environmental Have Learned About Want/Should Conflicts and How That Virtue Ethics, eds. Ronald Sandler and Philip Cafaro, New Knowledge Can Help Us Reduce Short-Sighted Decision York and Oxford, 135–58. Making,” Perspectives on Psychological Science, 3(4), DeMarree, Kenneth G., Cory J. Clark, S. Christian Wheeler, 324–38. Pablo Briñol, and Richard E. Petty (2017), “On the Pursuit Milkman, Katherine L., Todd Rogers, and Max H. Bazerman of Desired Attitudes: Wanting a Different Attitude Affects (2009), “Highbrow films gather dust: Time-inconsistent Information Processing and Behavior,” 70, 129–42. preferences and online DVD rentals,” Management Science, DeMarree, Kenneth G., S. Christian Wheeler, Pablo Briñol, 55(6), 1047-1059. and Richard E. Petty (2014), “Wanting Other Attitudes: Netemeyer, Richard G., Scot Burton, and Donald R. Lichtenstein Actual–Desired Attitude Discrepancies Predict Feelings of (1995), “Trait Aspects of Vanity: Measurement and Relevance Ambivalence and Ambivalence Consequences,” 53, 5–18. to Consumer Behavior,” Journal of Consumer Research, Dhar, Ravi and Klaus Wertenbroch (2012), “Self-Signaling and 21(4), 612–26. the Costs and Benefits of Temptation in Consumer Choice,” Journal of Marketing Research, 49(1), 15–25. Advances in Consumer Research (Volume 48) / 999

Peloza, John, Katherine White, and Jingzhi Shang (2013), “Good Thaler, Richard H., and Hersh M. Shefrin (1981), “An Economic and Guilt-Free: The Role of Self-Accountability in Influencing Theory of Self-Control,” Journal of Political Economy, 89(2), Preferences for Products with Ethical Attributes,” Journal of 392–406. Marketing, 77(1), 104–9. Van Harreveld, Frenk, Bastiaan T. Rutjens, Mark Rotteveel, Loran Peer, Eyal, Serge Egelman, Marian Harbach, Nathan Malkin, F. Nordgren, and Joop Van Der Pligt (2009), “Ambivalence Arunesh Mathur, and Alisa Frik (2019), “Nudge Me Right: and Decisional Conflict as a Cause of Psychological Personalizing Online Nudges to People’s Decision-Making Discomfort: Feeling Tense Before Jumping Off the Fence,” Styles,” working paper. Available at SSRN 3324907. Journal of Experimental Social Psychology, 45 (1), 167-173. Petty, Richard E (2006), “A Metacognitive Model of Attitudes,” Vartanian, Lenny R., C. Peter Herman, and Janet Polivy (2007), Journal of Consumer Research 33 (1), 22-24. “Consumption Stereotypes and Impression Management: How Ramanathan, Suresh, and Patti Williams (2007), “Immediate You Are What You Eat,” Appetite, 48 (3), 265-277. and delayed emotional consequences of indulgence: The Vosgerau, Joachim, Irene Scopelliti, and Young Eun Huh (2020), moderating influence of personality type on mixed emotions,” “Exerting Self‐Control ≠ Sacrificing Pleasure,”Journal of Journal of Consumer Research 34(2), 212-223. Consumer Psychology, DOI: 10.1002/jcpy.1142. Richins, Marsha L. (1991), “Social Comparison and the Idealized Wertenbroch, Klaus (1998), “Consumption Self-Control by Images of Advertising,” Journal of Consumer Research, 18(1), Rationing Purchase Quantities of Virtue and Vice,” Marketing 71–83. Science, 17(4), 317–37. Schwartz, Shalom H. (1994), “Are There Universal Aspects in the Yeomans, Michael, Anuj Shah, Sendhil Mullainathan, and Jon Structure and Contents of Human Values?,” Journal of Social Kleinberg (2019), “Making sense of recommendations,” Issues, 50(4), 19–45. Journal of Behavioral Decision Making, 32(4), 403-414. Shiv, Baba, and Alexander Fedorikhin (1999), “Heart and Mind in Zemborain, Martin R., and Gita Venkataramani Johar (2007), Conflict: The Interplay of Affect and Cognition in Consumer “Attitudinal Ambivalence and Openness to Persuasion: A Decision Making,” Journal of Consumer Research, 26 (3), Framework for Interpersonal Influence,”Journal of Consumer 278-292. Research, 33 (4), 506-514. Tangney, June P., Roy F. Baumeister, and Angie Luzio Boone (2004), “High Self‐Control Predicts Good Adjustment, Less Pathology, Better Grades, and Interpersonal Success,” Journal of Personality, 72 (2), 271-324. Conspicuous Consumption: Signaling and Compensatory Motivations Chair: Didem Kurt, Northeastern University, USA

Paper #1: I am Product, Hear Me Roar: Social Dominance and insecurity. The authors show that when observers feel insecure about Preference for Loud Products their power, they perceive the driver of a more conspicuous car as Michael Lowe, Georgia Tech, USA having lower status than that of a less conspicuous car. As such, tak- Morgan Ward, Emory University, USA ing the perspective of an observer helps consumers realize this un- Cem Ozturk, Georgia Tech, USA expected consequence of conspicuous consumption (i.e., being seen as engaging in compensatory consumption), preventing them from Paper #2: Obesity and Compensatory Consumption: Evidence paying a premium for a conspicuous item when they feel insecure. from Jewelry Shopping Finally, the fourth paper by Wang and John demonstrates an- Didem Kurt, Northeastern University, USA other unwanted cost of conspicuous consumption: self-control im- Paper #3: Conspicuously Insecure: When Conspicuous pairment after the use of luxury products. The authors argue that us- Consumption Backfires ing luxury products makes women feel more entitled to fulfill their Sean Blair, Georgetown University, USA wants, resulting in lower self-control. Their experiments involving Derek D. Rucker, Northwestern University, USA the use of real luxury products and consumption of unhealthy snack Monika Lisjak, Arizona State University, USA lend support for their thesis. Paper #4: Luxury and Consumer Well-being: Using Luxury Overall, this session brings together papers that use a variety of Products Impairs Women’s Self-Control methods––experiments, field studies, and archival data analysis–– to Yajin Wang, University of Maryland, USA shed more light on the benefits and costs of conspicuous consump- Deborah Roedder John, University of Minnesota, USA tion and status signaling via product purchases.

SESSION OVERVIEW I am Product, Hear Me Roar: Social Dominance and Conspicuous consumption has been on the rise around the Preference for Loud Products world (Deloitte, 2019). Some consumers buy visible and expensive products to communicate their social status (Han et al. 2010), while EXTENDED ABSTRACT others are motivated to bridge the gap between their current and ideal A billboard advertising Porsche shows a picture of the iconic selves through such purchases (Mandel et al., 2017). This special luxury car and the words “IT’S LOUD.” On the other hand, Bosch, session comprises four papers that examine the causes and conse- a luxury appliance brand, positions its dishwasher as the ‘quietest quences of status signaling via product purchases and compensatory model’ in the US. It is evident that the sounds products emit contrib- aspect of conspicuous consumption. The session broadly contributes ute to consumers’ perceptions of them. In this research, we examine to our understanding of how consumers’ self-discrepancies affect how products’ auditory noises function as signals of status. conspicuous consumption activity, and whether consumers’ status Expressions of dominance and prestige represent social strate- signaling efforts actually work and at what cost. The session also gies used to achieve status within a group. The dominance-prestige helps expand the current scope of research on status-driven and com- account of status-striving (Henrich and Gil-White 2001) suggests pensatory consumption by examining novel characteristics such as that social hierarchies arise from two systems of rank allocation: product noise and body-related products as well as by investigating prestige and dominance. The theory states that individuals may in- some unintended social and personal implications of engaging in crease their social status by engaging in behaviors that express one such consumption activity. or both of these qualities. Expressing dominance usually entails be- The first paper by Lowe, Ward and Ozturk analyzes the status haviors that induce fear through intimidation (Chase et al. 2002), signaling effect of product noise. This examination is important be- while expressing prestige refers to individuals who influence others cause related research to date has focused primarily on the role of by being “considered worthy of emulation” for their skills, access to product visibility and brand in status signaling. The authors demon- material resources, or knowledge (Cheng et al. 2013, pg. 105; Cheng strate that a product’s sound also plays an important role how observ- and Tracy 2014). ers make inferences about consumers’ personal qualities and status. In this research, we relate loud and quiet sounds to these two Specifically, the authors find that relatively louder products help con- status-striving strategies. Prior research on non-human primates has sumers signal their social dominance to others, whereas relatively established that having a particularly loud call or roar may establish quieter products communicate social prestige. an animal as being more dominant and consequently, having higher The second paper by Kurt examines the compensatory con- status within the pack (Neumann et al., 2010; Kitchen et al., 2003). sumption motivation behind people’s jewelry purchases. Based on Likewise, amongst humans, speaking louder or in a deeper voice the compensatory consumer behavior model, this paper argues that may afford one a more dominant position in a group (Carney 2020), jewelry shopping interest and spending will be higher in places with however, status may also be conferred to those who speak quietly greater obesity rate because wearing jewelry can help people feel insofar as it signals the refined demeanor of the speaker. To be sure, better about their looks and reduce their appearance-related self-dis- quiet sounds are more easily ignored or overlooked than loud sounds, crepancy. The data obtained from Google Trends, the U.S. Economic however, they are equally effective at signaling product attributes, Census on jewelry retail sales, the U.S. Consumer Expenditure Sur- including high quality (Lageat, Czellar, and Laurent 2003). vey, and Twitter support this theory. Because consumers use products strategically to send desired The third paper by Blair, Rucker and Lisjak highlights the risk social signals about themselves, we predict that consumers wish- that observers may actually perceive conspicuous consumption as a ing to express dominance will show a marked preference for louder compensation for one’s insecurities rather than as a signal of their product sounds, while those wishing to express prestige will prefer achievements. This risk, however, depends on observers’ feelings of quieter products.

Advances in Consumer Research 1000 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1001 The purpose of Study 1 is to demonstrate that using loud prod- Obesity and Compensatory Consumption: Evidence from ucts serves as a signal of social dominance, but not necessarily of Jewelry Shopping social prestige. In a 2 (product: loud vs. quiet) by 2 (user intent: purposeful vs. unintended) factorial design, participants imagined EXTENDED ABSTRACT observing another individual using a product in public (a truck or Obesity rates have been rising across the U.S. over the years, a radio). In one condition, participants read that this product was with four out of every ten adults having obesity (Hales et al., 2017). extremely loud, while in the other condition the product is described Yet, there is limited research on the link between obesity and con- as quiet. Participants also read that the reason for the unusual vol- sumption behavior in non-health domains. In particular, higher ume of the product was either intentional on the part of the driver obesity may fuel appearance-related compensatory consumption (a purposeful modification or setting), or unintentional (the result of because prior studies highlight that individuals with obesity exhibit manufacturer’s setting). After reading the scenario, participants were greater disparagement of body image and report lower self-esteem asked about their impressions of the other consumer (i.e. likeable, than normal weight individuals (e.g., Wadden and Stunkard 1985; dominant, submissive, humble, aggressive, high-status, and impor- Miller and Downey 1999; Wang et al. 2009). Motivated by this ob- tant they perceived this consumer to be). servation, the present research examines whether shopping interest An ANOVA interacting intent and product loudness on per- and spending on jewelry, which is regarded as the architecture of ceived social dominance was significant (F(1,443) = 5.66, p = .018). the body (Beckett 2015), are higher in places with high versus low Specifically, when the volume of the product was seen as uninten- obesity rates. tional, there was no difference between the loud and quiet conditions Obesity stigma is a well-documented effect (e.g., Crandall

(MLoud = 3.61, MQuiet = 3.58; F(1,443) = .018, p = .89). When product 1994; Miller et al. 1995; Hebl and Heatherton 1998; Carr and Fried- volume was seen as intentional, however, there was a significant dif- man 2005; Puhl and Heuer 2009). Prior research has shown that ference in perceptions of social dominance between two conditions people view obese individuals and those around them less favorably such that the individual was seen as more dominant when the indi- than others with normal weight (Hebl and Mannix 2003) and that vidual used the louder product (MLoud = 4.61, MQuiet = 3.95; F(1,443) obese individuals face discrimination due to weight (Sutin, Stephan, = 12.08, p = .001). and Terracciano 2015; Daly, Sutin, and Robinson 2019). People also In Study 2, we tested whether consumers use auditory sounds tend to associate obesity with disease and approach overweight in- to express social status. First, student participants were instructed to dividuals with disgust (Hoyt, Burnette, and Auster-Gussman 2014). imagine that that they were preparing a presentation that they would In turn, overweight individuals try to mitigate disease threats by give at a later date to win a Business Leaders of America compe- choosing to wear clean clothes (Neel, Neufeld, and Neuberg 2013). tition. They were instructed to design the look (e.g. colors, fonts) Despite the growing interest on obesity-focused consumer research, and feel of the presentation. Additionally, they were asked them how there has not been much progress on understanding whether and how quiet or loud they would set their ‘walk in’ music to be (scale of 1 – obesity plays a role in shaping people’s shopping interest and deci- 100). They listened to some EDM music and adjusted the volume to sions pertaining to visible products that may help boost their body their preferred audio level for the presentation and recorded the nu- images. Jewelry, for instance, has historically been viewed as a tool meric value. Lastly, they indicated whether they thought the winner for transforming human body (Gomelsky 2018). With over $68 bil- of this contest was likely to be more prestigious or more dominant lion annual sales in the U.S. (Branstrator 2019), jewelry consump- (1-7 scale). The results confirm that that those indicating the winner tion is therefore a suitable study setting in testing the compensatory should be more prestigious (vs. dominant) set their walk in music consumer behavior model (CCBM; Mandel et al., 2017) with a focus volume lower than those who thought that the winner would be a on appearance and body image threats (Park and Maner 2009; Hoegg dominant individual (F(1,144) = 4.73, p = .031). et al. 2014). In Study 3, we conducted an empirical model to show support According to the CCBM, self-esteem represents an important for these predictions using real marketplace data. Using a data set psychological asset that consumers desire to maintain at a certain a 2014 aggregation of Facebook data capturing consumers’ college level. When consumers’ self-esteem is threatened by social stigmas football team allegiances at a zip code level, we identified 482 zip and other external factors, they try to enhance the self via different codes where fan allegiances were the relatively homogenous among compensatory consumption strategies. Because wearing jewelry can the area’s population. For each zip code, data was provided indicat- help people feel better about their physical appearance and boost ing the 3 most popular teams in that area. Then team performance their self-esteem, one would expect that jewelry shopping interest data (win percentage) was gathered for all teams across 4 years and spending in a locality increases with obesity rate. The present (2011 to 2014). This data was matched with a complete dataset of research tests this hypothesis across four archival studies using on- car sales at the zip code level. Another dataset was used to find the line and field data. All the studies in the paper utilize U.S. data and actual operating volume (dB) of each model of automobile present measure obesity rate either at the state- or county-level depending on in the dataset. Finally, control variables including population, educa- data availability. tion and income were collected. Study 1 uses data obtained from Google Trends for the period We predicted that the better a zip codes’ preferred college foot- 2010-2017 to test whether people’s search interest in jewelry stores ball team performs, the more this population should feel relatively increases with obesity rate. The dependent variable is annual state- dominant, sharing in their team’s success. As such, more dominant level interest score for the search term “jewelry stores”. By construc- zip codes should show a preference for louder vehicles, all else held tion, the online search interest score is standardized by the total vol- equal. The results show that overall, that the better a college football ume of online searches in a given state and year. The independent team performed (resulting in increased feelings of dominance in the variable of interest is state-level obesity rate. The control variables fans), the more zip codes that associated themselves with that team, are median resident income, median resident age, the proportion of sought out louder cars. In addition, there was a main effect such that female residents, the proportion of white residents, the proportion of more educated zip codes (individuals who strive for status using college-graduate residents, and year dummies. The OLS regression prestige signals) showed a preference for quieter vehicles overall. results show that for a one-standard-deviation increase in obesity 1002 / Conspicuous Consumption: Signaling and Compensatory Motivations rate, online search interest for jewelry stores goes up by 12.9% (p < audience is never explicitly told what he means, it is readily under- .01). The results also show that people’s interest in the search term stood that Shrek is suggesting that the massive monument is com- “lose weight”––a proxy for weight- and appearance-related self-dis- pensating for Farquaad’s petite physique. crepancy–– mediates the positive relation between obesity rate and Both common sense and prior literature suggest that conspic- search interest in jewelry stores. Falsification tests conducted with uous consumption is often viewed as a symbol of an individual’s alternative search terms such as “appliance stores” and “hardware achievements. For example, merely wearing luxury brands can lead stores” did not yield any significant results. people to view individuals as having greater status (Cannon and Study 2 analyzes the 2012 U.S. Economic Census data on re- Rucker 2019; Nelissen and Meijers 2011). Yet, Shrek’s observation, tail sales (2017 data is not available yet) and tests the relationship and the audience’s response to it, suggests that conspicuous con- between jewelry store sales and obesity rate across counties. The sumption can sometimes be viewed as a symbol of an individual’s dependent variable is the total sales volume for jewelry stores in a shortcomings instead. The present research explores this latter possi- county (log-transformed). The independent variable of interest is bility. Specifically, this work tests when observers interpret conspic- county-level obesity rate. The control variables include the number uous consumption as an attempt to compensate for a deficit as op- of jewelry stores in a county and the same controls from Study 1 posed to a signal of achievement (Cutright 2012; Gao, Wheeler, and (measured at the county-level). The OLS regression results reveal Shiv 2009). To illustrate, consider a middle-aged man purchasing a that a one-standard-deviation increase in obesity rate is associated sports car. Do observers believe the purchase represents his success with a 13.0% increase in jewelry store sales (p < .01). Falsification and status as an elite businessman, or do they infer that the purchase tests based on sales volume for different types of stores such as elec- is driven by the psychological insecurities of a man in midlife crisis? tronics and appliance stores and hardware stores did not generate any We suggest that observers’ inferences are often influenced by significant results. their own psychological insecurities. Specifically, we propose that Study 3 employs the U.S. Consumer Expenditure Survey data observers are more likely to attribute an actor’s conspicuous con- for the period 2010-2017 to investigate whether consumers living in sumption to compensatory motives when they feel insecure (vs. states with greater obesity rates are more likely to buy jewelry. The not). This prediction builds on prior research showing that insecure data were collected through quarterly interviews with households. individuals desire conspicuous products primarily because they be- Gift purchases were excluded from the analysis. After controlling lieve the products will make them feel more confident in a desired for total spending and various state and household characteristics, self-view (Rucker and Galinsky 2008). Because people often impute the logistic regression results show that for a one-standard-deviation their own thoughts and feelings to others when making attributions increase in obesity rate, the likelihood of jewelry purchase goes up (Gordon 1986; Nickerson 1999), we argue that insecure (vs. secure) by 6.4% (p < .01). The OLS regression of dollar amount of jewelry observers are more likely to believe that an actor’s conspicuous con- spending yields similar results. sumption is similarly motivated by feelings of insecurity. Ironically, Study 4 performs a textual analysis of a large sample of jew- this suggests that although feelings of insecurity make actors more elry-related tweets downloaded between January, 2019 and April, likely to use conspicuous products to gain social acknowledgment, it 2019 to understand whether people from states with greater obesity simultaneously makes observers less likely to provide it. rates use more self and body words in their tweets. Tobit regression Study 1 tested our hypotheses by having participants complete results show that the use of self and body words increases by 5.6% a recall task manipulating power (insecure vs. secure vs. baseline). and 3.1%, respectively, when obesity rate is higher by one standard Subsequently, respondents were shown two Mercedes-Benz cars deviation (p’s < .05). This evidence suggests that as obesity becomes that differed in terms of how conspicuously they were branded. Re- more prevalent, self- and body-related themes stand out as more im- spondents were asked which driver they thought had more status. portant themes in consumers’ expressions about jewelry. Consistent with our theorizing, participants thought that the driver The present research demonstrates an additional cost of obesity. of the more conspicuous car had less status when they felt insecure Unlike weight loss, engaging in compensatory consumption through about their power but more status when they felt secure about their jewelry purchases does not provide a direct resolution to appearance- power (p < .05). This study provides initial support for our theoriz- related self-discrepancy stemming from obesity. Thus, resorting to ing that observers are more likely to attribute an actor’s conspicuous jewelry shopping as a compensatory consumption activity is likely to consumption to compensatory motives when feeling insecure (vs. be detrimental to consumers’ financial, physical, and psychological secure). well-being in the long-run. Study 2 explored the proposed psychological mechanism by showing that the focal effect is attenuated when observers take the Conspicuously Insecure: When Conspicuous perspective of the actor rather than their own perspective, as per- Consumption Backfires spective-taking reduces the tendency to impute one’s own mental states to others. After completing a recall task manipulating security EXTENDED ABSTRACT of their intelligence (insecure vs. secure), respondents were shown two Yale University t-shirts that varied in terms of the conspicuous- Shrek: “So, that must be Lord Farquaad’s castle.” ness of the university’s logo. Respondents were told to imagine they Donkey: “Uh-huh, that’s the place.” saw two people wearing the shirts and were asked to rate which per- Shrek: “Do you think maybe he’s compensating for something?” son felt more secure about his intelligence. Before responding, half the respondents were first asked to take the perspective of the actor. In this scene from the movie Shrek, the eponymous hero and his As expected, observers were more likely to report the conspicuous companion Donkey are approaching the evil Lord Farquaad’s castle. t-shirt signaled insecurity when they themselves were insecure than Having been powerless to prevent the relocation of grating fairytale when they were secure (p < .05). However, this effect was eliminated creatures to his once-tranquil swamp, Shrek has come to the for- when observers engaged in perspective taking (interaction p < .05). midable fortress to confront the man responsible. Laughter ensues, Study 3 tested an implication of observers’ attributions of com- however, when Shrek suggests that the large and conspicuous castle pensatory consumption. Specifically, if insecure observers are more is an indication of the despot’s personal shortcomings. Although the Advances in Consumer Research (Volume 48) / 1003 likely to see conspicuous consumption as driven by compensatory the handbag and were instructed to walk around the building with motives, then insecure actors should be less likely to engage in con- the handbag. Next, each participant was directed to an individual spicuous consumption if they take the perspective of the observer. room which had a computer and a bowl of M&M’s candies on the The idea is that calling an actor’s attention to the observer’s per- desk. The participants were then left alone in the room to complete spective might make salient the possibility that the purchase would the survey. As predicted, we found that participants in the Chosen be seen as compensatory. Thus, the actor might reduce their own Luxury condition consumed significantly more M&M’s (M = 23.33 compensatory behavior to avoid looking like a compensator. To grams, SD = 20.28) than participants in the Non-Luxury condition test this idea, we first manipulated participants’ power (insecure vs. (M = 12.38 grams, SD = 11.93; t(118) = 2.31, p = .023, ηp2 = .043). baseline). Next, respondents were shown a picture of an executive Women in the Given Luxury condition consumed significantly more pen and were asked either how powerful and respected owning the M&M’s (M = 21.81 grams, SD = 24.33) than women in the Non- pen would make them feel (actor condition) or how powerful and Luxury condition (M = 12.38 grams, SD = 11.93, t(118) = 2.08, p = respected others would think owning the pen would make them feel .039, ηp2 = .35). Finally, the amount of M&M’s consumed did not (observer condition). Respondents then provided their reservation differ for women in the Chosen Luxury versus Given Luxury condi- price for the pen. Replicating previous findings in the literature, re- tions (p = .73, ηp2 = .001). Study 1 also tested the role of mood and spondents in the actor condition tended to pay more for the pen if feelings of authenticity and did not find evidences for these effects. they were first led to experience insecurity about their power. How- In Study 2, 174 (Mage = 20.28, SD = 1.57) female students from ever, as predicted, this effect was eliminated when respondents took a large North American university participated and were randomly the perspective of the observer (interaction p < .05). assigned to one of the conditions in a 2 (Product Type: Luxury vs. In conclusion, we find that while engaging in conspicuous con- Non-Luxury) by 2 (Food Type: Healthy vs. Unhealthy) between- sumption may help actors feel more secure, it can also send the exact subjects design. The procedure was similar to study 1 except two opposite signal to observers. For consumers seeking to compensate types of food were provided. The results revealed that participants for a self-threat, this finding suggests that conspicuous consumption in the Luxury condition ate significantly more unhealthy food (M = may actually backfire if observers see it as reflecting psychological 30.12, SD = 17.08) than participants in the Non-Luxury condition insecurity and thus treat them as compensators. Furthermore, this (M = 22.12, SD = 23.84, (F(1,170) = 4.34, p = .039, ηp2 = .025). finding suggests that even when an actor’s behavior is not driven by However, for healthy food consumption, there was no significant dif- compensatory motives, observers may nevertheless make such an inference between participants in the Luxury condition (M = 13.98, ference. In doing so, this research sheds light on potentially unwant- SD = 17.34) and Non-Luxury condition (M = 16.47, SD = 20.85, ed costs of conspicuous consumption and highlights the importance (F(1,170) = .347, p = .56, ηp2 = .002). of considering the social context when studying both conspicuous In Study 3, 126 women (Mage = 39.81, SD = 16.72) attending and compensatory consumption. the State Fair were recruited to participate in the study. After participants walked around at the State Fair with a Louis Vuitton bag Luxury and Consumer Well-being: Using Luxury or a non-luxury bag, snacks were provided while they filled out the Products Impairs Women’s Self-Control survey. Entitlement was also measured (adopted from Campbell, et al. 2004). A one-way ANOVA revealed that women in the Luxury EXTENDED ABSTRACT condition ate significantly more M&M’s (M = 21.16, SD = 22.68) The growth of luxury consumption is one of the most important than women in the Non-Luxury condition (M = 13.49, SD = 17.99; consumer trends in the last decade. Sales of luxury goods have sky- F(1,124) = 4.45, p = .037, ηp2 = .035). Women in the Luxury con- rocketed to over $1.3 trillion globally, and luxury goods are in the dition felt more entitled (M = 3.04, SD = 1.21) than women in the hands of more consumers than ever. Despite the explosion of luxury Non-Luxury condition (M = 2.56, SD = 1.19; F(1,124) = 5.00, p consumption, there is a paucity of research aimed at understanding = .027, ηp2 = .039). Finally, a 5000 resample bootstrap test of the how the use of luxury goods affects our lives, especially how it might indirect effect of luxury use on the amount of M&M’s consumed negatively affect aspects of our well-being. revealed that feelings of entitlement statistically mediated this effect We provide women with an opportunity to use real luxury (b = 2.07, SE = 1.74, 95% CI: [.0297, 6.8526]). products, and find this experience makes them feel more entitled, In our final study, we examine an intervention that luxury con- more deserving of whatever they want and entitled to more than sumers can undertake to counteract the negative effects of luxury other people. Entitlement is known to be associated with a number use on self-control. The key is to identify an intervention that re- of negative psychological traits, such as narcissism and selfishness, duces feelings of entitlement. We propose and test that encouraging but we focus on an under researched consequence of entitlement that a grateful disposition among luxury users would serve this purpose, has particular relevance for consumer behavior—loss of self-control. by reducing feelings of entitlement that lessen self-control. 143 fe- We theorize that feelings of entitlement run counter to exercising male students (Mage = 19.73, SD = 1.54) from a large North Ameri- self-control, where one depresses their immediate wants and desires can university were assigned to one of three conditions: Luxury vs. to meet higher-order goals. Four studies are presented to test our Luxury with Gratitude vs. Non-Luxury. Replicating prior studies, hypothesis. we found that participants in the Luxury condition consumed sig- In Study 1, 121 female students were assigned to one of three nificantly more M&M’s (M = 17.75, SD = 19.82) than participants conditions: Chosen Luxury vs. Given Luxury vs. Non-Luxury. In in the Non-Luxury condition (M = 10.71, SD = 14.51; F(1,140) = the Chosen Luxury condition, women were shown a luxury handbag 4.26, p = .041, ηp2 =.030). Next, women in the Luxury with Grati- (Louis Vuitton) and a non-luxury handbag and were asked to choose tude condition consumed significantly less M&M’s (M = 9.79, SD = one of the handbags for the study. Participants in the Given Luxury 10.80) than women in the Luxury condition (M = 17.75, SD = 19.82; condition were given the Louis Vuitton handbag and were asked to F(1,140) = 4.05, p = .008; ηp2 =.049). In fact, the gratitude interven- use it for the study. Participants in the Non-Luxury condition were tion was so successful that consumption of M&M’s for women in the given a non-luxury cross-body handbag and were asked to use it Luxury with Gratitude condition was similar to that of women in the for the study. Participants were asked to imagine that they owned Non-Luxury condition (M = 9.79 vs. 10.71; F(1,140) = .10, p = .92). 1004 / Conspicuous Consumption: Signaling and Compensatory Motivations

Taken together, our results show that the experience of using Han, Young Jee, Joseph C. Nunes, and Xavier Drèze (2010), a luxury product can have unanticipated negative consequences—it “Signaling Status with Luxury Goods: The Role of Brand can lower consumer’s ability to exert self-control in everyday life. Prominence,” Journal of Marketing, 74 (4), 15-30. Hebl, Micheel R. and Todd Heatherton (1998), “The Stigma of REFERENCES Obesity in Women: The Difference is Black and White,” Beckett, Kathleen (2015), “Jewelry As ‘Architecture for the Personality and Social Psychology Bulletin, 24 (4), 417-426. Body’,” The New York Times, https://www.nytimes. Hebl, Michelle R. and Laura M. Mannix (2003), “The Weight of com/2015/12/07/fashion/jewelry-as-architecture-for-the-body. Obesity in Evaluating Others: A Mere Proximity Effect,” html. Personality and Social Psychology Bulletin, 29 (1), 28-38. Branstrator, Brecken (2019), “US Jewelry, Watch Spending Rose Henrich, Joseph and Francisco J. Gil-White (2001), “The Evolution 7% Last Year. National Jeweler,” https://www.nationaljeweler. of Prestige: Freely Conferred Deference as a Mechanism for com/majors/market-developments/7602-us-jewelry-watch- Enhancing the Benefits of Cultural Transmission,” Evolution spending-rose-7-last-year. and Human Behavior, 22 (3), 165-196. Cannon, Christopher and Derek D. Rucker (2019), “The Dark Side Hoegg, JoAndrea, Maura L. Scott, Andrea C. Morales, and Darren of Luxury: Social Costs of Luxury Consumption,” Personality W. Dahl (2014), “The Flip Side of Vanity Sizing: How and Social Psychology Bulletin, 45 (5), 767-779. Consumers Respond to and Compensate for Larger than Carney, Dana R. (2020), “The Nonverbal Expression of Power, Expected Clothing Sizes,” Journal of Consumer Psychology, Status, and Dominance,” Current Opinion in Psychology, 24(1), 70-78. forthcoming. Hoyt, Crystal L., Jeni L. Burnette, and Lisa Auster-Gussman Carr, Deborah and Michael A. Friedman (2005). “Is Obesity (2014), “’Obesity is a disease’: Examining the Self-Regulatory Stigmatizing? Body Weight, Perceived Discrimination, and Impact of This Public-Health Message,” Psychological Psychological Well-Being in the United States,” Journal of Science, 25 (4), 997-1002. Health and Social Behavior, 46 (3), 244-259. Gomelsky, Victoria (2018), “How Jewelry Makes Us Who We Are,” Chase, Ivan D., Craig Tovey, Debra Spangler-Martin, and The New York Times, https://www.nytimes.com/2018/11/30/ Michael Manfredonia (2002), “Individual Differences versus fashion/jewelry-metropolitan-museum-of-art.html. Social Dynamics in the Formation of Animal Dominance Gordon, Robert M. (1986), “Folk Psychology as Ssimulation,” Hierarchies,” Proceedings of the National Academy of Mind & Language, 1 (2), 158-171. Sciences, 99(8), 5744-5749. Kitchen, Dawn M., Dorothy L. Cheney, and Robert M. Seyfarth Cheng, Joey T., Jessica L. Tracy, Tom Foulsham, Alan Kingstone, (2003), “Female Baboons’ Responses to Male Loud Calls,” and Joseph Henrich (2013), “Two Ways to the Top: Evidence Ethology, 109 (5), 401-412. that Dominance and Prestige Are Distinct Yet Viable Avenues Lageat, Thierry, Sandor Czellar, and Gilles Laurent (2003), to Social Rank and Influence,”Journal of Personality and “Engineering hedonic attributes to generate perceptions Social Psychology, 104 (1), 103. of luxury: Consumer perception of an everyday sound,” Cheng, Joey T. and Jessica L. Tracy (2014), “Toward a Unified Marketing Letters, 14 (2), 97-109. Science of Hierarchy: Dominance and Prestige are Two Mandel, Naomi, Derek D. Rucker, Jonathan Levav, and Adam D. Fundamental Pathways to Human Social Rank,” in The Galinsky (2017), “The Compensatory Consumer Behavior Psychology of Social Status, New York, NY: Springer, 3-27. Model: How Self-Discrepancies Drive Consumer Behavior,” Crandall, Christian S. (1994), “Prejudice Against Fat People: Journal of Consumer Psychology, 27 (1), 133-146. Ideology and Self-interest,” Journal of Personality and Social Miller, Carol T. and Kathryn T. Downey (1999), “A Meta-Analysis Psychology, 66 (5), 882-894. of Heavyweight and Self-Esteem,” Personality and Social Cutright, Keisha M. (2012), “The Beauty of Boundaries: When Psychology Review, 3 (1), 68-84. and Why We Seek Structure in Consumption,” Journal of Miller, Carol T., Esther D. Rothblum, Diani Felicio, and Pamela Consumer Research, 38 (5), 775-790. Brand (1995), “Compensating for Stigma: Obese and Daly, Michael, Angelina R. Sutin, and Eric Robinson (2019), Nonobese Women’s Reactions to Being Visible,” Personality “Perceived Weight Discrimination Mediates the Prospective and Social Psychology Bulletin, 21 (10), 1093-1106. Association between Obesity and Physiological Dysregulation: Neel, Rebecca, Samantha L. Neufeld, and Steven L. Neuberg Evidence from a Population-Based Cohort,” Psychological (2013), “Would an Obese Person Whistle Vivaldi? Targets of Science, 30 (7), 1030-1039. Prejudice Self-Present to Minimize Appearance of Specific Deloitte (2019), “Global Powers of Luxury Goods 2019: Bridging Threats,” Psychological Science, 24 (5), 678-687. the Gap Between the Old and the New,” https://www2.deloitte. Nelissen, Rob M. A. and Marijn H. Meijers (2011), “Social Benefits com/global/en/pages/consumer-business/articles/gx-cb-global- of Luxury Brands as Costly Signals of Wealth and Status,” powers-of-luxury-goods.html. Evolution and Human Behavior, 32 (5), 343-355. Gao, Leilei, S. Christian Wheeler, and Baba Shiv (2009), “The Neumann, Christof, Gholib Assahad, Kurt Hammerschmidt, Dyah ‘Shaken Self’: Product Choices as a Means of Restoring Self- Perwitasari-Farajallah, and Antje Engelhardt (2010), “Loud View Confidence,”Journal of Consumer Research, 36 (1), Calls in Male Crested Macaques, Macaca Nigra: A Signal of 29-38. Dominance in a Tolerant Species,” Animal Behaviour, 79 (1), Hales, Craig M, Margaret D. Carroll, Cheryl D. Fryar, C.D., and 187-193. Cynthia L. Odgen (2017), “Prevalence of Obesity Among Nickerson, Raymond S. (1999), “How We Know—and Sometimes Adults and Youth: United States, 2015-2016,” NCHS Data Misjudge—What Others Know: Imputing One’s Own Brief, 288, 1-8. Knowledge to Others,” Psychological Bulletin, 125 (6), 737- 759. Advances in Consumer Research (Volume 48) / 1005

Park, Lora E. and Jon K. Maner (2009), “Does Self-Threat Sutin, Angelina R., Yannick Stephan, and Antonio Terracciano Promote Social Connection? The Role of Self-Esteem and (2015), “Weight Discrimination and Risk of Mortality,” Contingencies of Self-Worth,” Journal of Personality and Psychological Science, 26 (11), 1803-1811. Social Psychology, 96 (1), 203-217. Wadden, Thomas A. and Albert J. Stunkard (1985), “Social and Puhl, Rebecca M. and Chelsea A. Heuer (2009), “The Stigma of Psychological Consequences of Obesity,” Annals of Internal Obesity: A Review and Update,” Obesity, 17 (5), 941-964. Medicine, 103 (6), 1062-1067. Rucker, Derek D. and Adam D. Galinsky. (2008), “Desire to Wang, F., T. C. Wild, W. Kipp, S. Kuhle, and P. J. Veugelers (2009), Acquire: Powerlessness and Compensatory Consumption,” “The Influence of Childhood Obesity on the Development of Journal of Consumer Research, 35 (2), 257-267. Self-Esteem,” Health Reports, 20 (2), 21-27. Splash Out or Cheap Out? Motivational Influences on Consumers’ Management of Tradeoffs Chairs: Lawrence E Williams, University of Colorado, USA Nicole L. Mead, York University, Canada

Paper #1: Meaning Comes Cheap: The Pursuit of Meaning and loans. They find that refinancing and loan-type decisions are influ- the Preference for Low Price Options enced by the decisions of their neighbors. In particular, households Nicole L. Mead, York University, Canada that are relatively higher in value are more likely to exhibit the social Lawrence E Williams, University of Colorado, USA influence effects as compared to relatively cheaper households, suggesting that expensive houses exert a greater social influence effect. Paper #2: Hurts So Good: Status Products that Incur Questions emerging from this session might include the follow- Environmental Costs Are Preferred by Status Striving ing: Consumers • Which social processes are more likely to move consumers Morgan K Ward, Emory University, USA toward luxury versus frugality? Sara Loughran Dommer, Georgia Tech, USA • How might the pursuit of meaning (vs. pleasure) interact Darren Dahl, University of British Columbia, Canada with the felt need for status to drive consumption deci- Paper #3: Status Pivoting: Coping with Status Threats through sions? Motivated Trade-off Beliefs and Consumption in Alternative • How might households’ status concerns interact with nor- Domains mative social influences to drive economic decision-mak- Dafna Goor, London Business School, UK ing? Anat Keinan, Harvard Business School, USA Importantly, this collection of papers utilizes a diverse array of Nailya Ordabayeva, Boston College, USA methods to explore how consumers’ goals shape the tradeoffs they Paper #4: ‘I’ll Have What She’s Having:’ Neighbors Socially make, including field studies, lab experiments, incentive compatible Influence Households’ Mortgage Decisions choice designs, observational methods, interviews, and large-scale W. Benadict McCartney, Purdue University, USA panel data analysis. Given its multimethod nature, this session pro- Avni Shah, University of Toronto, Canada vides ideal conditions for a productive meeting of the minds, where researchers of different stripes can come together to generate new SESSION OVERVIEW insight into when and why consumers splash out versus cheap out. George Orwell’s Down and Out in Paris and London describes how he managed to live in Paris on pennies a day, a tale that stands Meaning Comes Cheap: The Pursuit of Meaning and the in sharp contrast to the City of Lights’ reputation for luxury and Preference for Low Price Options elegance. In Paris, Orwell found himself deciding between paying rent (and sleeping comfortably indoors) or sleeping on park benches EXTENDED ABSTRACT to afford wine on the weekends. Although few people experience Consumers strive to find and create meaning in their daily lives. the extremes Orwell faced, tradeoffs are inherent to any consumer Yet, despite this fundamental need, there is a paucity of research that choice, as people decide which options to consider, which ones to has examined how consumers use the marketplace to help them in choose, and how much to pay for those chosen options. For example, their pursuit of meaning. When consumers wish to add meaning to when planning vacations we decide whether to stay in expensive pre- their lives, what do they want? mium hotels or cheaper budget ones, or when commuting we decide Past research suggests that pursuing meaning may cause people whether to rely on public transportation or less ecofriendly single to “splash out’ by seeking higher-quality, more expensive products. passenger rides. Meaning is associated with a desire for self-growth (Baumeister et Collectively, these papers seek to add to the literature on how al. 2013) and the belief that larger investments will provide longer- consumers’ goals shape their management of tradeoffs. To kick lasting benefits (Percival Carter and Williams 2017). Those desires things off, in “Meaning Comes Cheap...” the authors show that par- are best fulfilled by high-quality products because those products can ticipants who were induced to derive meaning (vs. pleasure) from be incorporated into consumers’ sense of self over time (Belk 1988). consumption tend to cheap out – that is, they displayed a preference In this research, we posit (and find) the opposite: meaning for lower-end (vs. higher-end) products. The authors of the second causes people to “cheap out” by gravitating toward less expensive paper “Hurts So Good…” provide novel insight into why consumers options. This effect is robust across a variety of usage situations and sometimes want luxury products that are not only financially costly diverse material goods, experiences, and services, and this effect is but also environmentally costly. They show that consumers who are explained by an enhanced focus on opportunity costs. The distinct motivated to achieve high status desire luxury products that are en- thought processes associated with meaning (increased thoughtful- vironmentally costly because those products symbolically express ness, thinking about non-monetary ways to generate meaning, and consumers’ desire for dominance. Following on the status theme, in a focus on the future) encourage people pursuing meaning to think “Status Pivoting…” the authors provide new insight into why and about opportunity costs – the alternative ways they can spend their where consumers splash out when they experience a status threat. In limited time, energy, and money – which in turn makes them reluc- reaction to an upward status comparison, consumers are less likely to tant to devote their resources to any given high-price option. engage in within-domain compensatory consumption than they are to We report five studies which provided support for our hypoth- pivot to high status products in alternative domains. Finally, in “I’ll eses. Consistent with previous research studying meaning, we con- Have What She’s Having…” the authors examine how social influ- trasted the pursuit of meaning with the pursuit of pleasure. Not only ence changes households’ economic decisions regarding mortgage is this a conceptually and empirically valid approach, it is also eco-

Advances in Consumer Research 1006 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1007 logically valid as pleasure is a related but distinct motivation that participants reported a higher preference for less expensive prod- is considered to be one of the most important drivers of consumer ucts (p<.001). More important, this effect was mitigated when the choice (Alba and Williams 2013; Holbrook and Hirschman 1982). link between quality and durability was made salient to participants In experiments 1a (N=173; undergraduate participants) and 1b (p=.330). This attenuation was due to the durability reminder (vs. (N=181; Amazon Mechanical Turk (mTurk) workers), respondents no reminder) increasing preferences for more expensive products induced to purse meaning (vs. pleasure) reported a stronger prefer- among meaning participants (p<.001). Confirming that the moder- ence for the less-expensive alternatives across multiple product cat- ator was having an effect on meaning participants specifically, the egories (cars, coffee, skiing, dining, cameras, and water) and within durability reminder did not have an effect on participants in the single-use products (e.g., a cup of coffee) and multiple-use products pleasure condition (p = .740). (e.g., a coffee maker) of each product category (experiment 1a: This research provides a first look at how the pursuit of meaning F(1,171)=14.86, p<.001; experiment 1b: F(1,179)=20.460, p<.001). guides marketplace choices. Across five studies, we find that the pur- To check the robustness and external validity of our findings, suit of meaning (vs. pleasure) causes consumers to gravitate toward we conducted an incentive-compatible study which was pre-regis- less expensive options because of a heightened focus on opportunity tered (experiment 2; N=193 MTurkers). Participants were asked to costs. The findings thus point to a disconnect between what people go shopping on Amazon.com, and they were randomly assigned to say they want when engaged in meaningful consumption (more ex- choose something to create meaning (or pleasure). Participants were pensive products that last longer) and what they choose when pur- given a budget of $100, but otherwise they were free to select what- suing meaningful consumption (less expensive products that fail ever they wished. To make this study incentive compatible, partici- sooner). pants were told they would be entered into a draw to win the product selected, and that if they spent less than $100 on the product, they Hurts So Good: Status Products that Incur would receive the remainder in cash (i.e., MTurk bonus). This lottery Environmental Costs Are Preferred by Status Striving was executed. Conceptually replicating experiments 1a-1b, partici- Consumers pants who were induced to pursue meaning chose a less expensive product (M=$38.23, SD=$23.52) as compared to those induced to EXTENDED ABSTRACT pursue pleasure (M=$49.71, SD=$27.72), t(191)=3.100, p=.002. The motivation to strive for elevated status within one’s so- In experiment 3 (N=437; Prolific workers), we tested the hy- cial circle is evident across cultures (Barkow 1989). Although the pothesis that pursuing meaning (vs. pleasure) causes people to particular behaviors and objects that enable upward mobility differ “cheap out” because the pursuit of meaning heightens consum- considerably across culture (Bian and Forsythe 2012), the behavioral ers’ focus on alternative uses for their resources (i.e., opportunity strategies used to achieve high-status are universal. The dominance- costs). Consistent with predictions, we observed an indirect effect prestige account of status-striving (Henrich and Gil-White 2001) of the meaning (vs. pleasure) manipulation on increased preference states that individuals may increase their social status by expressing for less-expensive products through heightened focus on opportu- dominance, which usually entails displaying products or engaging nity costs for both single-use products (c’=.1666, SE=.0464, 95%CI in behaviors that represent an individual’s control over other entities [.0820, .2631]) and multiple-use products (c’=.1815, SE=.0514, through force or aggression (e.g. individuals, animals, surroundings; 95%CI [.0877, .2867]). Maeng and Aggrawal 2017; Chase et al. 2002) and by expressing In experiment 4 (N=344; Prolific workers), we tested a theo- prestige, which is accomplished by displaying material resources or retically relevant boundary condition. If pursuing meaning height- knowledge considered “worthy of emulation” (Cheng et al. 2013, pg. ens consumers’ focus on opportunity costs, then they may neglect 105; Cheng and Tracy 2014). to think of the longer-term benefits of more expensive products – Both prestige and dominance are effectively signaled to others namely, that high-quality products last longer. In this way, remind- vis a vis the products consumers use and display. Although modern ing consumers that higher-quality products tend to last longer should consumers seldom procure products via physical force, symbols of mitigate the effect of meaning (vs. pleasure) on increased preferenc- dominance are embedded in certain luxury products via the “envi- es for less-expensive products, but only when products can be used ronmental costs” that they incur in their sourcing or production (i.e. over time (i.e., multiple-use products not single-use products). In this the physical destruction, or damage that occurs to an environmental way, the experiment was a 2 (pleasure vs. meaning) by 2 (durability entity as an outcome of the procurement of a material or the produc- reminder: yes vs. no) by 2 (single-use product vs. multiple-use prod- tion of a good). Environmentally costly products symbolically ex- uct) mixed design, with consumption goal and durability reminder as press the producer’s (and by extension, the consumer’s) dominance between-subjects factors and product longevity as a within-subjects over another living being or resource. factor. We contend that products that incur environmental costs are We observed the predicted three-way interaction between particularly desirable to those engaged in status-striving. We theorize consumption goal, durability, and product longevity (F(1, 340) = that goods that are symbolic of both paths (i.e. prestige and domi- 7.022, p = .008). Examining preferences for single-use products, we nance) are particularly appealing to individuals who are seeking observed the predicted main effect of the goal manipulation (F(1, higher social status. The ego-enhancing benefits of luxury coupled 340)=32.046, p<.001). Conceptually replicating the previous stud- with the signals of physical threat makes the consumption of envi- ies, participants in the meaning (vs. pleasure) condition reported a ronmentally costly luxury items particularly effective tools in up- stronger preference for the lower-quality products. ward mobility (Sivanathan and Pettit 2010). When predicting preferences for multiple-use products, we While some might view scarcity as the inevitable result of the observed the predicted interaction between the durability reminder environmental costs inherent in the production of luxury products, and the consumption goal manipulation (F(1, 340)=8.149, p=.005). scarcity and environmental costs are orthogonal constructs. Thus, we The results of pairwise comparisons supported our theorizing. When acknowledge the importance of scarcity in consumers’ desire for luxury participants were not reminded that higher-quality products last but control for scarcity perceptions in our studies. longer, we replicated our previous findings: meaning (vs. pleasure) 1008 / Splash Out or Cheap Out? Motivational Influences on Consumers’ Management of Tradeoffs Study 1 was a 2 (environmental costs: high vs. low) x 2 (status- Status Pivoting: Coping with Status Threats through striving: high vs. low) behavioral study (n = 205). First, participants Motivated Trade-off Beliefs and Consumption in were instructed to write about why choosing status goods makes Alternative Domains sense (high status-striving condition) or does not make sense (low status-striving condition). Next, an RA asked participants to take part EXTENDED ABSTRACT in a taste test for a luxury cold-pressed juice brand. Those who agreed Prior research suggests that consumers cope with status threats read about juice made from Camu Camu fruit. Only in the high en- by engaging in compensatory consumption, i.e. by purchasing and vironmental cost condition were participants told that harvesting the displaying status-related products and brands that signal professional fruit has a detrimental effect on the fauna and flora surrounding it. success and affluence to restore their “shaken self” (Gao et al. 2009; Participants were then given samples of the juice. Participants en- Mandel et al. 2017). Going beyond compensatory consumption in the dorsed the degree to which they liked the juice, found it desirable domain of the status threat, we examine when and how consumers and thought it tasted good (1 = not at all/ 9 = very much; α = .94). cope with status threat by choosing to pivot to alternative domains. Participants in the high status-striving condition found the juice with We argue that upward comparisons to higher-status individuals mo- high environmental costs more appealing compared to the juice with tivate consumers to adopt beliefs about trade-offs between success in low environmental costs (MHEC = 6.97 vs. MLEC = 5.93, F(1, 202) = the domain of the threat (e.g., professional or financial success) and 9.38, p = .002). This pattern was reversed in the low status-striving alternative domains (e.g., personal life or physical fitness). These condition (MHEC = 6.01 vs. MLEC = 6.59, F(1, 202) = 2.64, p =.10). motivated trade-off beliefs, which are specifically pronounced when Study 2 manipulated one between-subjects factor (environmen- consumers experience a status threat, fuel consumption and display tal costs: high vs. low) and measured status-striving (Eastman et al. of brands associated with accomplishments in these alternative do- 1999). Participants (n = 131) were told that the marketing department mains. We investigate the domains consumers pivot to, compare the was working with the university bookstore to determine which kind of prevalence and appeal of the “status pivoting” behavior to compen- luxury chocolate squares they should carry. At each computer was 10 satory consumption within the threat domain, and examine when chocolate samples and a booklet offering information about the luxury consumers are more or less likely to engage in status pivoting. chocolate. In the high environmental cost condition participants read Eight studies combining field and lab experiments, observation- that cocoa farming contributes to deforestation and damage to wildlife al methods, incentive-compatible designs, and an analysis of social habitats. Before leaving, participants were told they could take as many media posts, explore the status pivoting phenomenon. chocolate samples from their stations as they liked. RAs recorded how Study 1 examined the prevalence and appeal of status pivoting many chocolates participants took (0-10). Two weeks later, a follow- in a real-world natural setting. We recorded and coded 217 responses up email survey measured status-striving (Eastman et al. 1999). For to a question posted on an online platform Quora.com: “How can high status-strivers (+1 SD), the presence of high environmental costs you overcome your envy of people who are your age but are far more caused an increase in the number of chocolate samples taken (β = .43, successful?” A greater proportion of posts mentioned pivoting to an p = .0002). For low status-strivers (-1 SD), the presence of high envi- alternative domain (48.8%) than restoring status within the threat do- ronmental costs caused a decrease in the number of chocolate samples main (25.3%; c2=6.031, p=.014). The most frequently mentioned al- taken (β = -.33, p = .03). ternative domain was social life and relationships (30.2%), followed Study 3 examines our hypothesized dominance process via mod- by mental wellness (23.6%), personal character and moral values eration. We conducted a 2 (environmental cost: high vs. low) X 2 (20.8%), and physical health (15.1%). Furthermore, comments about (dominance: validated vs control) between subjects design with status- status pivoting were viewed (p<.001) and liked (p=.006) more than striving measured. Participants (n = 323) were primed by writing about those about restoring status within the threat domain, indicating its a time that they had felt dominant (dominance validated condition) or greater prevalence and appeal. to describe their typical morning routine (control). Then, they were in- In Studies 2A-C, we explored how car bumper stickers can be structed to read a passage about how snakeskin is procured for luxury consumed as a form of status pivoting. In Study 2A, we examined accessories. In the high (low) environmental cost condition we included cars at golf club parking lots in Crans-Montana, a luxurious resort information that snakes were killed for their skins (molted their skins). town in Switzerland. Analysis of the cars that had bumper stickers Next, participants rated a luxury snakeskin bag on how attractive, ap- (N=98) revealed that, compared to owners of luxury cars, owners pealing, and desirable the bag was (1 = not at all/ 9 = very much; α = of non-luxury cars were more motivated to signal status in alterna- .91) and completed the same status-striving scale used in Study 2 (East- tive domains using stickers associated with collectable experiences man et al. 1999). (e.g., extreme sports; c2=28.63, p<.001) or social causes (c2=48.17; In the control condition, where dominance had not been validated, p<.001) over stickers associated with the status threat domain – golf). we continue to find a significant and positive effect of environmental To test causality, in Study 2B, we surveyed 73 runners costs at high status-striving (JN point at +1.66 SD), and significant and (Mage=45.1, 55.4% males) in the Boston Marathon pre-race exposi- negative effect of environmental costs at low status-striving (JN point tion. Runners who described someone they knew with a very expen- at -.45 SD). In the high environmental cost condition, the effect of the sive car (status threat) were more interested in displaying the “26.2” dominance manipulation on product appeal was negative and signifi- sticker on their car, laptop, or at work (1-7 scale; α=.56) (M=4.51) cant at high levels of status-striving (JN point at +.42 SD). than runners who did not describe such an individual (control) In sum, we contend that certain consumers (those striving for sta- (M=3.65; F(1,72)=13.84, p=.045). tus) are drawn to status products because they are destructive to envi- Study 2C manipulated status threat in a controlled setting and ronmental resources. Consuming these products enables consumers to tested the mediating role of motivated trade-off beliefs. Car owners access an object that represents an additional path to high social status: (N=179; Mage= 37.9; 47.5% female; MTurk) imagined that one of dominance. Thus, we find support for our prediction that objects that their coworkers had won a large amount of money ($100,000; status embody both signals (prestige and dominance) are a stronger represen- threat) vs. a smaller amount of money ($1,000; no threat) in the lot- tation of status and, as such, are particularly appealing to status-striving tery. Participants were more interested in putting stickers that display consumers. alternative domains on their cars in the threat (vs. no threat) condi- Advances in Consumer Research (Volume 48) / 1009 tion (p=.030). The effect was mediated by motivated trade-off beliefs unexamined in this research agenda. While this may be partially due (that financial success leads to problems in alternative life domains; to identification challenges that necessitate linking high-quality and a×b=.8421, 95% CI=[.4947,1.2206]). precisely located geographic data to outcome variables of interest, The following studies tested the effect of status threat on choice recent literature and labor trends have pointed to other concerns di- between status pivoting to an alternative domain versus status resto- minishing the importance of this group. ration within the threat domain. In his influential book, Bowling Alone, Putnam (2000) argues

In Study 3, participants (N=203, Mage=35.9, 41% female, that Americans are largely disconnected from their neighbors and lo- MTurk) imagined they met a former classmate at a high-school re- cal communities. Recent trends also find that Americans are increas- union. In the status threat (vs. no threat) condition, the classmate was ingly spending more time working and online (O’Boyle and Harter, described as more (vs. less) professionally successful than them. Af- 2013; Perrin, 2015). Yet there are clear reasons that neighbors should terwards, participants chose which mug they would like to win a real not be overlooked as a consequential source of influence, particularly raffle. Participants were more likely to choose a mug that highlighted in household mortgage decisions. Americans live in their homes for an alternative domain (“Keep Calm and Be a Friend”) over a mug an average of thirteen years, more than twice the average workplace that highlighted the domain of the status threat (“Keep Calm and tenure, and are living in their homes increasingly longer (BLS, 2019; Back to Work”) in the status threat (vs. no threat) condition (c2=4.51, Modestino and Dennett, 2013). And, finally, while people are work- p=.024). ing longer hours, more and more of these hours are spent working Studies 4A-B tested a boundary condition of status pivot- from home (BLS, 2017). Consequently, do social interactions from ing – attainability of high status in the domain of the threat. Study our neighbors have a meaningful impact on households’ economic

4A (N=355; Mage =34.7; 36.2% female; MTurk) showed that sub- decisions? If so, by what mechanism? Do they improve decision- jectively perceiving the success of the higher status person in the making or lead to worse outcomes? And in particular, does living in domain of the threat as attainable can attenuate the effect of status a relatively more expensive home lead you to be more or less likely threat on status pivoting and lead consumers to signal status in the to be socially influenced that those in relatively cheaper homes? 2 threat domain (c (3)=10.00, p=.019). Study 4B (N=236; Mage=48.6; We build a comprehensive panel dataset that details the own- males; Qualtrics panelists) revealed that consumers’ actual ability to ers, outstanding mortgages, and characteristics of every property in attain high status within the threat domain (i.e. income) attenuates Los Angeles between 1992 and 2012. Crucially, we know the precise the effect of status threat on status pivoting, as status threat (vs. no latitude and longitude of each property and can therefore map house- threat) increased status pivoting only at annual income levels below holds and their hyperlocal neighbors. With this fine level of detail,

$107,768 (bjn=.4057, p=.05). we can isolate and identify social influence effects occurring at a Study 5 examined an alternative possibility that pivoting could hyperlocal level. This level of precision affords the ability to identify occur because people merely want to escape the status threat. It also hyperlocal sources of influence in a way that is not possible using further tested the process (trade-off beliefs) by manipulating the sa- zip code-level or county-level data alone, where endogenous group lience of trade-offs across domains. 180 women (Mage=35, MTurk) formation concerns prevail. Our empirical strategy tests whether de- imagined that their friend enrolled in a professional course. She ei- cisions made by a household’s block peers, the sixteen neighbors on ther felt disadvantaged because her classmates were more accom- the household’s residential block, exert a significant influence on its plished (status threat), felt that she was doing well (no threat), or felt mortgage decisions even when controlling for the decisions made by disadvantaged compared to her classmates who were more accom- the household’s larger set of neighborhood peers, the 130 neighbors plished but did not have rewarding relationships (trade-off between on the household’s census block and those blocks adjacent to it, and domains highlighted). Participants evaluated bracelet charms to gift controlling for borrower and lender-specific fixed effects as well as to their friend. Participants were more likely to prefer charms associ- time quarter-by neighborhood level fixed effects. ated with an alternative domain (relationships) in the status threat We find households make refinancing and loan-type decisions and the trade-off salient conditions than in the no-threat condition especially like those of their hyperlocal, block-level neighbors, even (p=.028; p=.011), but their preferences for charms associated with after controlling for the decisions of those neighbors who live just the threat domain (p>.11) and neutral domains (p>.491) did not differ slightly farther away, on adjacent blocks. Specifically, we estimate across conditions. that households are 4% more likely to refinance and 2% more likely Our work adds to theory on status-signaling and compensatory to choose an adjustable rate mortgage (ARM) if the count of block consumption by enhancing the understanding of status pivoting, and peers who have recently refinanced or the share of their block peers it contributes to prior work on lay theories about cost-benefit heu- with an ARM for their last loan increases from the 10th percentile ristics by demonstrating that beliefs about trade-offs may be pro- block to the 90th percentile block, respectively. We further find that nounced under status threat. households making purchase loan decisions, who have not lived in the neighborhood before and therefore had no opportunity to socially ‘I’ll Have What She’s Having:’ Neighbors Socially interact with their new neighbors, show no signs of being socially Influence Households’ Mortgage Decisions influenced. However, though not initially socially influenced, we show that movers who later refinance become socially influenced EXTENDED ABSTRACT over time. A mortgage is the largest and most consequential financial Unlike real estate listings, adding solar panels (Bollinger and contract that the typical household makes. For most families, their Gillingham, 2012), and foreclosure (Gupta, 2018), refinancing and house is their largest asset and their mortgage their largest liability. ARM choice are not decisions typically visible at the hyperlocal Understanding the drivers of household mortgage choices therefore level, ruling out the possibility of a visual salience mechanism. In- remains critical to the study of consumer behavior and household stead, these hyperlocal social influence effects are more consistent finance. A growing body of work shows that decision-making and with a word-of-mouth mechanism. To provide further support of a risk-taking behavior can be influenced by various contextual factors, word-of-mouth transmission mechanism, we investigate whether including social influence. However, neighbors have been largely hyperlocal peer effects change as the likelihood of social - interac 1010 / Splash Out or Cheap Out? Motivational Influences on Consumers’ Management of Tradeoffs tions among neighbors increases. To do so, we estimate our model Barkow, Jerome H. Darwin, sex, and status: Biological Approaches across areas with varying walkability scores. Walkability scores to Mind and Culture. University of Toronto Press, 1989. measure accessibility to public transportation, the presence and qual- Baumeister, Roy F., Kathleen D. Vohs, Jennifer L. Aaker, and ity of footpaths, and traffic and road conditions, among other factors, Emily N. Garbinsky (2013), “Some Key Differences between therefore serving as a proxy for the likelihood that households living a Happy Life and a Meaningful Life,” Journal of Positive in that zip code interact with their neighbors, particularly those be- Psychology, 8 (6), 505-16. yond their specific block. We find that the estimated effect of block Belk, Russell (1988), Possessions and self. John Wiley & Sons, peer influence is steady and similar across all ten deciles of walkabil- Ltd. ity score. Social interactions between block neighbors are no more Bian, Qin, and Sandra Forsythe (2012), “Purchase intention for likely to occur as areas become more walkable. This is unsurprising Luxury Brands: A Cross Cultural Comparison.” Journal of since interactions between block-neighbors are already almost a cer- Business Research, 65, (10), 1443-51. tainty. Getting mail, taking out the trash, and shoveling snow all hap- Chase, Ivan D., Craig Tovey, Debra Spangler-Martin, and pen regardless of the area’s walkability. Importantly, this also means Michael Manfredonia (2002), “Individual Differences versus that block-level social interaction effects are always in play, regard- Social Dynamics in the Formation of Animal Dominance less of neighborhood attributes correlated with walkability. We also Hierarchies.” Proceedings of the National Academy of find that the estimated effect of adjacent-block peers’ decisions do Sciences, 99 (8), 5744-49. increase as a function of increased walkability. While we cannot at- Cheng, Joey T., and Jessica L. Tracy (2014), “Toward a Unified tribute all of this increase to an increase in social interactions as there Science of Hierarchy: Dominance and Prestige are Two are other correlated unobservable forces, the notion that more walk- Fundamental Pathways to Human Social Rank,” In The able neighborhoods increase the neighborhood-level social influence Psychology of Social Status, pp. 3-27. Springer, New York, effect is consistent, once again, with a word-of-mouth, social interac- NY. tion mechanism. Gao, L., Wheeler, S.C., & Shiv, B. (2009). The “Shaken Self”: Finally, we use our data to speak to the welfare implications of Product Choices as a Means of Restoring Self-View hyperlocal neighbor social interactions. First, we examine whether Confidence. Journal of Consumer Research, 36(1), 29-38. households are better/worse off from social interaction by examin- Henrich, Joseph and Francisco J. Gil-White (2001), “The Evolution ing optimal refinancing. We do this by determining what the current of Prestige: Freely Conferred Deference as a Mechanism for prevailing interest rate to the interest rate prevailing when the loan Enhancing the Benefits of Cultural Transmission,” Evolution was originated and measure how long it has been since the house and Human Behavior, 22 (3), 165-96. has refinanced. By looking just at the extremes, we can say what a Hirschman, Elizabeth C., and Morris B. Holbrook (1982), given household “ought” or “ought not” to do (Keys et al., 2016). We “Hedonic Consumption: Emerging Concepts, Methods and find that the effects of hyperlocal social influence are significantly Propositions,” Journal of Marketing, 46, 92-101. stronger for households who ought to refinance compared to house- Maeng, Ahreum and Pankaj Aggarwal (2017), “Facing Dominance: holds who ought not to. This indicates that interacting with neigh- Anthropomorphism and the Effect of Product Face Ratio on bors might encourage learning whereby households who ought to be Consumer Preferences,” Journal of Consumer Research, 44 making a certain choice are more likely to do so if their block peers (5), 1104-21. have recently made that same decision. Second, we examine whether Mandel, N., Rucker, D.D., Levav, J., & Galinsky, A.D. (2017). there is heterogeneity in the ‘type’ of households who are more likely The Compensatory Consumer Behavior Model: How to be socially influenced. We find that relatively larger households on Self‐Discrepancies Drive Consumer Behavior. Journal of a block (as determined by estimated value) are, in fact, more likely Consumer Psychology, 27(1), 133-146. to exhibit social influence effects while households that are relatively Percival Carter, Erin and Lawrence Williams (2017), “Meaningful cheaper on a block are less likely to show these effects. This implies Consumption Provides Long Lasting Benefits at a High Cost,” that households who are relatively wealthier in terms of visual sa- in NA – Advances in Consumer Research, Vol. 45, ed. Ayelet lience are more likely to show hyperlocal social influence effects in Gneezy, Vladas Griskevicius, and Patti Williams, Duluth, MN: their refinance decisions relative to those households whose visual Association for Consumer Research, 63-67. display of wealth is comparatively less. Ryan, Richard M. and Edward L. Deci (2001), “On Happiness and In summary, we demonstrate hyperlocal social interactions Human Potentials: A Review of Research on Hedonic and may be an important mechanism for information flow in household Eudaimonic Well-Being,” Annual Review of Psychology, 52, finance and mortgage markets, spurring households to learn about 141-66. prevailing interest rates and new mortgage products, leading them Sivanathan, Niro, and Nathan C. Pettit (2010), “Protecting the towards “better” decisions. Self Through Consumption: Status Goods as Affirmational Commodities.” Journal of Experimental Social Psychology, REFERENCES 46, (3), 564-70. Alba, Joseph W., and Elanor F. Williams (2013), “Pleasure Principles: A Review of Research on Hedonic Consumption,” Journal of Consumer Psychology, 23 (1), 2–18. Consumer Interactions with Voice Technology Chairs: Kurt Munz, Bocconi University, Italy Johann Melzner, New York University, USA

Paper #1: Verba Volant Scripta Manent: Communication can elicit country of origin associations that boost evaluations of Modality Affects Privacy Expectations products from that country. Johann Melzner, New York University, USA Finally, Munz and Morwitz explore the cognitive consequences Andrea Bonezzi, New York University, USA of shopping by voice. Voice shoppers have greater difficulty mak- Tom Meyvis, New York University, USA ing comparisons, leading them to choose recommended items more often, but also making them more likely to choose nothing. They Paper #2: Experiential Effects of Device-Dependent Voice rely more heavily on information they can understand apart from a Interactions: Resolving the Power Imbalance between Voice- comparative context. assistants and Users This session should be of interest to a broad audience of re- Ana Valenzuela, Baruch College, USA and Ramon Llull searchers with substantive interests in voice technology, human- Esade, Spain technology communication as well as a theoretical interest in norma- David Luna, University of Central Florida, USA tive beliefs, associative processes, and information processing. The Christian Hildebrand, University of St. Gallen, Switzerland consequential nature of the dependent variables provides insights for Jessie Du, Baruch College, USA practitioners. Paper #3: Is That an Accent I Hear? How a Digital Voice Assistant’s Accent Affects Consumer Perceptions and Intentions Verba Volant Scripta Manent: Communication Modality Rebecca Chae, Santa Clara University, USA Affects Privacy Expectations James A. Mourey, DePaul University, USA Carolyn Yoon, University of Michigan, USA EXTENDED ABSTRACT Paper #4: Not-so Easy Listening: Roots and Repercussions of Consumers increasingly interact with technology by speaking Auditory Choice Difficulty in Voice Commerce instead of typing. This voice technology boom is predicted to affect Kurt P. Munz, Bocconi University, Italy marketing on many levels (Dawar 2018). In this research, we fo- Vicki G. Morwitz, Columbia University, USA cus on targeted advertising. Specifically, we propose that oral versus written communication increases consumers’ privacy expectations SESSION OVERVIEW with negative effects for targeted advertising. Voice-enabled technologies enjoy widespread popularity and Privacy expectations are social norms about how information are being adopted at a staggering rate. Sixty-six million US adults should flow within and outside of the contextual boundaries of an own a smart speaker (Levin and Lowitz 2019), 49% of US smart- information exchange (Nissenbaum 2009). Oral communication is phone users use voice-technology at least once a week (Statista ephemeral and transient and thus tends to be contextually bounded. 2018), and 20% of mobile Google searches are done by voice (Ster- That is, spoken content typically remains within the social (i.e., with ling 2016). Some industry experts predict that in the coming decade the audience involved directly in the communication), spatial (i.e., in 30% of online browsing will be done entirely without screens (Gart- the same location), and temporal (i.e., no record) boundaries of the ner 2016), and it cannot be disputed that voice-enabled technologies information exchange setting (Jahandarie 1999). Written communi- are becoming a part of consumers’ daily lives with myriad marketing cation, however, is a contextually impoverished modality, as content implications (Dawar 2018). can easily be saved and transmitted to third-parties at remote place As we are entering the era of voice (Campos 2017), this ses- at later times (Sproull and Kiesler 1986). Thus, consumers may have sion combines four papers that shed light on various dimensions on internalized the norm that orally versus written disclosed informa- consumer interactions with voice technology with important down- tion should not be accessible to parties not involved in the immediate stream consequences for advertising effectiveness, product prefer- information exchange and not be transmitted to other places or saved ences, and choice. and later used. Even though, ephemerality of the spoken word does The first paper investigates differential privacy perceptions in not hold in online communications with technology (as content is re- online interactions with technology as a function of communication corded, sent to remote servers, and often disseminated; e.g., Acquisti, modality. Specifically, Melzner, Bonezzi, and Meyvis find that con- Brandimarte, and Loewenstein 2015), the physical ephemerality of sumers expect oral versus written interactions with technology to be consumers’ spoken output is highly salient to consumers, while the more private with negative consequences for the receptivity of tar- digital translation of the disclosed information is arguably not. Fur- geted advertising based on orally as compared to written disclosed thermore, consumers tend to apply well-established offline norms to information. interactions with technology online (Moon 2000) even though they The second paper by Valenzuela, Luna, Hildebrand, and Du ex- do not apply in this context. plores the use of imperative language while interacting with voice as- In four studies, we show that consumers display increased sistants. Experienced voice assistant users adapt their speech toward privacy expectations in oral versus written communication with greater use of commands, while inexperienced users feel uncomfort- technology online. Furthermore, consumers perceive contextual able doing so, making them less likely to seek status products. boundary breaches as higher privacy violation in oral versus writ- Next, Chae, Mourey, and Yoon show that a subtle cue, such as ten communication with negative downstream consequences for the a programmed accent in the voice of an artificial intelligence agent effectiveness of targeted ads, which necessitate contextual boundary can influence willingness to pay and purchase intention–an effect not breaches of provided information. For brevity, we report two studies intuited by consumers. That is, accents in smart technology voices in this report.

Study 1: Undergraduate students were each seated in private Results. Participants indicated marginally higher privacy vio- rooms and presented with an online survey on a topic of relevance lations in the oral than written condition. Conversely, participants to them, namely dating. Participants were assigned to one of three indicated lower ad effectiveness in the oral than written condition. conditions. In the oral (written) condition, participants heard (read) However, the indirect effect of oral versus written condition via pri- the questions recorded with a female Google Assistant voice and re- vacy violation on ad effectiveness was only marginally significant. sponded by speaking into a built-in laptop-microphone (typing). The Discussion. The results provide first evidence in an ecologically oral transcription condition closely resembled the oral condition, valid context that targeted ads based on oral versus written informa- with the exception that participants were specifically notified that no tion from consumers are received less favorably due to increased audio files of their recordings are saved and that their responses will consumer privacy expectations in oral communication. be automatically transcribed (participants in the oral condition received no further information even though no voice files were saved Experiential Effects of Device-Dependent Voice either). Interactions: Resolving the Power Imbalance between Measures. As measures of privacy expectations, participants Voice-assistants and Users were presented with the following five statements and asked to indicate the extent they would find social (three items: the research- EXTENDED ABSTRACT ers get access to their answers, the researchers give others access to Amazon’s Echo (a voice-activated device) is a smart home their answers, a third-party company gets access to their answers), speaker embedded with artificial intelligence. DubbedAlexa , Ama- temporal (their answers are stored indefinitely), and spatial (their zon’s Artificial Intelligence has gained popularity in many house- answers are transmitted to remote servers) boundary breaches ac- holds. The U.S. installed base of smart speakers has hit 66 million ceptable. Since all items show highly similar results and load on a in 2018 (Levin and Lowitz 2019) with 252 million U.S. adults using single factor, they will be reported as a single composite measure voice-assistants in their homes or cars (Voicebot 2019). Alexa only of privacy expectations for brevity. Next participants indicated the responds when summoned, and consumers usually talk to her using degree to which they felt that their answers made them identifiable imperative language. In this paper, we investigate how the consum- as individual. er’s experience originating from device-dependent voice interactions Results. Conditions differed in acceptability of contextual may have an effect on downstream decisions. boundary breaches, such that participants in the oral condition in- The chameleon effect is a specific form of synchrony that takes dicated lower acceptability (and thus higher privacy expectations) place among dyads rather than within and between groups (for a re- than participants in both the oral transcription than written condition) view, see Chartrand & Lakin 2013). As adults, we engage in mim- with no difference between the latter two conditions. Additionally, icry to establish and maintain social relations with others (Chartrand conditions differed in perceived identifiability, with the oral condi- & Bargh 1999). In linguistics, accommodation theory investigates tion indicating higher identifiability than the oral transcription and the mechanism underlying verbal mimicry (Gallois, Ogay, and Giles marginally higher identifiability than the written condition with no 2005; Koslow, Shamdasani, and Touchstone 1994). Convergence, difference between the latter two conditions. However, identifiability within linguistic accommodation, explores the motivation for verbal did not mediate the effect of oral versus oral transcription and written imitation. Convergence arises when speakers alter their linguistic condition on acceptability. patterns to adopt styles more like that of their interaction partners Discussion. The results support the hypothesis that consumers (Wisniewski, Mantell and Pfordresher 2013). find boundary breaches less acceptable in oral versus written com- In this paper, we focus on the dynamics of language accom- munication. The effect does not seem to be a function of feeling more modation in the dyad “consumer - voice activated assistant” (assem- identifiable through speaking than writing. Furthermore, it is not the blage theory; Hoffman and Novak 2018). We posit that technology act of speaking itself that evokes increased privacy expectations but devices’ inherent constraints make digital assistants less conversa- the ephemerality of spoken output as indicated by the fact that par- tional, and, consequently they potentially limit consumers’ use of ticipants in the oral transcription condition displayed privacy expec- language. Specifically, in their working paper, Hildebrand, Hoffman tations similar to the written condition. and Novak (2019) link phonetic signals in the human voice when Study 2: Undergraduate students were seated in private rooms interacting with voice-activated assistants with consumer experi- and asked to imagine that they had to Google things to do for a week- ences, and differentiate between command-based (“Alexa, length of end-trip (amongst which was eating at a restaurant). Participants marathon”) and request-based (“Alexa, can you tell me the length of were assigned to one of two conditions (oral and written) and entered a marathon?”) voice inputs. search terms either by typing or by speaking into the built-in laptop- Our research question centers on the idea that differential ex- microphone using the Google voice functionality. The keyboard was periences of the interaction with a voice interface may have down- disabled in the latter condition. Afterwards, participants completed stream consequences. Consumers may experience this interaction unrelated studies for about 30 minutes. On a different laptop, partici- differently depending on whether they have already started their ac- pants were informed that the search terms they entered previously commodation process towards talking “with a Machine”. Consumers were used to generate a personalized ad for them and were then pre- that already use voice-activated devices would be accommodated to sented with an ad for a restaurant. using Commands instead of Requests, while the opposite would be Measures. Upon seeing the ad, participants were asked to indi- true for those that have never used such devices. This differential cate the degree to which they felt like their privacy has been violated. extent of accommodation may create friction in their experience, They further indicated the degree to which they were interested in which may lead to compensatory processes to resolving the power visiting the website of and eating at the advertised restaurant. These imbalance between the voice-assistant and the user. In this paper, items were averaged into measure of ad effectiveness. Additionally, we examine the impact of device-dependent voice Interactions on they were asked how identifiable they felt based on the search terms consequential marketing outcomes like the evaluation of “status”- they provided and how good a match the ad was for them. related cues. A set of three studies tests these ideas. Advances in Consumer Research (Volume 48) / 1013

In the first study (N= 114), participants worked in individual lab assistants, which may lead to compensatory behavior based on the rooms with an Amazon Echo dubbed “Computer.” Participants were consumers’ conversational experience. given a 3-minute practice run. At that point, participants were given 8 tasks to fulfill with their Echo and told they would get rewarded Is That an Accent I Hear? by their performance. We then measured their level of language ac- How a Digital Voice Assistant’s Accent Affects Consumer commodation (“I changed the way I spoke to make it easier for X Perceptions and Intentions to understand me”; 0-100) and their success communicating (1-7). Individual differences in their experience (yes/no) with Alexa and EXTENDED ABSTRACT with voice-activated phone systems were recorded. Is a Chinese horoscope more likely to come true if the fortune Study 1: We identify a significant two-way interaction between teller speaks with a Mandarin accent? The country-of-origin (COO) whether participants had prior experience with Alexa or with other literature (Hong and Wyer 1989, 1990) suggests it is possible to voice-operated systems in the extent of language accommodation produce positive market outcomes by leveraging existing associa- (F (1,113)=3.87, p <0.05) and communication effectiveness (F tions between cultures and products, such as the French and wine. (1,113)=6.73, p<.01). Participants accommodated less and experi- However, when most digital devices–like Amazon’s Alexa–feature enced less effectiveness in their communication when they had not standard voices in a user’s native language, would adding accents to particular responses help or hurt user perceptions? Two studies sug- used Alexa or Siri before (Macommod;success= 45; 5.7) or had used Siri but gest that a native accent can dominate, but that deliberately adding not Alexa (Macommod;success = 35; 6.2), compared with when they had accents to particular responses can affect user-related outcomes via already used Alexa and other voice-activated systems (Macommod;success = 55; 6.6). perceptions of fit and competence. Study 2 (N= 159) and 3 (N= 80) followed a similar procedure To date, the majority of smart devices feature female voices as study 1. Participants were given 8 tasks to fulfill. However, when speaking a user’s native language without any regional accent (Kin- interacting with the voice-activated assistant (either just repeating sella 2019; West, Kraut, and Chew 2019). For some devices, con- the sentence to the screen - study 2 - or interacting with the actual sumers are able to customize the specific language and gender of a Echo dubbed “Computer” - study 3), participants were directed to device’s voice, but regional accents, foreign accents, and dialects are use either Commands or Requests (e.g. “Alex, Turn on the TV” in- not yet customizable features (Fingas 2018). For English-speaking stead of “Alex, Can you turn on the TV?”). After that, they answered consumers, the fact that English tends to be perceived as the “inter- a set of questions about language accommodation, communication national” language and even as the expected voice for companies en- effectiveness, and prior experience with Alexa and other voice-acti- gaged in global business may explain why English is the default lan- vated devices. In study 2, they also rated whether their experience guage setting for most smart devices (Krishna and Ahluwalia 2008). was that of a 1 = Master; 7 = Servant. In study 3, we included some When it comes to smart devices, English is the expected voice. downstream DVs: a) Attitude towards status-cuing ads (“A symbol However, English as the default language ignores the vast lit- of status. The Office Pen is designed to attract attention with its so- erature on the COO effect in which consumers link positive associa- phisticated design”) vs. a non-status driven ad (“A symbol of quality. tions between particular products and their source country (Bilkey The Office Pen is designed to perform consistency with its extraordi- and Nes 1982; Erickson, Johansson, and Chao 1984; Han 1989, Has- nary quality“; 7-point scale: “I would like/share this ad”); b) Evalu- tak and Hong 1991). These associations typically result in greater ation of status-cuing products (WTP (10% to 120% of retail price) liking, purchase intentions, and satisfaction for the product, such for status (e.g. briefcase, silk tie) vs. non-status (e.g. sofa, washing as paying more for pasta from Italy (Hong and Wyer 1989, 1990; machine)). Johansson, Douglas, and Nonaka 1985). Once a particular COO as- Study 2: We identify a significant two-way interaction between sociation has been established, deviations or incongruities with that whether participants had prior experience with Alexa and whether association can produce negative outcomes that could have been they were using Commands/Requests in the extent in which they felt avoided by omitting any COO reference in the first place (Melnyk, like a Servant (F (1,158) =3.49, p<0.05), their language accommo- Klein, and Volckner 2012). dation (F (1,158) =8.66, p<0.01) and communication effectiveness That English is generally considered the default “voice” for (F (1,113)=3.41, p<.07). With prior Alexa experience, participants global business poses an interesting question with respect to smart device voices: what happens when a smart device response, provided felt more like a Master using Commands (MCommand= 5 vs. MRequest= in English, is delivered with a foreign accent? The goal of the current 3.8), while those without experience using Requests (MCommand= 4.2 research is to explore the extent to which an accent for a smart de- vs. MRequest= 4.6). When using Commands, those with experience ac- vice’s voice predictably affects consumer outcomes. We predict that commodated much less than those without experience (MAlexa_experi- the standard, American English accent--recognized as the default, ence= 35vs. MAlexa_no_experience= 62). Study 3: We identify a significant two-way interaction between global business language and accent--is likely to yield positive out- whether participants had prior experience with Alexa and were using comes given its ubiquitousness and familiarity. However, we argue Commands/Requests in their relative evaluation of status items (F that English spoken with particular foreign accents--like Mandarin (1,80) =4.09, p<0.05) and marginal for attitude towards status claims or French--can help or hinder consumer outcomes depending on (F (1,80) =2.40, p<.10). Without prior Alexa experience, participants context. Specifically, when a positive COO association exists, the exhibited lower relative evaluations towards status items when using relevant accent should bolster consumer outcomes; when no such association exists, the accent should harm consumer outcomes. Commands (MCommand= .85 vs. MRequest= 1.15), while those with expe- A pilot study tested people’s beliefs about our predictions. Par- rience when using Requests (MCommand= 1 vs. MRequest= .90). Taken together, device-dependent voice Interactions seem to ticipants (N =198) rated whether they thought accents could affect potentially have consequential marketing outcomes like the evalu- their perceptions and behaviors. Participants were significantly like- ation of “status”-related items. This results could be driven by the ly to say that accent, gender, and vocal age would certainly not affect pattern of language accommodation when talking to voice-activated their perceptions of truth, product quality, or behavioral intentions. 1014 / Consumer Interactions with Voice Technology

We next tested how accents actually influence people’s perceptions area has boomed since voice assistants like Amazon Alexa opened and behavioral intentions. a new shopping channel. Yet research into auditory consumer be- In Study 1, participants (N = 183) heard a vocal instruction that havior is relatively scarce (Krishna 2010; Meyers-Levy, Bublitz, was created using Amazon’s Polly Artificial Intelligence. A smart and Peracchio 2010). While there are undoubtedly aspects to shop- home device gave recipe instructions for grilled cheese or French ping by voice that make it more like a human social interaction (e.g. toast in a voice with an American or French accent. As predicted, Schroeder and Epley 2016), we compare text presented visually (in the interaction of recipe and accent was significant on participants’ writing) to the same text spoken aloud and explore the underlying recommendations (p = .02). When the AI had an American accent, cognitive processes. French toast (M = 4.79) versus grilled cheese (M = 5.23) recipes did In the context of persuasive messaging, some studies have di- not influence recommendations (p = .12). However, speaking with a rectly compared spoken words to their equivalent written form, as French accent for French toast (vs. grilled cheese) increased recom- we do in this research (e.g. Chaiken and Eagly 1976, 1983; Sparrow mendations (M = 4.53 vs. M = 3.99; p = .05). The interaction effect 2007; Unnava, Burnkrant, and Erevelles 1994). However, we focus was also significant on the perceived fit of voice (p < .001). While on the novel circumstance of choices made from options presented a French accent increased the perceived fit with French toast more by voice, suggesting that this context may be particularly challeng- than with grilled cheese (M = 4.51 vs. M = 3.21; p < .001), an Ameri- ing for consumers. can accent increased the perceived fit with grilled cheese more than It is not necessarily the case that spoken information is always with French toast (M = 5.62 vs. M = 4.69; p = .001). Importantly, the harder to process. Indeed, processing auditory information is gener- perceived fit was a significant underlying mechanism for a French, ally easier than reading text (Liberman 1989)prior to all others: why not American, accent (95% CI for index of moderated mediation = is it easier to perceive speech than to read, and why is it easier to [.83, 1.78]). Only when the AI spoke with a French accent, providing speak a word than to spell it? My aim is to repair these omissions. To the French (vs. American) recipe increased the perceived fit between that end, I divide my talk into two parts. First, I say why we should product and voice and subsequently increased participants’ recom- consider that the greater ease of perceiving and producing speech is mendations. paradoxical, by which I mean to suggest that the reasons are not to Study 2 further examined the underlying processes in a new be found among surface appearances. Then I propose how, by going context. Participants (N = 193) heard an audio clip of a Chinese or beneath the surface, we can find the reasons, and so resolve the para- Zodiac horoscope read in a voice with either a Chinese accent or an dox. THE PARADOX Before developing the paradox. I should first American accent. They rated the extent to which they believe the remind you that perceiving and produCing speech are easier than horoscope would come true, the amount they are willing to pay for reading or writing. for thiS is the point from which I depart and to a 1-year subscription, the perceived fit of voice, and the perceived which I will. at the end. return. The relevant facts include the follow- competence of the AI. As predicted, horoscope and accent resulted ing. (1, a sentiment anyone who has hesitated to read subtitles in a in a significant interactive pattern that was similar to Study 1, on all foreign-language film can appreciate. For proficient readers howev- of the measures (ps < .04). A significant moderated serial media- er, differences in processing difficulty between text and speech may tion showed that only when the AI had a Chinese accent, reading the be minimal, with comprehension of written text (books) and auditory Chinese (vs. Zodiac) horoscope increased the perceived fit, which text (audiobooks) being generally similar (Gernsbacher, Varner, and enhanced the perceived competence and subsequently the level of Faust 1990). However, for particularly difficult passages, reading belief and willing to pay. can be easier than listening due to an ability to re-read (Daniel and Despite the fact that individuals do not think a smart device’s Woody 2010). accent can affect their perceptions and behavioral intentions, two We propose that voice shopping is challenging due to difficulty experiments provide initial evidence to the contrary. Although an making auditory comparisons. Voice comparisons require a greater American English accent consistently yielded positive effects on reliance on memory. Spoken words persist only in the instant of their consumer outcomes and intentions, ethnic accents bolstered con- utterance while typeset words can be re-visited with ease. A rich lit- sumer outcomes in COO-relevant contexts and diminished those erature in consumer behavior has demonstrated that holding informa- outcomes in non-COO-relevant contexts. Further, the better the per- tion in memory can interfere with processing new information due ceived fit of the voice, the higher the perceived competence of the to increased cognitive load (Shiv and Fedorikhin 1999)compared to smart device. This better fit and greater competence had important a second (fruit salad. Limitations on memory may also enhance the consequences on consumer outcomes that users did not think were likelihood of ignoring some of the information (e.g. Tversky 1972) possible. Together, results suggest that simply altering the accent of an aspect is selected (with probability proportional to its weight, and a smart device’s voice can have interesting consumer implications. consumers might rely on summary information or on important attributes when choosing, rather than integrate all available information. Not-so Easy Listening: We predict that these processes will lead auditory consumers to Roots and Repercussions of Auditory Choice Difficulty in have greater difficulty processing choice options compared to visual Voice Commerce consumers with three consequences. First, auditory difficulty may make accepting a recommendation more likely, as difficulty while EXTENDED ABSTRACT choosing can increase acceptance of advice (Gino and Moore 2007). Will shopping by voice resonate with consumers or will the Second, auditory consumers may be more likely to “defer” choice many upbeat forecasts (e.g. Hayllar and Coode 2018; Simms 2019) by choosing none of the available options, which also becomes more prove to be mostly chatter? We take an information-processing per- likely as choice difficulty increases (Dhar 1997) or even as the sub- spective to suggest that voice shopping can be cognitively challeng- jective feeling of choice difficulty increases (Novemsky et al. 2007). ing for consumers, particularly when they need to make compari- Third, because cognitive load disrupts attitude formation (Cronley, sons. Mantel, and Kardes 2010), we predict that auditory consumers will While presenting information to consumers by voice is not new have less polarized attitudes toward items they evaluate in a com- (e.g. radio advertisements or spoken daily specials), interest in this parative context. Advances in Consumer Research (Volume 48) / 1015

Experiments 1a-1i confirmed the above hypotheses. In Experi- Bilkey, Warren J. and Eric Nes (1982), “Country-of-origin effects ment 1a, in-lab participants chose a lip balm as a free gift. One of on product evaluations,” Journal of International Business two options was recommended. Based on random assignment, half Studies, 13 (1), 89–100. the participants heard the options read aloud to them in the voice Campos, Ivan (2017), “The Era of Voice: From Keyboards to Vocal of “Alexa” and half read the same information. Participants chose Chords,” Medium (June 25), https://medium.com/slalom- one and were given it to keep. As predicted, a significantly greater technology/the-era-of-voice-from-keyboards-to-vocal-cords- percentage of auditory participants chose the recommended option 5f891705a5ea (χ2(1;N=74)=6.77,p=.009). Chaiken, Shelly and Alice H. Eagly (1976), “Communication In Experiment 1b (preregistered), participants faced a simi- Modality as a Determinant of Message Persuasiveness and lar (hypothetical) choice, but could also defer. Auditory par- Message Comprehensibility.,” Journal of Personality and ticipants were more likely to choose the recommended option Social Psychology, 34(4), 605–14. (χ2(1;N=603)=17.06,p<.001). They also deferred more often ——— (1983), “Communication Modality as a Determinant of χ2(1;N=603)=15.43,p<.001), as predicted. Attitudes toward both Persuasion: The Role of Communicator Salience.,” Journal of choices revealed less polarized attitudes among auditory consumers Personality and Social Psychology, 45(2), 241–56. (F(1,601)=19.24,p<.001), also as predicted. A processing difficulty Chartrand, Tanya L. and John A. Bargh (1999), “The Chameleon scale (“complex,” “complicated,” “easy to comprehend” (reverse); Effect: The Perception-behavior Link and Social Interaction,” Anand & Sternthal, 2006) mediated the effect of presentation mode Journal of Personality and Social Psychology, 76 (6), on choice. Additional measures suggested the effects were not due 893–910. differences in anthropomorphism or trust. Chartrand, Tanya L. and Jessica L. Lakin (2013), “The Antecedents Experiments 1c through 1i demonstrated robustness across and Consequences of Human Behavioral Mimicry,” Annual various stimuli and voices, controlling for order effects, and when Review of Psychology, 64, 285-308. the visual stimuli were presented sequentially or in an ephemeral Cronley, Maria L., Susan Powell Mantel, and Frank R. Kardes manner mimicking speech. (2010), “Effects of Accuracy Motivation and Need to Evaluate Experiments 2 through 4 provided evidence that the effects on Mode of Attitude Formation and Attitude–Behavior are specifically related to difficulty making comparisons in auditory Consistency,” Journal of Consumer Psychology, 20(3), mode. In Experiment 2, we crossed presentation mode (voice vs. 274–81. text) with a classic joint versus separate evaluation paradigm (Hsee Daniel, David B. and William Douglas Woody (2010), “They Hear, 1996), reasoning that preference reversals should be less likely by but Do Not Listen: Retention for Podcasted Material in a voice due to difficulty learning from context. As predicted, we rep- Classroom Context,” Teaching of Psychology, 37(3), 199–203. licated the classic preference reversal only in visual mode. Auditory Dawar, Niraj (2018), “Marketing in the Age of Alexa,” Harvard consumers’ evaluations were similar whether considering the op- Business Review,” (June), https://hbr.org/2018/05/marketing- tions jointly or each option separately. Experiment 3 (preregistered) in-the-age-of-alexa extended this idea to comparisons of items previously encountered. Dhar, Ravi (1997), “Consumer Preference for a No-Choice Consumers rated a single novel product less favorably when it was Option,” Journal of Consumer Research, 24(2), 215–31. described by voice and they compared to one they had encountered Erickson, Gary M., Johny K. Johansson, and Paul Chao (1984), previously, but evaluations were similar when consumers were not “Image Variables in Multi-attribute Product Evaluations: making comparisons. In experiment 4, manipulating the difficulty of Country-of-Origin Effects,”Journal of Consumer Research, 11 making comparisons eliminated the differences that naturally exist (2), 694–99. between the conditions and provided stronger causal evidence. Fingas, Jon (2018, July 19), “Voice Assistants still Have Problems Experiments 5 and 6 explored the role of memory. In experi- Understanding Strong Accents,” Engadget, https://www. ment 5, working memory was measured with a “digit span” task (re- engadget.com/2018/07/19/voice-assistant-problems- calling strings of numbers). As consumers’ working memory capac- understanding-accents/ ity improved, the differences in choice outcomes between auditory Gallois, Cindy, Tania Ogay and Howard Giles (2005), and visual consumers diminished. Experiments 6a and 6b used the “Communication Accommodation Theory: A Look Back insight that the effects are related to memory to design interventions and a Look Ahead,” in Theorizing about Intercultural useful for marketing managers. Negative effects of voice presenta- Communication, ed. W. Gundykunst, Thousand Oaks: Sage, tion can be mitigated by explicitly stating the comparative value of 121-148. an option, eliminating the need to make comparisons (Experiment Gartner (2016), “Top Strategic Predictions for 2017 and Beyond: 6a) or by describing the choice options in a by-feature (rather than Surviving the Storm Winds of Digital Disruption,” Gartner by-option) manner (Experiment 6b), which reduces the time and (October 14), https://www.gartner.com/binaries/content/ amount of information needed to be held in memory. assets/events/keywords/cio/ciode5/top_strategic_predictions_ This research represents one of the first explorations of voice fo_315910.pdf commerce and contributes to our understanding of auditory judg- Gernsbacher, Morton A., Kathleen R. Varner, and Mark E. Faust ment and decision making. (1990), “Investigating Differences in General Comprehension Skill.,” Journal of Experimental Psychology: Learning, REFERENCES Memory, and Cognition, 16(3), 430–45. Acquisti, Alessandro, Laura Brandimarte, and George Loewenstein Gino, Francesca and Don A. Moore (2007), “Effects of Task (2015), “Privacy and Human Behavior in the Age of Difficulty on Use of Advice,”Journal of Behavioral Decision Information,” Science, 347 (6221), 509-14. Making, 20(1), 21–35. Anand, Punam and Brian Sternthal (2006), “Ease of Message Han, C. Min (1989), “Country Image: Halo or Summary Processing as a Moderator of Repetition Effects in Construct?,” Journal of Marketing Research, 26 (2), 222–29. Advertising,” Journal of Marketing Research, 27(3), 345. 1016 / Consumer Interactions with Voice Technology

Hastak, Manoj and Sund-Tai Hong (1991), “Country-of-Origin Melnyk, Valentyna, Kristina Klein, and Franziska Völckner (2012), Effects on Product Quality Judgments: An Information “The Double-Edged Sword of Foreign Brand Names for Integration Perspective,” Psychology & Marketing, 8 (2), Companies from Emerging Countries,” Journal of Marketing, 129-143. 76 (6), 21–37. Hayllar, Will and Matt Coode (2018), The Talking Shop: The Rise Meyers-Levy, Joan, Melissa G. Bublitz, and Laura A. Peracchio of Voice Commerce, London. (2010), “The Sounds of the Marketplace: The Role of Audition Hoffman, Donna L. and Thomas P. Novak (2018), “Consumer and in Marketing,” in Sensory Marketing, ed. Aradhna Krishna, Object Experience in the Internet of Things: An Assemblage New York: Routledge, 137–56. Theory Approach”, Journal of Consumer Research, 44 (6), Moon, Youngme (2000). “Intimate Exchanges: Using Computers to 1178-1204. Elicit Self-Disclosure From Consumers,” Journal of Consumer Hong, Sung-Tai and Robert S. Wyer Jr. (1989), “Effects of Country- Research, 26 (4), 323-39. of-Origin and Product-Attribute Information on Product Nissenbaum, Helen (2009), Privacy in Context: Technology, Policy, Evaluation: An Information Processing Perspective,” Journal and the Integrity of Social Life. Stanford University Press. of Consumer Research, 16 (2), 175-87. Novemsky, Nathan, Ravi Dhar, Norbert Schwarz, and Itamar Hong, Sung-Tai and Robert S. Wyer Jr. (1990), “Determinants Simonson (2007), “Preference Fluency in Choice,” Journal of of Product Evaluation: Effects of the Time Interval between Marketing Research, 44(3), 347–56. Knowledge of a Products Country of Origin and Information Schroeder, Juliana and Nicholas Epley (2016), “Mistaking Minds about its Specific Attributes,” Journal of Consumer Research, and Machines: How Speech Affects Dehumanization and 17 (3), 277–88. Anthropomorphism,” Journal of Experimental Psychology: Hsee, Christopher K. (1996), “The Evaluability Hypothesis: General, 145(11), 1427–37. An Explanation for Preference Reversals between Joint Shiv, Baba and Alexander Fedorikhin (1999), “Heart and Mind in and Separate Evaluations of Alternatives,” Organizational Conflict: The Interplay of Affect and Cognition in Consumer Behavior and Human Decision Processes, 67(3), 247–57. Decision Making,” Journal of Consumer Research, 26(3), Jahandarie, Khosrow (1999), “Structural Differences between 278–92. Speech and Writing: The Linguistic Evidence,” in Spoken Simms, Kane (2019), “How Voice Assistants Could Change the and Written Discourse: A Multi-Disciplinary Perspective, K. Way We Shop,” Harvard Business Review, May. Jahandarie ed., Stamford, CT: Ablex, 131-51. Sparrow, Nick (2007), “Quality Issues in Online Research,” Johansson, Johny K., Susan P. Douglas, and Ikojiro Nonaka (1985), Journal of Advertising Research, 47(2), 178–82. “Assessing the Impact of Country of Origin on Product Sproull, Lee and Sara Kiesler (1986), “Reducing Social Context Evaluations: A New Methodological Perspective,” Journal of Cues: Electronic Mail in Organizational Communication,” Marketing Research, 22 (4), 388-96. Management Science, 32 (11), 1492– 1512. Kinsella, Bret (2019), “U.S. Consumers Do Express a Preference Statista (2018), “Frequency with which Smartphone Owners Use for Female Gendered Voice Assistants According to New Voice-Enabled Technology in the United States in 2017,” Research,” Voicebot.ai, https://voicebot.ai/2019/11/23/u-s- Statista (July 12), https://www.statista.com/statistics/787369/ consumers-do-express-a-preference-for-female-gendered- voice-technology-utilization-in-us/ voice-assistants-according-to-new-research/ Sterling, Greg (2016), “Google Says 20 Percent of Mobile Queries Kinsella, Bret (2019), “Twice the Number of U.S. Adults Have Are Done by Voice Searches,” Search Engine Land (May Tried In-Car Voice Assistants as Smart Speakers,” Voicebot.ai, 18), https://searchengineland.com/google-reveals-20 percent- https://voicebot.ai/2019/01/15/twice-the-number-of-u-s-adults- queries-voice-queries-249917 have-tried-in-car-voice-assistants-as-smart-speakers/ Tversky, Amos (1972), “Elimination by Aspects: A Theory of Krishna, Aradhna (2010), “An Introduction to Sensory Marketing,” Choice,” Psychological Review, 79(4), 281–99. in Sensory Marketing, ed. Aradhna Krishna, New York: Unnava, H. Rao, Robert E. Burnkrant, and Sunil Erevelles (1994), Routledge, 1–13. “Effects of Presentation Order and Communication Modality ——— and Rohini Ahluwalia (2008), “Language Choice on Recall and Attitude,” Journal of Consumer Research, 21(3), in Advertising to Bilinguals: Asymmetric Effects for 481. Multinationals versus Local Firms, Journal of Consumer West, Mark, Rebecca Kraut, and Han Ei Chew (2019), “I’d Blush Research, 35 (4), 692–705. if I Could: Closing Gender Divides in Digital Skills through Koslow, Scott, Prem N. Shamdasani and Ellen E. Touchstone Dducatio,” UNESCO and EQUALS Skills Coalition. (1994), “Exploring Language Effects in Ethnic Advertising: A Wisniewski, Matthew G., James T. Mantell, and Peter Q. Sociolinguistic Perspective,” Journal of Consumer Research, Pfordresher (2013), “Transfer Effects in the Vocal Imitation 20(4), 575-85. of Speech and Song,” Psychomusicology: Music, Mind, and Levin, Michael R. and Joshua N. Lowitz (2019), “Smart Speaker Brain, 23, 82–99. Market Takes Off in Holiday Quarter,” CIRP.com, https://files. constantcontact.com/150f9af2201/5c84cae5-7ff0-403d-acd5- 9065db944222.pdf Liberman, Alvin M. (1989), “Reading Is Hard Just Because Listening Is Easy,” in Wenner-Gren International Symposium Series: Brain and Reading, ed. C. von Euler, Hampshire, England: Macmillan, 197–205. Words and Beyond: Using Language to Better UnderstandConsumer Experiences and Goals Chair:Gizem Ceylan, University of Southern California, USA

Paper #1: Emotional Echo Chambers: Observed Emoji Clarify sual aspects of the experience. Internal factors (i.e., caring about vi- Individuals’ Emotions and Responses to Social Media Posts sual aspects) or external factors (i.e., choosing photos ahead of text) Sarah G. Moore, University of Alberta, Canada heighten visual focus in consumers’ writing and increase the likeli- Gopal Das, Indian Institute of Management Bangalore, India hood of sharing a photo. Anirban Mukhopadhyay, HKUST, Hong Kong Humphreys, Isaac, and Wang examine how people’s transactional (vs. informational) goals impact their word choice in online Paper #2: Is a Photo Worth a 1000 Words? The Usage of Images search queries. They find that people use more concrete words when and Text to Express Opinions in Online Reviews they are in a concrete mindset. Further, consumers respond more Gizem Ceylan, University of Southern California, USA favorably to search engine results and ad content whose language Kristin Diehl, University of Southern California, USA matches their construal level. Paper #3: Construal Matching in Online Search: Applying Text Yeomans et al . explores how words indicating receptiveness to Analysis to Illuminate the Consumer Decision Journey opposing opinions can affect readers’ perceptions, engagement, and Ashlee Humphreys, Northwestern University, USA willingness to cooperate. While using receptiveness words fosters Mathew S. Isaac, Seattle University, USA trust and cooperation, this effect is often underappreciated by writers Rebecca Jen-Hui Wang, Lehigh University, USA who do not anticipated the positive consequences of receptiveness. Paper #4: Conversational Receptiveness: Improving Taken together, these presentations identify the antecedents and Engagement with Opposing Views consequences of consumers’ word choices by adopting novel meth- Mike Yeomans, Imperial College, UK odologies in a variety of consumption settings. All papers are at an Julia Minson, Harvard Kennedy School, USA advanced stage, employing a range of methodologies and contexts. Hanne Collins, Harvard Business School, USA None of these papers has been previously presented at ACR. We ex- Frances Chen, University of British Columbia, Canada pect this session will generate significant interest among researchers Francesca Gino, Harvard Business School, USA studying word of mouth, online sharing, natural language processing, and social media. SESSION OVERVIEW Consumers produce written content for themselves or other Emotional Echo Chambers: Observed Emoji Clarify consumers every day. They write reviews, enter search queries, and Individuals’ Emotions and Responses to Social Media create information on websites like Wikipedia. Doing this, people Posts use words alone or in combination with visuals to express their opinions or goals. Past research shows that the words people use can re- EXTENDED ABSTRACT veal important aspects of their social and psychological worlds (Pen- Imagine an individual browsing on Facebook—a daily activity nebaker, Mehl, and Niederhoffer 2003). With advances in computer for more than 1.5 billion people (Facebook 2019). This individual technology, text analysis allows researchers to reliably and quickly sees a friend’s post expressing mixed emotions about their grand- gain insights into what people say. In this session, we examine how mother’s passing. She notes that others have already responded to the people choose the words they use as a function of their supplemen- post, some with a sad emoji, and some with a love emoji. How are her tary use of visuals and their goals, and examine novel consequences feelings and responses to the post shaped by others’ prior responses? of verbal-visual communication. Jointly, these presentations uncover We find that if others have responded to the post with predominantly insights about antecedents and consequences of consumers’ word sad (love) emoji, the individual in question—and subsequent oth- choices as they move through their decision journey, contribute to ers—will feel sadder (happier), will be more likely to respond to the discussion forums, and write recommendations for others. Papers in post, and will be more likely to respond with a sad (love) emoji. This this session ask: Do consumers express themselves differently when process may create “emotional echo chambers,” contributing to the they use emoji (presentation1) or pictures (presentation2) together polarization of opinion and emotion on social media (Toubiana and with text? Do consumers’ internal goals impact their word choice Zietsma 2017). when they create search queries (presentation3) or generate Wikipe- Emoji are icons used to convey emotion (e.g., ; Oxford Eng- dia entries (presentation4)? Jointly, these papers examine novel fac- lish Dictionary 2018), and are a form of textual paralanguage (TPL), tors that affect consumers’ word choices when using different ways which expresses audible, tactile, or visual elements of face-to-face to express themselves and their subsequent decision. communication in text (Luangrath, Peck, and Barger 2017). In 2016, Moore, Das, and Mukhopadhyay distinguish how people use in addition to the like emoji, Facebook began offering love, haha, emojis to complement or substitute text when responding to others’ wow, sad, and angry emoji reactions to posts. In the subsequent posts. They find when people respond to posts with mixed emotions, twelve months, these emoji were used 800 million times per day the emoji predominantly used by previous posters drives their own (Cohen 2017). Further, nearly half of Americans (48.9%) use emoji emoji choice and felt emotions. Further, the emotional contagion on social media or in text messages (Pilon 2015), and over half of from such an anchor emoji increases the likelihood of using both text Instagram posts contain emoji (Morrison 2017). and emoji when responding. Despite their popularity, it is unclear how use of—or exposure Ceylan and Diehl examine whether people use photos to com- to—emoji affects consumers. Thus, we examine how social media plement or substitute text when communicating their experiences. users are impacted by “anchor emoji”—others’ prior emoji responses They find that photos complement text the more people focus on vi- to a post. In doing so, we identify the potential for anchor emoji to

Advances in Consumer Research 1017 Volume 48, ©2020 1018 / Words and Beyond: Using Language to Better UnderstandConsumer Experiences and Goals create emotional echo chambers—a substantive issue for consum- the free choice condition who were exposed to a sad anchor emoji ers, firms, and policy makers. Further, we investigate emoji asa felt the most negative emotion, and were the most likely to select unique type of TPL, which can be used independent of text. Finally, both an emoji and a comment as a response. This suggests that when we contribute to theory on emotional contagion. Recent work has anchor emoji are present, observers’ felt emotions mirror the anchor, established that emoji can lead to contagion (Smith and Rose 2019). and that emoji and text responses supplement one another. The online context allows us to build on this work by: 1) exploring To sum, consumers’ emotional responses to mixed emotional contagion arising from multiple, anonymous emoji; 2) investigating posts are clarified by anchor emoji, leading them to make corre- contagion from mixed emotions; 3) testing a new source of conta- sponding emoji choices and potentially creating “emotional echo gion: others’ emoji responses to a post, rather the post’s content. Our chambers”. results establish anchor emoji as a novel form of unintended and indirect contagion. Is a Photo Worth a 1000 Words? The Usage of Images We predict that consumers will use the distribution of anchor and Text to Express One’s Opinions emoji in response to a given social media post as an input to their own responses. Hence, the same post may generate different emo- EXTENDED ABSTRACT tional and emoji responses from consumers, depending on the distri- People have more ways than ever to share and communicate bution of anchor emoji. We suggest that anchor emoji (e.g., sad vs. thoughts and experiences to other people. In particular, camera- love) clarify consumers’ emotions in response to mixed emotional enabled phones have created the ability to easily share experiences content (e.g., sad and happy), resulting in felt emotions and emoji not just verbally but also visually. Hence our research asks: How do choices that correspond with the anchor emoji. A field data sample people use visual and verbal information when communicating their of 100 posts (with ~200,000 emoji reactions) from CNN’s Facebook experiences? Specifically, we examine whether people use pictures page supported this conjecture: the first emoji response to a post sig- to complement or substitute text. Lay belief may suggest that “a pic- nificantly predicted subsequent emoji reactions. ture is worth 1000 words” and hence that pictures may substitute Study 1 provided a causal test of this notion. Participants (N text. Alternatively, people may want to “show and tell”, in which = 389; MTurk) imagined viewing a friend’s mixed emotional Face- case pictures may complement text (and vice versa). In this project, book post about their grandmother’s passing (i.e., happy and sad; we investigate how and why people combine these different forms Williams and Aaker 2002). We manipulated the anchor emoji that of expression. appeared with the post. Participants saw no anchor emoji (all emoji Study 1 examines the relationship between text and pictures in were zero), one heart anchor emoji (all others zero), or one sad an- the context of online reviews. We examined whether greater focus chor emoji (all others zero). They indicated which of the six emoji on visual aspects of the experience would increase or decrease the they would choose if they were to respond to their friend’s post, and usage of photos in two large restaurant review datasets (from Yelp! reported their feelings on 14 emotion items. Anchor emoji condition and TripAdvisor). We found that people were more likely to post predicted participants’ chosen emoji: those in the sad (love) anchor a photo when they provided greater visual detail (measured by the condition were more likely to select a sad (love) emoji. Further, par- number of words from the “see” category in LIWC) in both the Yelp! ticipants in the love (sad) anchor emoji conditions felt more (less) (β = 0.11, t = 44.5, p < .001) and TripAdvisor (β = 0.05, t = 15.9, p positive and less (more) negative emotion than those in the control < .001) datasets. This relationship also occurred for the number of condition. Felt emotion mediated the relationship between anchor photos included (Yelp!: β = 0.06, t = 27.7, p < .001; TripAdvisor: β = emoji condition and emoji choice. A supplementary study (1A) repli- 0.03, t = 15.3, p < .001) and held for expert and non-expert review- cated these effects and demonstrated that anchor emoji impacted not ers and controlling for word-count (excluding see words) as well only emoji choice, but participants’ overall likelihood of responding. as restaurant-specific factors. These analyses suggest that reviewers Using the same mixed emotional post as Study 1, Study 2 aimed use pictures to complement words when they focus more on visual to replicate these effects using multiple anchor emoji (N = 311; un- aspects of the experience. dergraduates). In the sad (love) anchor emoji condition, participants In study 2, we manipulated the number of see words in a saw a like emoji and a sad (love) emoji with the number 107, in- product review and tested whether greater focus on visual aspects dicating that 107 others had responded to the post, some with like heightens photo use. This study followed a three-group design. We reactions and some with sad (love) reactions. The control condition recruited 598 participants (41.2% female; meanage = 35.8) on MTurk showed no anchor emoji. This study also manipulated the order in (preregistration: osf.io/jd23r/). Participants imagined they had writ- which participants chose an emoji and reported felt emotion, to con- ten a review of a computer mouse and read “their” 140 words- re- firm that choices were driven by felt emotion (and not vice versa). view. Participants in the no-see words condition read a review that We found that regardless of order, participants in the sad (love) an- included no words from LIWC’s “see” category. In the low-see chor emoji condition were more likely to choose a sad (love) emoji, [high-see] words condition “see” words constituted 10% [30%] of relative to the control condition and to the other anchor emoji condi- the review. Participants indicated whether they would consider add- tion. Felt emotion showed corresponding effects, and mediated paring a photo when posting this review (Yes = 1, No = 0), our indepen- ticipants’ emoji choices. dent variable. In line with the correlational data, a greater percentage Study 3 (N = 484; MTurk) explored the relationship between of participants in the high-see words condition (86%) indicated that emoji and text responses. We tested whether emoji responses sup- they would include a photo compared to the no-see words condition plement or replace commenting on a post (Luangrath et al. 2017). (75%), t(598) = 3.09, p = .002. Participants in the low-see words Our prediction was that the emotional contagion from anchor emoji condition were directionally but not statistically significantly more would increase the likelihood of responding with both. Using a dif- likely to include a photo with their review (82%) compared the no- ferent mixed emotional post (a dog dying), we manipulated anchor see words condition (75%), t(598)=1.4, p=.16. This study provides emoji condition (control, sad) and whether participants were: 1) causal evidence that using more visually-focused words heightens forced to choose either an emoji or a comment as a response; or 2) photo sharing. free to choose either response, or both. We found that participants in Advances in Consumer Research (Volume 48) / 1019

Studies 1 and 2 suggest a complementary relationship between posit a fluency-driven matching effect in online search such that context and photos. In study 3, we manipulated people’s focus within an sumers respond more favorably to search engine results and ad con- experience. We recruited 400 respondents on MTurk (41.2% female; tent containing language that matches their mindset. meanage = 35.8) but excluded 66 people due to non-compliance In the pilot study, we analyzed a data set of 24,039 search que- with instructions (preregistration: https://osf.io/472xy). Participants ries. We calculated a concreteness score for each search query in our imagined writing an email to a friend about their positive experience database using the concreteness dictionary developed by Paetzold at a restaurant. Participants first identified 3-5 elements that posi- and Specia (2016). Our analysis established that, contrary to con- tively influenced their experience that either related to the look of ventional wisdom, substantial variation exists in the abstractness/ the food and the design of the restaurant (visual-focus condition) or concreteness of language used by consumers when performing an to the staff and service (human-focus condition). Subsequently, all online search. participants wrote an email to their friend focusing on these elements Study 1 is a survey in which university students (N = 369) re- and then indicated whether they would consider adding a photo to flected on the last query they made using a search engine. We found this email (Yes = 1, No = 0). Importantly, participants were more that participants with a transactional goal executed queries that were likely to share a photo in the visual-focus (63%) than the human- more concrete (less abstract) than participants with an informational focus condition (46%), z=3.1, p=.002, again pointing to a comple- goal. Furthermore, we obtained correlational evidence that partici- mentary relationship. pants at an earlier stage of the decision journey (i.e., informational One aspect inherent to the real-world data but also the lab stud- goals) are relatively less likely to visit e-commerce websites and ies is the order in which text versus pictures are produced: text is corporate websites than consumers at a later stage (i.e., transactional written first, then photos are added. This structure may have con- goals). tributed to the observed complementary relationship which may not Study 2 tests our proposition that construal matching between exist if photos are selected first. In study 4, we manipulated the or- consumers’ stage in the decision journey and the concreteness of the der of writing (text-first vs. picture-first) and altered whether text language in a search result will increase their engagement. In this 2 and pictures were produced on the same page or different pages. (journey stage: informational vs. transactional) x 2 (search results: We recruited 466 participants on MTurk (preregistration: https:// abstract vs. concrete) between-participants design, we informed 490 osf.io/hmfzc). Procedures followed those of study 3 except that par- MTurkers that they were repainting their living room. The language ticipants indicated 3-5 elements of their choice that impacted their used in the description of their journey stage or in the search results experience positively. In the text-first condition, participants first was either abstract or concrete. Results confirmed our prediction that wrote an email to their friend and subsequently picked a photo of participants were more likely to choose a search result whose con- that restaurant (available on Yelp!) to include in their email. In the creteness matched with the consumer journey stage. picture-first condition, these steps were reversed. In the different Study 3 assesses how construal level varies for multiple stages pages condition, these steps were requested on separate pages, in the of the decision journey. Recruited by a market research firm, 357 same page condition they were requested on one page. We measured U.S. participants were informed that they were considering purchas- participants’ focus by their usage of “see” category words. Partici- ing a laptop computer, and that consumers tend to progress through pants themselves also indicated to what extent they focused on the three sequential stages – information, comparison, transaction – look of the food and design elements when writing the email. Par- when making a purchase decision. We assigned the participants to ticipants used more see words in the picture-first (M=0.73) than the one of the three stages, asked them to execute an online search, and text-first condition (M=0.65, t(466)=1.83, p=.06), suggesting again showed them advertisements whose language varied in concrete- a complementary relationship. They also used more see words when ness. We found that participants in the transactional stage executed text and picture were presented on the same (M=0.77) than on differ- searches that were more concrete (less abstract) and preferred more ent pages (M=0.61, t(466)=2.14, p=.03). Self-reported measures of concrete results than participants in the other two stages. visual focus parallel these findings. Study 4 explores how the construal matching effect influences Taken together, we find that people use photos complementary consumers’ perceived progression through the decision journey. to text in communicating their experiences. This is particularly true In a 2x2 design, 427 MTurkers were informed that they wanted to when they focus on visual aspects of an experience, either due to an learn more about grilling (informational stage) or that they needed to internal or external focus. While much research has examined how choose a charcoal grill (transactional stage). Then, after being shown purely verbal communications are perceived by readers, our findings a list of either abstract or concrete results, they were asked to indicate shed light on an important but unexplored area, namely how both the amount of progress they would make on their decision journey visual and verbal information are used jointly by communicators. if they visited these websites. Participants in the informational stage reported that they would make greater progress when they encoun- Construal Matching in Online Search: Applying Text tered abstract search results as opposed to concrete search results. Analysis to Illuminate the Consumer Decision Journey However, participants in the transactional stage reported that they would make greater progress when they encountered concrete search EXTENDED ABSTRACT results. This research incorporates Construal Level Theory (Liberman Finally, in a field experiment with over 128,000 ad impressions, and Trope 1998; Trope, Liberman, and Wakslak 2007) and text anal- we investigated actual click behavior in a real-world context. We bid ysis (e.g., Humphreys and Wang 2017; Berger et al. 2020) in order on a set of abstract and concrete keywords that appeared in search to improve outcomes in online search. We propose that as consum- queries for a particular product category. We created two versions ers move through the decision journey, they adopt different goals of an online advertisement that were written using either abstract or that can be detected through text analysis of their search queries. We concrete language. The field study corroborated the construal match- hypothesize that consumers with informational [transactional] goals ing effect; click-through rates were higher when the language of the are likely to generate search queries that exhibit different levels of search query matched the concreteness of language in the advertise- mental construal, or mindsets, i.e., abstract [concrete]. We further ment. 1020 / Words and Beyond: Using Language to Better UnderstandConsumer Experiences and Goals

We make three contributions to research on search and con- trust in judgment and willingness to work together (r=.289, p<.001). sumer journey. First, we demonstrate the influence of construal level Furthermore, the receptiveness model from Study 1 predicted these in online search and show that firms can increase engagement by positive interpersonal outcomes (r=.232, p<.001), and also showed matching their content to consumers’ mindsets. Second, we illus- that partners’ receptiveness converges over time, indicating that one trate the important substantive consequences of matching construal of the benefits of receptiveness is that it is returned in kind (r=.335, for improving the search experience and making search advertising p<.001). However, people could not predict how receptive their part- more cost- and time-efficient. This work outlines a scalable method ner would rate them (r=.048, ns). Our language model held their own for firms to identify consumers’ search goals in text and facilitate speech to a different standard than their partner’s speech, focusing their progression through the consumer journey. Finally, we high- more on formality (titles, gratitude, etc.) than demonstrations of lis- light the contextual and changing nature of the decision journey. By tening. outlining a straightforward and implementable process by which In Study 3 we extend this result to conversations within glob- consumers’ mindsets can be reliably categorized based on their own ally-distributed organizations where disagreement naturally arises, search queries, we add to the burgeoning literature in marketing that and where people are free talk about many different topics, with shows how text analysis can provide nuanced insight into consum- many different people. In Study 3A we examine receptiveness er needs and goals. Digital marketers can leverage this newfound among 3,303 students in policy-themed massive open online courses knowledge to serve content that corresponds with consumers’ con- at HarvardX. We collect ideology measures and compare them to strual level, resulting in greater, customized consumer engagement the contents of the class discussion forums. We find that on average, at a lower cost to advertisers. students were less receptive to students they disagreed with (r= .099, p < .05). However, the receptiveness of students’ posts predicted the Conversational Receptiveness: Improving Engagement receptiveness of the replies they received from other students who with Opposing Views disagreed with them (r= .226; p<.001). This suggests that receptiveness is often, and individual choices to be more receptive can foster EXTENDED ABSTRACT a more receptive dialogue going forward. In Study 3B, we measure Disagreement is a fundamental feature of social life, in civic receptiveness during the editorial process of correcting Wikipedia spaces, in professional organizations, and in personal relationships articles. We borrowed a dataset of talk page threads, in which 585 at home. Opposing viewpoints are often inevitable in the pursuit threads ending in personal attacks were each matched to similar of more important organizational and interpersonal goals. Engage- thread without an attack (Zhang et al., 2018). We found that editors ment with diverse perspectives can also help us increase the accu- who were less receptive were more likely to be attacked themselves racy of our own belief. However, disagreement can also give rise to (59.9%; p<.001). biased processing, negative inferences, and conflict. While engage- Study 4 was similar to Study 1 except that some writers were ment with opposing viewpoints can be beneficial, its effects will be first taught a “recipe for receptiveness”, which was developed as a tempered by the contents of those interactions (see Bail et al. 2019; 100-word summary that can be cheaply deployed as a nudge during Paluck, Green and Green 2018). Here we examine whether “con- conversation. Opponents who read the responses from writer who versational receptiveness” can foster co-operative goals during dis- saw the recipe thought they were more trustworthy and persuasive agreement and prevent conversational conflict spirals. than writers who responded naturally. However, writers did not al- In this research we conduct four studies, and all data, analy- ways predict this effect, and expressed surprising hesitation to be sis code, stimuli, and preregistrations from each study are available receptive in the future. (anonymously) at https://bit.ly/2QwyiuL. In Study 1 we instructed Overall, our results suggest that receptiveness is measurable and 1,102 participants to write responses to statements written by peo- has meaningful interpersonal consequences but can be under-utilized ple with whom they disagree, on one of two controversial issues. in part because speakers can misjudge their own receptiveness. Our A separate group of 1,322 participants read responses from people results also highlight an under-discussed element of recent efforts to with opposing viewpoints, and evaluated how receptive the writer improve civic discourse: The linguistic behavior that people exhibit had been. We parsed the text of the responses into features from the in conversation can powerfully affect their partners’ perceptions, en- politeness R package (Yeomans, Kantor and Tingley 2018), and we gagement, and willingness to cooperate. trained a supervised machine learning algorithm (Friedman, Hastie and Tibshirani 2010) to build a receptiveness detection model that REFERENCES was generalizable (for datasets from other domains) and interpreta- Bail, C., Argyle, L., Brown, T., Bumpus, J., Chen, H., Fallin ble (to design interventions). This model was just as accurate (pair- Hunzaker, M., . . . Volfovsky, A. (2018). Exposure to opposing wise accuracy = 66.8%; p<.001), as any one human rater (65.2%; views on social media can increase political polarization. p<.001). The model focuses on the structural, domain-general ele- Proceedings of the National Academy of Sciences of the ments of the language (hedges, acknowledgment, negation, reason- United States of America, 115(37), 9216-9221. ing), and the model’s accuracy was unaffected when it was trained Berger, Jonah, Ashlee Humphreys, Stephan Ludwig, Wendy W. and tested on different topics (65.2%; p<.001). In Studies 2 and 3, we Moe, Oded Netzer, and David A. Schweidel (2020), “Uniting apply the model in conversations from organizational contexts where the Tribes: Using Text for Marketing Insight,” Journal of disagreement naturally arises. Marketing, 84 (1), 1-25. In Study 2, we collected conversations between 238 senior lo- Cohen, David (2017), “Facebook’s emoji reactions are now cal government officials in an executive education program, who one year old, and they have been used 300 billion times.” were paired up to discuss controversial policy topics (using a nega- Retrieved August 5, 2018 from: https://www.adweek. tive assortative matching algorithm to ensure they all disagreed com/digital/facebook-reactions-one-year-old-300-billion/. with their partner). After the conversation, participants rated their Facebook.com (2019), “Q4 2019 Results.” Retrieved February own and their partner’s receptiveness. Partner-rated receptiveness 27, 2020 from: https://s21.q4cdn.com/399680738/files/doc_ was associated with a range of positive interpersonal benefits, like financials/2019/q4/Q4-2019-Earnings-Presentation-final.pdf. Advances in Consumer Research (Volume 48) / 1021

Friedman, J., Hastie, T., and Tibshirani, R. (2010). Regularization Pilon, Anne (2015), “Emoji Survey: Users Want New Rolling Eye paths for generalized linear models via coordinate descent. Face Most.” Retrieved September 27, 2018 from: https://aytm. Journal of statistical software, 33(1), 1. com/blog/emoji-survey/. Humphreys, Ashlee, and Rebecca Jen-Hui Wang (2017), Smith, Leah and Randall Rose (2019). Service with a smiley face: “Automated Text Analysis for Consumer Research,” Journal Emojional contagion in digitally mediated relationships. of Consumer Research, 44 (6), 1274-306. International Journal of Research in Marketing. Liberman, Nira, and Yaacov Trope (1998), “The Role of Feasibility Toubiana, Madeline, and Charlene Zietsma (2017), “The message and Desirability Considerations in Near and Distant Future is on the wall? Emotions, social media and the dynamics of Decisions: A Test of Temporal Construal Theory,” Journal of institutional complexity,” Academy of Management Journal, Personality and Social Psychology, 75 (1), 5-18. 60 (3), 922-953. Luangrath, Andrea Webb, Joann Peck, and Victor A. Barger (2017), Trope, Yaacov, Nira Liberman, and Cheryl Wakslak (2007), “Textual paralanguage and its implications for marketing “Construal Levels and Psychological Distance: Effects on communications,” Journal of Consumer Psychology, 27(1), Representation, Prediction, Evaluation, and Behavior,” 98-107. Journal of Consumer Psychology, 17 (2), 83-95. Ma, Jingjing, and Neal J. Roese (2014), “The Maximizing Mind- Williams, Patti, and Jennifer L. Aaker (2002), “Can mixed emotions set,” Journal of Consumer Research, 41 (1), 71-92. peacefully coexist?,” Journal of Consumer Research, 28(4), Morrison, Kimberlee (2017), “Emojis Lead to Higher Engagement 636-649. on Instagram.” Retrieved September 27, 2018 from: https:// Yeomans, M., Kantor, A., and Tingley, D. (2018). The politeness www.adweek.com/digital/emojis-lead-to-higher-engagement- package: Detecting politeness in natural language. R Journal, on-instagram-report/. 10(2). Paetzold, Gustavo, and Lucia Specia (2016), “Inferring Zhang, J., Chang, J., Danescu-Niculescu-Mizil, C., Dixon, L., Psycholinguistic Properties of Words,” In Proceedings of Hua, Y., Taraborelli, D., and Thain, N. (2018). Conversations the 2016 Conference of the North American Chapter of the gone awry: Detecting early signs of conversational failure. In Association for Computational Linguistics: Human Language Proceedings of the 56th Annual Meeting of the Association for Technologies, pp. 435-440. Computational Linguistics, 1, 1350-1361. Paluck, E. L., Green, S. A., and Green, D. P. (2018). The contact hypothesis re-evaluated. Behavioural Public Policy, 1-30. Price, Purchases, and Beyond: A Multidimensional Perspective Chair: Manissa P. Gunadi, ESADE, Univ. Ramon Llul, Spain

Paper #1: Choice Bracketing and Consumer Choice under ‘pay-what-you-want.’ Using a large-scale panel dataset, Jung, Liu, Uncertainty and Nelson found that consumers voluntarily contribute a positive Liat Hadar, Tel Aviv University, Israel amount, but that this contribution gradually declines over time. Shai Danziger, Tel Aviv University, Israel, and the University Collectively, this session strives to answer two broad questions. of Sydney, Australia Firstly, how are consumer purchase decisions influenced by price Vicki Morwitz, Columbia University, USA information, be it past prices, aggregate price perceptions, or prices of other, related products? Secondly, how do consumers behave Paper #2: The Impact of Historic Price Information on Purchase under specific pricing schemes and strategies? The papers arguably Deferral show the multidimensionality of pricing, and how diverse research Manissa P. Gunadi, ESADE, Univ. Ramon Llul, Spain on this topic can be. While price plays a key role in the consumers’ Ioannis Evangelidis, ESADE, Univ. Ramon Llul, Spain decision-making process, at the same time, it is also an important in- Paper #3: The Negative and Positive Consequences of Placing strument that retailers have managerial control over. Accordingly, we Products Next to Promoted Products believe these papers—which are wide-ranging but complementary in Christina Kan, University of Connecticut, USA nature—offer important theoretical and practical insights. Whether Donald Lichtenstein, University of Colorado Boulder, USA companies are choosing to use an elective pricing scheme, or to in- Chris Janiszewski, University of Florida, USA corporate price promotional deals, or to convey price changes, or to Yan (Lucy) Liu, Texas A&M University, USA opt for depth or frequency pricing strategies—this session highlights Paper #4: An Examination of Consumers’ Long-term Social how they could all have substantial impact on consumers’ purchas- Preferences Under Elective Pricing ing behavior. Minah H. Jung, New York University, USA Xiao Liu, New York University, USA Choice Bracketing and Consumer Choice under Leif D. Nelson, University of California, Berkeley, USA Uncertainty

SESSION OVERVIEW EXTENDED ABSTRACT Price can arguably be considered one of the most important at- Consumers’ inclination to use narrow bracketing, that is, to contributes for consumers’ purchase decisions. Importantly, besides ac- sider choices independently of other similar choices, or broad brack- tual prices, consumer price perceptions play a crucial role in impact- eting, that is, to consider multiple, similar choices jointly, may sys- ing purchase decisions (Zeithaml 1988; Gupta 1988; Monroe 2003; tematically influence choice. For example, consumers may choose Krishna et al. 2002). This session was put together with the aim of between two retailers based on an isolated assessment regarding deepening our understanding of how consumers’ decisions and be- which retailer will be cheaper on a particular purchase or based on an haviors are influenced by factors related to pricing. Each paper will aggregate assessment of which retailer will be cheaper on average, approach this from a different theoretical angle—making the session across multiple purchases. These choices may differ because the re- as a whole highly relevant and appealing to a wide range of con- tailer more likely to be cheaper on a particular purchase is not neces- sumer researchers. Importantly, on top of using data from behavioral sarily the one that will be cheaper on average. We examine the effect experiments, the papers also present evidence drawn from incentive- of choice bracketing on preference for retailers using frequency (i.e., compatible paradigms or field studies, large-scale panel data, and offering many small discounts) versus depth pricing strategies (i.e., secondary data from various industries. We contend that this con- offering fewer, but larger discounts). tributes to the overall robustness and eventual generalizability of our Consumers generally buy more often at frequency retailers collective findings. than at depth retailers because they rely on trial-by-trial predictions The first two papers focus on how purchase decisions are guided of which retailer will be cheaper on the next shopping trip, rather by how consumers make use of price information. When engaging in than on their beliefs regarding which retailer is cheaper on average purchases, consumers can rely on narrow or broad bracketing. Mor- (Danziger, Hadar, and Morwitz 2014). We examine conditions under witz, Hadar, and Danziger document conditions under which choice which choice is likely to be guided by average price perceptions. We is likely to be guided by average price perceptions versus predictions find that consumer reward programs that reward multiple purchases of specific future prices, and how these perceptions and predictions or set minimum purchase requirements encourage broad bracketing, impact purchase for consumers of frequency or depth pricing retail- prompt reliance on average price judgments and preference for retailers. ers who consumers believe offer a lower price on average, whether it Next, Gunadi and Evangelidis evaluate how consumers’ pur- is the depth or frequency retailer. chase decisions are influenced by historic price information. They fo- Participants in our studies were incentivized to minimize overall cus on the interaction between frequency and direction of past price costs when making multiple choices between two options. Choices changes on purchase decisions for a given product. were made without knowing the options’ prices in advance. In study In the third paper, we turn to examine whether the purchase of 1, 120 students made 21 daily choices between two fictitious coffee a specific product could be influenced by price promotional informa- shops and received an incentive-compatible payoff. Unbeknownst to tion for another product. Kan, Lichtenstein, Janiszewski and Liu participants, one coffee shop used frequency pricing while the other document the deal proximity effect, and find that this crucially de- used depth pricing. We manipulated, between-subjects, which coffee pends on substitutability of the respective products. shop was cheaper on average. After participants completed 20 choic- We round out the session with a longitudinal assessment of es between the two coffee shops the bracketing manipulation was how consumers behave under a consumer elective pricing scheme introduced. Participants in the narrow bracketing condition received

10 coupons offering a 10% discount on the price of coffee. Each tion, the judged average price difference did not mediate the effect of coupon could be redeemed at either coffee shop; participants in the Cheaper price on choice on choice (B=-.02, 95% CI=-.06,.02). broad bracketing condition received one coupon offering a 10% dis- Our results suggest that companies whose products or services count on the price of coffee, forone of the coffee shops, valid for the are perceived as relatively cheaper on average should offer reward next 10 purchases. They indicated which coffee shops’ coupon they programs that trigger broad bracketing, for example, by rewarding wanted and received the discount only if they chose to buy at this multiple purchases. In contrast, companies whose products or ser- shop. Supporting our hypothesis, we found a Bracketing by Cheaper vices are perceived as relatively expensive on average are better off distribution interaction effect (χ²(1)=4.95;p =.026): participants were rewarding individual purchases. more likely to shop at the cheaper shop in the broad bracketing condition (χ²(1)=1.9, b=1.4, p=.057), but not in the narrow bracketing The Impact of Historic Price Information on Purchase condition (χ²(1)=-1.3, b=-.97, p=.184). Deferral Study 2 (N=202) examines the effect of choice bracketing on the propensity to choose the option consumers judge as cheaper on EXTENDED ABSTRACT average. Broad bracketing is prompted by having participants made In current retail settings, prices of goods in the market change multiple choices before receiving any feedback (while in the narrow frequently over time and consumers now have increasing access to bracketing condition feedback is provided following every choice; historical prices. We examine how frequency and direction of past Read, Loewenstein, and Rabin 1999). Actual average prices were the price changes influence consumers’ decision to buy later versus same across the depth and the everyday low price (EDLP) retailers. now. We advance an expectation-based framework—proposing that The Bracketing by judged average price difference interaction was frequency of price changes impacts the expectations people form significant (χ²(1)=5.09; p=.024): in the broad bracketing condition about future price changes. We set forth two key predictions. For the cheaper the judged average price of the depth grocer, relative price decreases, we predict that consumers would be more likely to to the EDLP grocer, the higher the choice likelihood of the depth defer purchase upon observing a higher frequency of past decreases grocer (B =.34; 95% CI=.05,.64). In the narrow bracketing conditon, in price. For price increases, consumers would be less likely to defer however, average price difference did not predict choice (B= .004; upon seeing a higher frequency of past increases. 95% CI=-.01,.02). Research shows that consumers tend to expect a sequence of Study 3 replicated this effect with non-dichotomous distribu- events to represent alternations instead of runs (Tversky and Kahne- tions of time. Participants (N=201) made repeated choices between man 1974). Upon observing a series of identical outcomes, consum- two roads and were instructed to minimize overall commute time. ers may believe an alternative outcome would be due. However, in One road was associated with a frequency time distribution and the the price changes context, we predict that consumers may be more other with a constantly short time distribution. In the narrow brack- likely to expect the same change (decrease or increase) as previ- eting condition participants received 10 vouchers to a fast lane that ously observed to repeat in the future (Gilovich et al. 1985; Carlson reduces the commute time by up to five minutes; participants in the and Shu 2007). We posit that these expectations will be stronger, broad bracketing condition chose a single voucher for the fast lane, the more changes in the same direction consumers have previously valid for the next 10 trips, for one of the roads. We, again, found a encountered. Specifically, for price decreases, consumers will hold Bracketing condition by judged average time difference interaction stronger beliefs that the price will decrease further in the future, upon (B=.008; 95% CI=.002,.01; p=.015): in the broad bracketing condi- observing relatively higher frequency of decreases in price. Similar- tion, the lower the judged average time of the frequency road, rela- ly, for price increases, consumers will hold stronger beliefs that the tive to the constant road, the higher the choice proportion of the fre- price will increase further, upon observing higher frequency of in- quency road (B =.012; 95% CI=.008,.017; p<.0001). In the narrow creases. In turn, these expectations that consumers form about future bracketing conditon, however, the judged average time difference price changes will affect their decision to make or defer purchase. did not predict choice (B=.004; 95% CI=-.001,.009; p=.082). Consumers should be more (less) likely to defer purchase, the more In Study 4 (N=400), holding reward magnitude constant, par- they believe that the price of the product they consider purchasing ticipants in the narrow bracketing condition were rewarded for will decrease (increase) in the future. individual choices whereas participants in the broad bracketing We found collective support for our predictions in eight pre- condition were rewarded only if they completed a fixed number registered experiments (N=5933) and three studies using secondary of purchases at a given retailer. We manipulated which option was data (N=3537). In most of the studies, we employed a 2 (direction cheaper on average. Our hypothesis was supported by a moderated of price change: decrease vs. increase) by 2 (frequency of change: mediation analysis (PROCESS model 14; Hayes 2013) in which the low vs. high) between-participants design. Participants were asked independent variable was the Cheaper Retailer, the mediator was to imagine that they were buying a specific product, and were pro- judged average price difference, the dependent variable was Depth vided with historic price information (in numeric form) based on choices, and the moderator was Bracketing. We found a significant the experimental condition. We manipulated frequency of change by Bracketing by judged average price difference interaction (B=.03, altering how many changes in the ticket price were communicated 95% CI=.01,.05; p=.008) and a significant moderated mediation ef- to the participants. In the low frequency condition, there was one fect (B=.05, 95% CI=.01,.11): average price difference mediated the change in price. In the high frequency condition, there were four effect of Cheaper price on choice in the broad bracketing condition changes. Importantly, the final price and the absolute magnitude of (B=.04, 95% CI=.01,.07), such that the the judged average price of change were held constant across conditions. Our dependent variable the depth retailer relative to the frequency retailer was lower when was participants’ decision to buy the product now or later. the depth grocer was cheaper than when the frequency grocer was Study 1a established the expected interaction between frequen- cheaper. In addition, the lower the average price of the depth grocer cy and direction of change on purchase deferral in the context of was judged to be (relative to the frequency retailer), the more the flight tickets. As predicted, when price decreased, participants were depth grocer was chosen. In contrast, in the narrow bracketing condi- more likely to defer purchase when the frequency of price changes was high (27.4%), compared to low (10%), Waldχ2 = 18.50, p<.001. 1024 / Price, Purchases, and Beyond: A Multidimensional Perspective When price increased, participants were less likely to defer purchase The Negative and Positive Consequences of Placing when the frequency of price changes was high (26.4%) versus low Products Next to Promoted Products (46.3%), Waldχ2 = 16.87, p<.001. We replicated this pattern of results when holding the amount of information constant across low/ EXTENDED ABSTRACT high frequency conditions (Study 1b) and when price information Consumers often select from a variety of competing products was presented visually in graphs (Study 1c). in multi-product displays. Some products might be discounted while Study 2 tested our expectations-based account. Consistent with others are regular price, and products are positioned at various loca- our theoretical framework, we found that controlling for the mag- tions within the display. Do price deals differentially impact demand nitude of change, frequency strengthens expectations about future for other products depending on their location within a display? price changes. Consumers hold stronger expectations that prices will We find evidence for a deal proximity effect that differs depend- decrease further in the future (F(1,598) = 46.84, p<.001), and are ing on product substitutability. When the proximal and distal items more likely to defer purchase when the frequency of price decreases are strong substitutes for the promoted item, a price promotion de- is high versus low. Conversely, consumers hold stronger expecta- creases the sales of proximal products relative to distal products: a tions that prices will increase further in the future, and are less likely negative proximity effect. We surmise that this occurs because the to defer purchase when the frequency of price increases is high ver- promoted product attracts attention that then spills over to the proxi- sus low F(1,598) = 31.12, p<.001, (interaction F(1,598) = 77.16, mal product. Because both products are strong substitutes, they enter p<.001). Further, a conditional process analysis showed that expec- the same consideration set. The reduced price on the promoted prod- tations mediated the effect of frequency on deferral both when price uct results in the promoted product being more attractive, leading to decreased and increased (index of moderated mediation = -2.77, sales losses for the proximal product. SE=.42, 99%LLCI =-3.94, ULCI=-1.82). However, when the proximal and distal items are weak substi- Subsequent studies test potential moderators. Studies 3a and tutes for the promoted item, the promoted product increases the sales 3b document the role of monotonicity. The presence of an incon- of proximal products relative to distal products: a positive proxim- sistent change—a change that goes in the opposite direction, com- ity effect. Because the proximal product is a weak substitute for the pared to previous changes—attenuated the impact of frequency on promoted product, it is unlikely to enter the same consideration set purchase deferral. In these studies, we had low frequency, high fre- as the promoted product and thus, the reduced price on the promoted quency monotonic, and high frequency nonmonotonic conditions. item is not a relevant referent for the proximal product. However, the In Study 3a, when prices were decreasing, participants in the high proximal product still benefits from the increased attention, which frequency monotonic condition were less likely to defer the purchase positively impacts its sales. (29.6%) compared to their counterparts in the low frequency con- Our first study explored whether grocery store managers would dition (60%), Waldχ2 = 36.48, p<.001. Importantly, this result was intuit the proposed pattern of effects. Grocery store managers were eliminated when changes were not monotonic (high nonmonotonic = recruited by Qualtrics to participate in this study. They were shown 64.2%, Waldχ2 = .74, p=.389). We replicated these findings for price a planogram from a yogurt shelf and told that a product was being increases (Study 3b). promoted at a price discount. They were first asked to estimate the Study 4 showed that the impact of frequency on purchase de- impact of this promotion on the percentage change in sales of four ferral is stronger when the magnitude of the change in price is rela- other products on the shelf in a 2 (Proximity to discounted product: tively larger. We added a new factor, magnitude of change (small vs. proximal v. distal) x 2 (Substitutability: strong substitute v. weak large) to our usual 2x2 design. The three-way interaction (direction, substitute) design. Results indicate managers predict that a price frequency, and magnitude) on deferral was marginally significant promotion has a significantly stronger impact on stronger substitutes (Waldχ2 = 2.77, p=.096). Importantly, the interaction between our than weaker substitutes, but do not anticipate that price promotion direction and frequency factors was stronger when the magnitude of effects differ by proximity or the interaction of proximity by substi- the change was large (Waldχ2 = 46.10, p<.001) versus small (Waldχ2 tutability. = 16.22, p<.001). Study 2 analyzes a secondary dataset of yogurt sales obtained Study 5 adds to the existing body of evidence for our effect from a retail grocer. The dataset provided information on each yo- by replicating our findings (interaction: Waldχ2 = 6.01, p=.014) in a gurt’s sales, regular price and whether it was being promoted at a dis- setting where decisions are incentivized and consequential for par- counted price. We use the grocer’s planogram to determine how far ticipants. apart each yogurt is located from one another. We also ran a pretest Lastly, to complement the previous studies which were con- on mTurk to determine perceived substitutability between yogurts ducted in controlled experimental settings, Studies 6a-c document based on various product attributes. A fixed effects analysis reveals our basic effects using data from online cryptocurrency, housing that cross price promotion effects differ significantly depending on market, and automobile transactions. substitutability, distance within the planogram and their interaction. In sum, we found that in the case of price decreases, partici- For highly substitutable products, a sales promotion had a negative pants were more likely to defer purchase upon observing a relatively cross price promotion effect that diminished as the distance between higher frequency of changes. On the contrary, in the context of price the two products increased. For weakly substitutable products, a increases, participants were less likely to defer purchase upon ob- sales promotion had a positive cross price promotion effect that di- serving a relatively higher frequency of changes. These effects are minished as the distance between the two products increased. explained by variations in consumers’ expectations about future Studies 3a and 3b were designed to address limitations due to prices. the correlational nature of the study 2 dataset. We ran two field studies at a campus bookstore in which we manipulated the proximity of products to a discounted product within a display. In study 3a, we used two different product categories to serve as weak substitutes for one another; plush toy keychains and car decals. We observe a positive deal proximity effect: when a plush toy keychain (car emblem) Advances in Consumer Research (Volume 48) / 1025 is promoted at a discounted price, the proximally located car emblem The extant research offered important insights into how con- (plush toy keychain) sells significantly more than the distally located sumers determined how much to pay under elective pricing, but car emblem (plush toy keychain). In study 3b, we used five product much less known about the stability of consumers’ social preferenc- category replicates (energy drinks, keychains, lanyards, socks, and es. Most studies that document consumers’ decisions under elective trail mix), each with its own display, to test for a negative proxim- pricing used the data that typically involved either single purchases ity effect. For each product category, one product was chosen to be or on aggregated data across time. This limitation is notable because the discounted product, and two products were chosen to be strong it leaves open the possibility that apparently strong social prefer- substitutes for the discounted product. Collapsing across product ences might actually be fleeting. A few studies looked at consumers’ replicates, a test of proportions showed that the choice share of the repeated purchases under elective pricing but showed mixed results proximal product was significantly lower than the choice share of on their long-term behavior. Some found that consumers’ payments the distal product. increased over time (Kim et al. 2010), whereas the others document- Our next study used two different operationalizations of prod- ed a decreasing pattern of payments over time (Levitt 2006; Riener uct substitutability to provide further evidence that the effects re- and Traxler 2012). Although these papers provided new insights into sult because of differences in perceived substitutability. We used a the long-term dynamics of consumers’ social preferences, their data simulated internet shopping experience with participants on mTurk lacked a critical feature-they could not observe individual-level so- (n = 559). The two substitutability operationalizations were goal- cial preferences over time. A few studies that recorded individual- based substitutability and product attribute-based substitutability. level payments considered behavior in relatively short duration. Our These conditions were run across eight product category replicates goal is to examine the stability of consumers’ long-term social pref- (cookies, chocolate covered snacks, savory snacks, candies, jams, erences in a market exchange environment by addressing the key chocolate bars, sodas and soaps) in a repeated measures latin-square weaknesses of the previous research. design. We find evidence for a negative deal proximity effect when We analyze panel data from an online media retailer that sells substitutability is highest, a null effect when substitutability is medi- bundles of electronic books (eBooks) written by independent au- um and a positive deal proximity effect when substitutability is low. thors. The data set includes 163,551 transactions from 57,196 cus- Across five studies, we observe evidence for a negative deal tomers, from the company’s launch on August, 6th, 2012 to June, proximity effect when substitutability is high and a positive deal 26th, 2017. Our data contains the individual-level transactions for proximity effect when substitutability is low. This pattern of effects the sales of 114 bundles of eBooks, which contain a total of 1,147 is novel to marketing practitioners and can offer actionable insights; books, written by 748 independent authors. retailers can use product proximity to actively manage the consider- Given the firm’s pricing features, we use consumer elective ation sets and purchases of consumers. contribution as a proxy measure of customers’ social preferences. These results contribute to literature on 1) price promotions We define consumer elective contribution to be how much more con- (e.g., Blattberg and Wisniewski 1989, Rooderkerk, Van Heerde, and sumers pay than the minimum or bonus price for a specific bundle Bijmolt 2013) by offering shelf proximity as an important modera- that they purchase. The notion is that the entirety of this marginal tor of cross price promotion effects, 2) consideration sets (e.g., Ne- contribution is offered without any additional direct benefit to the dungadi 1990, Shocker et al., 1991) by providing a proximity-based customer. explanation for influences on brand co-occurrence in the same versus Consumers’ elective contribution decreases with each subse- different consideration sets, and hence contexts in which brands are quent purchase; falling from an average initial elective contribution more/less likely to directly compete, and 3) retail shelf optimization of over $2.50 to about $1.00 by the 25th purchase. More important, (e.g., Corstjens and Doyle 1981, Eisend 2014) by offering insight however, is that even by the 25th purchase, customers are still mean- into how to organize shelves so that multiple brands (e.g., promoted ingfully contributing well beyond the bonus or minimum price. The brand, weak substitute proximal brand) might benefit most from the pattern is similar when we look at consumers’ elective contribu- promotional activity of a single brand. tion as percentage of the price paid. We looked at the sub-sample of 2,634 customers who purchased more than 10 bundles and find An Examination of Consumers’ Long-term Social that customers decrease their elective contribution with each subse- Preferences Under Elective Pricing quent purchase. The results are consistent when we control for the bonus price and multiple product quality factors variations across EXTENDED ABSTRACT the bundles. Standard economic models assume that people maximize mate- We observe that frequent buyers indeed contribute less than rial self-interests, ignoring any concern for others. Accordingly, con- infrequent buyers over time. Customers who only bought one bun- sumers want to pay as little as possible and tend to seek for deals and dle contributed an average elective contribution of $2.66, whereas discounts to exploit in marketplaces. Yet, people also give money consumers who bought two bundles contributed an average elective to charities or volunteer their time to help others. People care about contribution of $2.29. When a consumer buys more than 30 bundles, fairness and reciprocity even at the expense of their own material the average elective contribution decreases to $1.26, less than half self-interests (Andreoni and Miller 2002, Charness and Rabin 2005, of the level of elective contribution of those who bought only once. Rabin 1993). The average level of elective contribution and purchase frequency Consumer elective pricing applies in any transaction in which are negatively correlated such that frequent buyers contribute less buyers can decide their own prices for the goods and services that than infrequent buyers. sellers offer. Consumer elective pricing has been shown to be a valu- Consumers with increasing elective contribution exit sig- able tool for studying social preferences in real market contexts. It nificantly earlier than those with decreasing elective contribution. has been well documented that even when consumers are allowed to Therefore, the declining pattern of customers’ elective contribution be maximally selfish, most of them choose not to be (Kim et al. 2009, seems to be due to attrition of a certain segment. Customer exit is Mak et al. 2010, Gneezy et al. 2010, Jung et al. 2014, 2017). consequential for the firm. As we previously reported, frequent buyers start with a smaller elective contribution and their contributions 1026 / Price, Purchases, and Beyond: A Multidimensional Perspective decrease over time, but they also stay with the company for a longer Hayes, Andrew F. (2013), Introduction to Mediation, Moderation, duration. Although customers who contribute increasingly pay more and Conditional Process Analysis, New York: Guilford. per bundle, because they leave the company earlier, they contrib- Jung, Minah H., Leif D. Nelson, Ayelet Gneezy, and Uri Gneezy ute only an average cumulative elective contribution of $6.55. Al- (2014), “Paying More When Paying for Others,” Journal of ternatively, the customers who contribute decreasingly pay less per Personality and Social Psychology, 107 (3), 414-31. bundle, but offer substantially more cumulative elective contribution Jung, Minah H., Leif D. Nelson, Uri Gneezy, and Ayelet Gneezy to the company. (2017), “Signaling Virtue: Charitable Behavior Under In sum, most customers voluntarily contribute a positive Consumer Elective Pricing,” Marketing Science, 36 (2), 187- amount with their first purchases and continue to do so over subse- 94. quent purchases. Furthermore, although on average, customers al- Kim, Ju-Young, Martin Natter, and Martin Spann (2009) “Pay What ways pay more than they have to, their voluntary payments slowly You Want: A New Participative Pricing Mechanism,” Journal decline with each subsequent purchases. Lastly, we further identify of Marketing, 73 (1), 44-58. customer heterogeneity in the stability of their elective contributions ____ (2010), “Kish: Where Customers Pay as They Wish,” Review over time. We observe that generous customers are less likely to be of Marketing Science, 8 (2). loyal to the firm and therefore make fewer purchases over time. For Krishna, Aradhna, Richard Briesch, Donald R. Lehmann, and a firm that leverages customers’ social preferences for its revenue, Hong Yuan (2002), “A Meta-analysis of the Impact of Price customers who pay relatively less, yet loyal, contribute more to the Presentation on Perceived Savings,” Journal of Retailing, 78 firm in the long run than customers with larger initial contributions (2), 101-18. but less stable. Levitt, Steven D. (2006), “White-collar Crime Writ Small: A Case Study of Bagels, Donuts, and the Honor System,” American REFERENCES Economic Review, 96 (2), 290-94. Andreoni, James, and John Miller (2002), “Giving According Mak, Vincent, Rami Zwick, and Akshay R. Rao (2010), “Pay to GARP: An Experimental Test of the Consistency of What You Want as a Profitable Pricing Strategy: Theory Preferences for Altruism,” Econometrica, 70 (2), 737-53. and Experimental Evidence,” University of Cambridge, Blattberg, Robert C. and Kenneth J. Wisniewski (1989), “Price- Cambridge, UK. Induced Patterns of Competition,” Marketing Science, 8 (Fall), Monroe, Kent B. (2003), Pricing: Making Profitable Decisions, 3rd 291-309. ed., Boston: McGraw-Hill/Irwin. Carlson, Kurt A., and Suzanne B. Shu (2007), “The Rule of Three: Nedungadi, Prakash (1990), “Recall and Consumer Consideration How the Third Event Signals the Emergence of a Streak,” Sets: Influencing Choice Without Altering Brand Evaluations,” Organizational Behavior and Human Decision Processes, 104 Journal of Consumer Research, 17 (3), 263-76. (1), 113-21. Rabin, Matthew (1993), “Incorporating Fairness into Game Theory Charness, Gary, and Matthew Rabin (2004), “Expressed and Economics,” The American Economic Review, 1281-1302. Preferences and Behavior in Experimental Games”, Available Read, Daniel, George Loewenstein, and Matthew Rabin (1999), at SSRN 369402. “Choice Bracketing,” Journal of Risk and Uncertainty, 19 Corstjens, Marcel and Peter Doyle (1981), “A Model for (1-3), 171-97. Optimizing Retail Space Allocations,” Management Science, Riener, Gerhard, and Christian Traxler (2012) “Norms, Moods, 27 (7), 822-33. and Free Lunch: Longitudinal Evidence on Payments from Danziger, Shai, Liat Hadar, and Vicki Morwitz (2014), “Retailer a Pay-What-You-Want Restaurant,” The Journal of Socio- Pricing and Consumer Choice under Price Uncertainty,” Economics, 41 (4), 476-83. Journal of Consumer Research, 41, 761-74. Rooderkerk, Robert P., Harald J. van Heerde, and Tammo H.A. Eisend, Martin (2014), “Shelf Space Elasticity: A Meta-analysis,” Bijmolt (2013), “Optimizing Retail Assortments,” Marketing Journal of Retailing, 90 (June), 168-81. Science 32 (5), 699-715. Gilovich, Thomas, Robert Vallone, and Amos Tversky (1985), “The Shocker, Allan D., Moshe Ben-Akiva, Bruno Boccara, and Hot Hand in Basketball: On the Misperception of Random Prakash Nedungadi, (1991). “Consideration Set Influences on Sequences,” Cognitive Psychology, 17 (3), 295-314. Consumer Decision-Making and Choice: Issues, Models, and Gneezy, Ayelet, Uri Gneezy, Leif D. Nelson, and Amber Brown Suggestions,” Marketing Letters, 2 (3), 181-97. (2010), “Shared Social Responsibility: A Field Experiment in Tversky, Amos, and Daniel Kahneman (1974), “Judgment under Pay-What-You-Want Pricing and Charitable Giving,” Science, Uncertainty: Heuristics and Biases,” Science, 185 (4157), 329 (5989), 325-27. 1124-31. Gupta, Sunil (1988), “Impact of Sales Promotions on When, What, Zeithaml, Valerie (1988), “Consumer Perceptions of Price, Quality, and How Much to Buy,” Journal of Marketing Research, 25 and Value: A Means-End Model and Synthesis of Evidence,” (4), 342-55. Journal of Marketing, 52 (3), 2-22. Consuming Extraordinary Experiences: Personal Transformationin the Antistructural Realm Chairs: Chris Hackley, University of London - Royal Holloway, UK Rajesh Nanarpuzha, Indian Institute of Management Udaipur, India

Paper #1: Dissecting Consumption: The Temple Festival as an – Brazil). The study investigates how the extraordinary experience of Extraordinary Experience pilgrimage is experienced as a transformative sensory phenomenon. Rajesh Nanarpuzha, Indian Institute of Management Udaipur, Finally, paper 4 takes a Turnerian approach to the theorization of solo India travelling. Sixteen solo travelers are interviewed in order to generate Sridhar Samu, Great Lakes Institute of Management, India insight into the personally transformative extraordinary experiences Prakash Satyavageeswaran, Indian Institute of Management they sought and gained in the anti-structural realm of adventurous Udaipur, India solo travel. Jossin Shaji, Indian Institute of Management Udaipur, India The papers have in common a focus on consumer experiences that can be deemed extraordinary in the dual sense that they are Paper #2: The Extraordinary Experience of Death and Rebirth a) far from quotidian or routine and b) there is an expectation that in a Thai Temple Ritual undertaking these experiences will yield benefits of a spiritual and/ Rungpaka Amy Hackley, Birkbeck College, University of or transformative character. Each paper discusses extraordinary ex- London, UK periences that go beyond the realm of everyday life (Bhattacharjee Chris Hackley, University of London - Royal Holloway, UK and Mogilner, 2014) and hold the potentiality of a personal change Paper #3: Bodily Experiences and Self-transformation and renewal, even bringing joy (Arnould and Price, 1993, p. 25) and Jannsen Santana, ESCP Europe, France magic (Schouten and McAlexander, 1995). Olivier Badot, ESCP Europe, France In this session we explore three main research questions: Paper #4: Personal Transformation Through Extraordinary 1. In what ways can extraordinary consumer experiences be Experiences empirically and theoretically extended/re-framed, particu- Isadora Matsuda Sanchez de Rojas, EAESP-FGV, Brazil larly in the area of personally transformative experiences? Benjamin Rosenthal, EAESP-FGV, Brazil 2. What is the role of socio-historical institutions and rituals in creating extraordinary experiences? SESSION OVERVIEW 3. How does consumer participation affect extraordinary ex- Extraordinary experiences as consumption phenomena remain periences? a fascinating area of academic research (Beverland, Lindgreen, and Vanhamme 2012; Caru & Cova, 2006; Lanier Jr., Clinton, & Rader, Dissecting Consumption: The Temple Festival as an 2015; Turner and Bruner 1986). Within marketing, multiple perspec- Extraordinary Experience tives have been adopted to look at extraordinary experiences (Ar- nould & Price, 1993; Belk & Costa, 1998; Kozinets, 2002; Tum- EXTENDED ABSTRACT bat and Belk 2011). Even as extraordinary experiences have been The consumption of extraordinary experiences has been studied explored in diverse contexts (Tumbat and Belk 2011), these studies as service encounters (Arnould and Price 1993), as a contrast to the have been largely restricted to the consumption sphere in the United consumption of ordinary experiences (Bhattacharjee and Mogilner States and other developed markets (Arnould & Price, 1993; Belk 2014), as a means of consumer emancipation (Kozinets 2002), and & Costa, 1998, Kozinets, 2002). Even as this stream of research has as socially constructed fabrications (Tumbat and Belk 2011). This provided seminal insights about consumer behavior in the consump- is representative of the different lenses through which extraordinary tion of extraordinary experiences, it has left open the opportunity experiences have been conceptualized and examined in consumer to explore this phenomenon using differing consumer cultural per- research. However, a unifying thread connecting these research spectives. In this special session, we aim to explore diverse facets endeavors is an assumption of experiences being consumed by an of extraordinary experiences from hitherto under-explored cultural archetypal consumer, unshackled by the social context. Even as re- perspectives, with the broad aim of theoretically and empirically ex- search in the consumer domain has looked at consumer motivations, tending consumer research on extraordinary experiences. In doing the role of social institutions in shaping consumption experiences has so, we consider diverse empirical contexts, different methodologi- been overlooked. This is in line with Askegaard & Linnet’s (2011, cal approaches to studying extraordinary experiences, and disparate p. 381) contention that what is missing in Consumer Culture Theory types of extraordinary experiences. (CCT) research is the lack of a ‘context of context’. In paper 1, we consider the context of a single temple festival In this paper, we investigate the impact of social institutions in in India, and examine how socio-historical institutions shape con- the consumption of an extraordinary experience. In doing so, we at- sumption. The study takes a naturalistic ethnographic approach to tempt to answer the following broad research question explore the Nemmara-Vallangi vela, a temple festival in Kerala, In- 1. How is consumption of an extraordinary experience shaped dia. In paper 2, we consider a specific Buddhist ritual that occurs at by the social institutions of gender, social class, and caste? the intersection of spirituality and consumption- the Thai Theravada Buddhist Death Ritual for the Living. The theorization seeks to ex- To investigate this, we conducted a naturalistic inquiry of the tend Victor Turner’s idea of liminoid consumption into the spiritual Nemmara-Vallangi vela, a temple festival in Kerala, India. Planned sphere. Paper 3 entails a multi-sited sensory ethnography of four dif- as a multi-year participant observation of the week-long temple fes- ferent pilgrimages in Europe (Lourdes – France, Camino de Santiago tival, we have completed the first round of data collection and analy- – Spain) and Latin America (Juazeiro do Norte and Belém do Pará sis. A total of 25 ethnographic interviews of festival participants have been conducted. We have complemented this with still photographs

Advances in Consumer Research 1027 Volume 48, ©2020 1028 / Consuming Extraordinary Experiences: Personal Transformationin the Antistructural Realm and videos of the festival. In keeping with the tenets of constant ing from within the temple, taking occasional photographs, the lead comparison (Glaser and Strauss 1967), data collection, analysis and researcher translating and reflexively understanding the observed informant selection were done iteratively. events in the light of her own personal life experience of being a We have chosen the Nemmara-Vallanghy vela (temple festival) Thai Buddhist since birth. The devotees gathered in the temple are primarily for the information richness that it offers. This annual fes- invited by the monks to lie down in a row of neatly arranged coffins. tival is ostensibly fashioned as a competition between two adjacent Their palms are pressed together holding flowers, a candle and three villages (Nenmara and Vallanghy, in Palakkad district, Kerala state, incense sticks. The monks’ mesmeric chanting of the funeral ritual India) that share a common temple. Aspects of ritual rivalry as dis- chant Ānisong Sīa Sop: “The Blessings of Disposing of Corpses” cussed by Tarabout (1993) is prevalent in Nenmara-Vallanghy vela. interspersed with the striking of the resonant ceremonial gong adds Held annually, seemingly for the residents of the two villages, the to the sense of religious gravitas of the occasion. In spite of the simi- temple festival is a major tourist attraction, bringing in thousands of larities between this ritual and genuine death rituals in the temple, visitors. Complexity in the context is evidenced in the disparity in the death in this ritual, is symbolic. The coffins and their occupants consumption experiences, which we focus on. are covered with a shroud. When the chanting is finished, the shroud Our findings suggest that the impact of diverse socio-historical is removed and the devotee rises from the coffin. The literal transla- forces at play in shaping consumption experiences is clearly visible tion of Norn-Loeng-Sa-Dor-Cro is ‘laying down in a coffin you can in the Nemmara-Vallanghy vela. Visitors to the vela travel long dis- get rid of bad luck’. Devotees believe that the ritual will cleanse their tances to attend and consume the festival. People of Nenmara and soul by tricking bad spirits into thinking they have died, and hence Vallangy travel home in a process somewhat similar to Thanksgiving the bad spirit will leave them alone. Through the symbolic rebirth, in the USA or Christmas in some other countries. They bring their the devotee can start afresh with a new identity shorn of the negative families (spouse/children) so that they can share the history of the karma of their old identity. As a consequence, their fortune in this festival and can initiate them into the different activities. A mosaic life will improve. of different castes takes up specific roles, performing mandated ritu- In Thai Theravada Buddhism, religious doctrine and practice is als during the festival. One dominant caste (The Nair community) interwoven with folk beliefs and practices, and the realm of death is monopolizes the organization of festivities for one of the competing interwoven with that of the living. Spirits, ghosts and ancestors are villages (Nemmara) while a loose federation of members from dif- a vivid presence in material life, for good or for ill. Monks are often ferent castes takes up organizing activities for the other competing asked to intercede between the spirit world and the material world village (Vallanghy). Consumption of the festival is also gendered, to help devotees, in exchange for gifts of goods, food or money that with women taking up traditional household roles, and also occupy- are donated to the temple. For example, monks might convey propi- ing ‘safe space’ vantage points while consuming the festivities. Men, tious lottery numbers for those devotees hoping for a lottery win to on the other hand, ‘perform’ the festivities, and consume them, by improve their circumstance in life, or the monks might be asked to taking center-stage. Though ostensibly egalitarian in consumption, intercede for an ancestor to improve the ancestor’s lot in the next economic status hierarchies also play a role in festival consumption. life. This, in turn, would bring merit to the devotee who commis- The best elephants and the best music ensembles are bid for, and sioned the intervention, which could benefit them both spiritually, perform for the village with the higher financial capability. Festival and materially. rituals are also broken down and chunks are sold to the highest bid- The Thai Death Ritual for the Living requires some translation ders. For example, the house from which the main festival proces- for Western audiences. Firstly, the notion of death in the East differs sion begins is decided based on a bidding process. It is therefore, the from the Western notion of death as a finality and end point (Tumbat consumption of the overall festival or specific events in the festival and Belk, 2011: Bonsu and Belk, 2003). In Buddhism, there is a viv- that we investigate in this study in the context of the role played by id sense of the continuity of life and death. Life on earth itself, then, social institutions including gender, social class, and caste. is a liminal state (Turner, 1969) in the sense that it is but one of the Askegaard & Linnet (2011) argue that consumer research is Wheels of Life. In the Death Ritual for the Living, the liminality of largely characterized by a focus on the lived experiences of consum- this state, the sense of life being an unfolding state of potentiality, is ers (Thompson, Locander, & Pollio, 1989) without adequate focus rendered especially vivid through the intercession of the monks who on “macro-social explanatory frameworks.” In this paper, we heed mediate between this life and the next. Whilst the continuity of life the call by Askegaard & Linnet (2011) and investigate consumption and death is a reality in many practices of Thai Buddhist daily life, within the context of experiencing a temple festival in Kerala. We such as the belief in ghosts and the presence of a spirit house in most study this context through the different social institutional lenses of dwellings and business premises, there is still a need for frequent for- gender, caste, and social class. We find evidence for consumption mal ritual intercessions mediated by monks in order to connect with shaped by the contextual socio-historical development of these so- the spiritual world for particular reasons and on significant occa- cial institutions, and for intersectionality at the individual consumer sions. Some rituals, such as this one, are exceptional in that they fall level. Our findings point to the fallacy of considering an archetypal outside the usual rites of passage or calendrical rites in the religious consumer of extraordinary experiences and motivate us to look at life of Thai Buddhists and, instead, fall broadly within the realm of differing consumption experiences of the same extraordinary experi- commissioned intercessions. That is, they are consumer experiences ence, driven by larger socio-historical forces. that are paid for by the devotees and undertaken for specific reasons of advancement or benefit in the material world. The Extraordinary Experience of Death and Rebirth in a In this project we extend Turner’s (1969) ideas on the liminal Thai Temple Ritual phase of ritual process to a death ritual in order to draw out new theorizations of extraordinary consumer experiences (Arnould and EXTENDED ABSTRACT Price, 2003). For Turner (1969), liminal experience was confined to The ritual is visually striking, within the dramatic and beautiful compulsory rites, while he characterized consumer experiences as setting of the temple with its giant golden Buddha statues, the huge liminoid (Turner, 1974: 1982) rather than liminal. Liminoid experi- ceremonial gong and the robed monks. The researchers are observ- ences are freely entered into in a spirit of playfulness- they are not Advances in Consumer Research (Volume 48) / 1029 trivial, but neither are they permanent. Just as consumer experiences Our findings demonstrate the recurrent mentions to self-trans- can be disposed and repeated, so too can devotees of the Death Rit- formations in participants’ speech and practices not only in Lourdes, ual for the Living repeat the experience many times if they choose. but at the four pilgrimage contexts. Pilgrims reported different types “The liminoid is more like a commodity- indeed, often is a com- of transformations they’ve been through. They can address their modity, which one selects and pays for, than the liminal” (Turner, transformations to religious and non-religious causes, such as: 1982 p.55). Hence, Turner (1982) explicitly linked liminoid experi- To achieve their aimed transformations, pilgrims engage them- ence to consumer experience. The simulation of death and rebirth in selves in bodily enactments mainly related to (i) the place, and (ii) this case also offers opportunities for cross-cultural insights into the using the materials around. People going to these experiences feel a profoundly different understanding of spirituality, death and escha- connection with the very ground of the pilgrimage site and the con- tology that obtains in the Judeao-Islamic-Christian West, and in the suming interactions between their bodies and the place enables well- Buddhist and Hindu influenced East. In their moments in the coffin, being and self-transformations as expressed in the following quote: the devotees subsist in a state between life and death, and it is the liminal character of this state that opens up the potentiality of a new I think that the very ground of the pilgrimage site reveals a power life, a new spiritual identity and new possible futures both as a hu- that can help people to find their healing. And that is the reason man being, and subsequently as a being on a different Wheel of Life. people come to these religious sites. (Fatima, Lourdes) We frame this ritual as an extraordinary consumption experience because it is consumed as a commercial transaction and trans- Individuals engage their bodies in enactments involving mate- ports the consumer into the realm of the extraordinary. The case rials, such as in the words of Lara explaining how the doing of the illustrates a liminal dimension that we suggest is latent in many Camino de Santiago and the difficulties involved at carrying a 11 extraordinary consumer experiences but foregrounded in spiritual kg backpack paradoxically helps her to leave behind the emotional consumer experiences. The theorization also challenges the use of weight that she was carrying so far. Turner’s ideas in consumer research. “I had around 11kg in my backpack. When you’re walking with this material weight you are letting behind the emotional weight you Bodily Experiences and Self-transformation were carrying with you before arriving here” (Lara, Camino de EXTENDED ABSTRACT Santiago)

Many people come to Lourdes facing difficulties. They are going Religious sites mobilize about 380 million people per year through very difficult situations in life. However, in Lourdes they around the world, which represents a market of around 400 USD learn their lesson. They return home transformed. (Michael, million (UNWTO, 2018). These results are interesting for the com- Lourdes) prehension of the patterns of consumption of such religions and experiences and how they are connected to well-being and self-trans- Consumer’s senses have been investigated from the provider’s formation. view in consumer research and psychology, such as in the concept of sensory marketing, defined as marketing that engages the consum- Personal Transformation Through Extraordinary ers’ senses and affects their behaviors (Hultén, 2011; Krishna, 2011; Experiences 2012). In consumer culture, incipient line of research has also been dedicated to understanding the role of body and senses in experi- EXTENDED ABSTRACT ences (Scott & Uncles, 2018), such as through the connection with Extraordinary experiences are characterized as uncommon, nature (Canniford & Shankar, 2013), dancing (Hewer & Hamilton, infrequent, emotionally intense (Tumbat and Belk 2011; Arnould 2010), impacted by time (Woermann & Rokka, 2015), or submitted and Price 1993; Scott, Cayla, and Cova 2017), providing a sense of to pain (Scott, Cayla, & Cova, 2017; Cova & Cova, 2019). Yet there community (Arnould and Price 1993; Celsi, Rose and Leigh 1993), is still space to understand how consumers live extraordinary ex- with identity manifestation being the result of belonging (Arnould periences and how they enable consumer transformations and well- and Price 1993). Extraordinary experiences go beyond the realm of being. We believe that it can be better explained when showing that everyday life (Bhattacharjee and Mogilner 2014), bringing a sense the body sensoriality (Kirmayer, 2003; Roux & Belk, 2018; Vom of newness, originality, and freshness, often changing the individ- Lehn, 2006) is shaped by the place (Castilhos, Dolbec, & Veresiu, ual through “…personal integration, personal control, awareness of 2017) and other materials (Miller, 2005), especially during pilgrim- power, joy and valuing…” (Arnould and Price 1993, p. 25), some- age experiences. times even entering into the realm of the spiritual, the magical and How do consumers live pilgrimage experiences seeking self- supernatural (Fernandes and Lastovicka 2011; Schouten and McAl- transformation? This question guided the design of a multi-sited exander 1995). sensory ethnography approach (Pink, 2015) conducted in four differ- Extraordinary experiences allow participants to create a new ent pilgrimages in Europe (Lourdes – France, Camino de Santiago identity facet (Celsi et al. 1993), to “escape from the rationality, – Spain) and Latin America (Juazeiro do Norte and Belém do Pará rules, and stresses of everyday life” (Tumbat and Belk 2011, p. 44), – Brazil). Sensory ethnography is an emplaced ethnography devoted to temporarily invert one’s social order by engaging in new social to understanding an experience by addressing the relationships be- groups with a certain degree of homogeneity (Schouten and McAl- tween bodies, minds, and the materiality and sensoriality of the en- exander 1995; Kozinets 2002). vironment (Pink, 2015). We employed three main techniques of data Often, the kind of impact and transformation that extraordinary collection, which were (i) interviews, (ii) participant observations, experiences provide is portrayed by researchers in a specific time- and (iii) visual data. We carried out data analysis by coding each in- space, located within the experience per se. As illustrated by one of terview, looking for cultural categories and individual and collective Tumbat and Belk’s (2011) alpinists “I think it [the money] would interpretations mainly related to the concepts of embodiment, place only be worth it if I get to the top. I’ve had a really good time, but and materiality (Spiggle, 1994). [reaching the summit is] the only reason I’m actually here” (p. 53). 1030 / Consuming Extraordinary Experiences: Personal Transformationin the Antistructural Realm

Also, we know that it is through post-fact narratives that individu- through extraordinary experiences using the lenses of rites of pas- als make sense of their experience, for instance for those that are sage, with separation from social roles and status, transition into new physically intensive and immersive (Scott et al., 2017). Traveling roles, and incorporation of new selves, roles and status (Schouten to off-track places often leads to self-accomplishment and personal 1991) marking the lives of our informants. development (Zinelabidine et al. 2018). While some researchers have provided some extended effects REFERENCES of extraordinary experiences, such as learning new skills (Arnould Arnould, E. J., & Price, L. L. (1993). River magic: Extraordinary and Price 1993; Celsi et al. 1993), personal growth and renewal (Ar- experience and the extended service encounter. Journal of nould and Price 1993), personal development (Zinelabidine et al. Consumer Research, 20(1), 24-45. 2018), and identity expansion (Arnould and Price 1993; Celsi et al. Askegaard, S., & Linnet, J. T. (2011). Towards an epistemology of 1993), there is still a lack of studies that investigate the process that consumer culture theory: Phenomenology and the context of leads to the enduring consequences of engaging with extraordinary context. Marketing Theory, 11(4), 381-404. experiences. In this research, we expand the literature on the ex- Beverland, M. B., Lindgreen, A., & Vanhamme, J. (Eds.). (2012). traordinary experience by investigating its transformational effects, Memorable Customer Experiences: A Research Anthology. specifically by theorizing how it can act as a rite of passage (Rook Gower Publishing, Ltd.. 1985; Schouten 1991), and as symbolic marks of a permanent status Bhattacharjee, A., & Mogilner, C. (2014). Happiness from ordinary change (Rook 1985). The research question we ask here is how can and extraordinary experiences. Journal of Consumer Research, extraordinary experiences transform individuals? 41(1), 1-17. Methodologically, we investigated the solo travel extraordi- Belk, R. W., & Costa, J. A. (1998). The mountain man myth: nary experience, which we define here as traveling alone for some A contemporary consuming fantasy. Journal of consumer time (although the meaning of the experience is more important than research, 25(3), 218-240. its temporal length) and engaging with meaningful situations that Bonsu, S.K., Belk, R.W. (2003) Do Not Go Cheaply into That Good transform the individual in significant ways. Travel has long been Night: Death-Ritual Consumption in Asante, Ghana, Journal considered in the tourism literature as a rite of passage with transfor- of Consumer Research, 30(1), 41–55. mative capacities (Grabowski et al. 2017; Decrop et al. 2018). The Carù, A., & Cova, B. (2006). How to facilitate immersion in a first author conducted in-depth interviews (Kvale 2008) with sixteen consumption experience: Appropriation operations and service informants who have passed through meaningful experiences while elements. Journal of Consumer Behaviour: An International traveling alone outside of their original country (Brazil), so they Research Review, 5(1), 4-14. could describe their motivations, the most significant situations, their Canniford, R. & Shankar, A. (2013). “Purifying Practices: How feelings, reflections, meanings and the outcomes of the solo travel Consumers Assemble Romantic Experiences of Nature,” experience. Data analysis was first conducted in an inductive man- Journal of Consumer Research, 39(5), 1051–69. ner (Spiggle 1994), coding for the situations, reflections, learnings, Celsi, R. L., Rose, R. L., & Leigh, T. W. (1993). An exploration of and transformations while and after the solo travel. After that, we high-risk leisure consumption through skydiving. Journal of used conceptual elements of rites of passage (Rook 1985; Schouten consumer research, 20(1), 1-23. 1991) to consolidate our theoretical analysis of the transformational Cova, V., & Cova, B. (2019). Pain, suffering and the consumption aspects of the solo travel experience. of spirituality: a toe story. Journal of Marketing Management, Our findings indicate that the main motivation to engage in a 35(5-6), 565-585. solo travel experience is the desire to temporarily depart from the Decrop, A., Del Chiappa, G., Mallargé, J., & Zidda, P. (2018). social structure in which one is immersed and to live functional and “Couchsurfing has made me a better person and the world extraordinary experiences in a new place. The lived experience of a better place”: the transformative power of collaborative solo travel is marked by performance experiences and stochastic ex- tourism experiences. Journal of Travel & Tourism Marketing, periences (Lanier and Rader 2015), which constitute an important 35(1), 57-72. source of reflection and transformation. Our informants learned and Fernandez, K. V., & Lastovicka, J. L. (2011). Making magic: reflected upon ordinary functional instances of the solo travel, such Fetishes in contemporary consumption. Journal of Consumer as taking care of one’s clothes, housing matters, transport, money, Research, 38(2), 278-299. without any support, and getting immersed in new cultures, and upon Glaser, B. G., & Strauss, A. L. (1967), The Discovery of Grounded stochastic (non-functional) experiences. The condition of being solo Theory, Chicago: Aldine. frequently contributed to the reflexive element of the experience, as Grabowski, S., Wearing, S., Lyons, K., Tarrant, M., & Landon, A. it allowed time and space to think and reflect, even when establishing (2017). A rite of passage? Exploring youth transformation new and occasional social ties. Traveling solo is an experience rich and global citizenry in the study abroad experience. Tourism in stochastic situations, such as walking in the city, seeing something Recreation Research, 42(2), 139-149. interesting on the streets, getting to know someone new and forming Hewer, P., & Hamilton, K. (2010). On emotions and salsa: unplanned ties, which leads to unplanned situations, visiting extraor- Some thoughts on dancing to rethink consumers. Journal of dinary nature places, and learning about rituals and practices of other Consumer Behaviour: An International Research Review, 9(2), cultures. The condition of having no previous social ties and space 113-125. knowledge in the new places also contributes to a constant state of Hultén, B. (2011). Sensory marketing: the multi-sensory brand- situation-awareness, immersion, and self-reflection. experience concept. European Business Review, 23(3), 256- In short, solo travel allows the experience of learning - acquir- 273. ing knowledge and skills, playing with possible selves, questioning Kirmayer, L.J. (2003). “Reflections on embodiment”, in Jr, one’s former values and practices, developing new goals and life J.M.W. (Ed.), Social and Cultural Lives of Immune Systems, models, planning how one wants to be and to live. We contribute to Routledge, London, pp. 282-302 the literature by theorizing the process of personal transformation Advances in Consumer Research (Volume 48) / 1031

Kozinets, R. V. (2002). Can consumers escape the market? Thompson, C. J., Locander, W. B., & Pollio, H. R. (1989). Putting Emancipatory illuminations from burning man. Journal of consumer experience back into consumer research: The Consumer research, 29(1), 20-38. philosophy and method of existential-phenomenology. Journal Krishna, A. (2012). An integrative review of sensory marketing: of consumer research, 16(2), 133-146. Engaging the senses to affect perception, judgment and Tumbat, G., & Belk, R. W. (2011). Marketplace tensions in behavior. Journal of Consumer Psychology, 22(3), 332-351. extraordinary experiences. Journal of Consumer Research, Krishna, A. (Ed.). (2011). Sensory marketing: Research on the 38(1), 42-61. sensuality of products. Routledge. Turner, V. (1969) The Ritual Process: Structure and Anti-Structure, Kvale, S. (2008), Doing Interviews. London: Sage. Transaction Publishers, Rutgers- the state university, New Lanier Jr, C. D., & Rader, C. S. (2015). Consumption experience: Jersey. An expanded view. Marketing Theory, 15(4), 487-508. Turner, V. (1974). Liminal to liminoid, in play, flow, and ritual: An Miller, D. (2005). Materiality. Duke University Press. essay in comparative symbology. Rice Institute Pamphlet-Rice Pink, S. (2015). Doing sensory ethnography. Sage. University Studies, 60(3), 53-92. Rook, D. W. (1985). The ritual dimension of consumer behavior. Turner, V. (1982) From Ritual to Theatre- The Human Seriousness Journal of consumer research, 12(3), 251-264. of Play. New York, PAJ Publications. Roux, D., & Belk, R. (2018). The Body as (Another) Place: Turner, V. W., & Bruner, E. M. (Eds.) (1986). The Anthropology of Producing Embodied Heterotopias Through Tattooing. Journal Experience. Urbana: University of Illinois Press. of Consumer Research. United Nations World Tourism Organization (UNWTO). (2018). Schouten, J. W. (1991). Selves in transition: Symbolic consumption UNWTO Global Tourism Dashboard. Retrieved from https:// in personal rites of passage and identity reconstruction. www.unwto.org/global-and-regional-tourism-performance Journal of consumer research, 17(4), 412-425. Vom Lehn, D. (2006). Embodying experience: a video-based Schouten, J. W., & McAlexander, J. H. (1995). Subcultures of examination of visitors’ conduct and interaction in museums. consumption: An ethnography of the new bikers. Journal of European Journal of Marketing, 40 (11/12), 1340-1359. consumer research, 22(1), 43-61. Woermann, N., & Rokka, J. (2015). Timeflow: How consumption Scott, R. O., & Uncles, M. D. (2018). Bringing sensory practices shape consumers’ temporal experiences. Journal of anthropology to consumer research. European Journal of Consumer Research, 41(6), 1486-1508. Marketing. Zinelabidine, B. B. G., Touzani, L., Dahmane, N. B., & Touzani, Scott, R., Cayla, J., & Cova, B. (2017). Selling pain to the saturated M. (2018). How off-track tourists create their own event: a self. Journal of Consumer Research, 44(1), 22-43. customer-dominant logic perspective. Qualitative Market Tarabout, G. (1993). Ritual Rivalry in Kerala. In H. Brückner, L. Research: An International Journal., 21 (4), 549-66. Lutze, & A. Malik, Flags of fame: Studies in South Asian folk culture (pp. 81-108). New Delhi: Manohar. Cognitive Influences on Consumption Experiences Chairs: Adrian Ward, University of Texas at Austin, USA Tito L. H. Grillo, University of Texas at Austin, USA

Paper #1: The “Next” Effect The fourth paper shifts focus from a currently prevalent source Ed O’Brien, University of Chicago, USA of product-related knowledge (Google), to explore an emerging option — virtual reality (VR). Chun, Baek, and MacInnis look at a Paper #2: Unintentional Inception: Why Unintentionality context where knowledge acquisition and consumption experience Increases Quality Perceptions of Artistic Products merge into a single event: product sampling. VR-based sampling ex- Alexander Goldklank Fulmer, Yale University, USA periences facilitate the mental simulation and conjuring of stories Taly Reich, Yale University, USA about the consumption of sensory-rich experiential products (e.g., Paper #3: The Effect of Google-Induced Confidence on luxury hotels), thus increasing behavioral intentions toward them. Consumption Experiences This session demonstrates how complementary facets of cogni- Tito L. H. Grillo, University of Texas at Austin, USA tion – content (i.e., what individuals know) and formation (i.e., how Cristiane Pizzutti, UFRGS, Brazil they obtain knowledge) – influence consumers’ experiences. More- Adrian Ward, University of Texas at Austin, USA over, the session takes a timely perspective on the topic. It takes into Paper #4: Glimpse into the Future Experience: When is Virtual account that product information (e.g., their history [Fulmer and Reality Sampling Beneficial? Reich] and their future [O’Brien]) is increasingly accessible, how Helen Chun, Cornell University, USA consumers tend to access it (Grillo, Pizzutti, and Ward), and the ten- Eunsoo Baek, Hong Kong Polytechic University, China dency of evolving technologies to provide yet new forms of product Deborah MacInnis, University of Southern California, USA knowledge acquisition (Chun, Baek, MacInnis).

SESSION OVERVIEW The “Next” Effect As we meet this year in Paris, a center of the Enlightenment and birthplace of new paradigms to thinking and knowing about the EXTENDED ABSTRACT world, the setting seems particularly appropriate for discussing con- Imagine you are toying with a fun new gadget for the first sumer knowledge. Consumer research has extensively studied how time. Assuming things work as intended, experiencing a fun stimu- knowledge affects decision processes. However, little research has lus you have never experienced before should be highly enjoyable. examined how knowledge shapes experiences. The current papers Yet, consider one small tweak: You learn beforehand that an even contribute to a nascent area of research addressing this important better version is in the works. Despite still proceeding to play with gap. These papers provide insight into the question of, “How can the same gadget, and despite nothing objectively changing within it, differences in consumer knowledge — in terms of both knowledge the current research explores whether mere expectations of a more itself and the way in which knowledge is acquired — influence eval- enjoyable future (e.g., upcoming versions of a technology) might uations and enjoyment of consumption experiences?” The research undermine one’s experience of an enjoyable present (e.g., a current in this session is relevant not just to academics, but also to market- version)—despite that present experience being brand new for you. ers seeking to enhance consumer experiences by delivering the right Why? Existing research might suggest the opposite: Better information, in the right format. future versions of a good should, if anything, foster more positive First, O’Brien draws attention to the consequences of learning impressions of targets subordinately associated with that good (e.g., about the development of an upgraded version of a product. He finds assimilation effects: Bless & Schwarz, 2010). This idea is echoed that knowing that an improved version of the product will be avail- by broader research on positive expectancies (e.g., studies highlight- able in the future affects consumers’ subjective experience with the ing benefits of dispositional optimism: Carver & Scheier, 2014). product in the present. The key finding is that merely knowing this However, the current research advances these literatures by consid- fact has a negative effect on the enjoyment of the experience with the ering the additional possibility of inference-based effects. Indeed, current product. assessing novel goods can be difficult. People reduce ambiguity by Fulmer and Reich also explore how perceptions of the same drawing inferences from other available cues, which can shape their stimulus change as function of what consumers know about it. How- consumption experience (Lee, Frederick, & Ariely, 2006). Put in cur- ever, instead of examining the effect of knowledge about the future rent terms, one salient inference people likely draw from learning of the product, they look at the effect of knowledge about its incep- of future improvement is that there must have been something to tion. In the context of arts consumption, they find that the perceived improve upon (Grice, 1975)—but this need not be true, particularly quality of the same artistic product improves when consumers learn in today’s world of strategic change (e.g., planned obsolescence: that its creation was an unintentional (as opposed to intentional) ef- Gershoff, Kivetz, & Keinan, 2011). Thus, despite being able to draw fort. on one’s bottom-up experience to inform one’s judgment, bottom-up The third paper starts exploring the notion that cognitive influ- experiences may be tainted: Mere knowledge of a better future may ences on experiences arise not only from what consumers know, but prompt people to shift their attention to present problems warranting also from how they obtain information. Grillo, Pizzutti, and Ward the change. Improvement may be self-fulfilling: Merely knowing a test the idea that accessing information about a product category better future is coming may make the present worse (e.g., by prompt- through Google search (compared to the same information without ing people to search for, and alas find, more “bugs”). online search) can subjectively enhance subsequent experiences with Three preregistered experiments tested this hypothesis. In Ex- products of that category. This effect stems from decision confidence periment 1, 806 participants played “Art Time,” a novel art-making boosts induced by Google-based access to information. game. All participants played the same game, dated from the same time, but we manipulated knowledge of upcoming versions: Some

Advances in Consumer Research 1032 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1033 participants first viewed a flyer for “Art Time: October 2016re- improvement across so many domains (Pinker, 2018), our findings lease,” then played “Art Time: October 2016 release” (control); oth- hint at a vicious cycle: Ever-newer offerings are not necessarily ever- ers viewed a visually-identical flyer for “Art Time 2” announcing up- better, yet people may fail to learn from experience if their experi- dates to the original, then played “Art Time: October 2016 release” ence is psychologically tainted at each fresh start. Mere awareness of (better-future); still others viewed a visually-identical flyer for “Paint a better future often may risk spoiling an otherwise satisfying pres- Wars 2” announcing updates to an unrelated stimulus, then played ent. “Art Time: October 2016 release” (unrelated-future). This latter condition isolates inference-based mechanisms (rather than any positive Unintentional Inception: Why Unintentionality Increases future spoiling the present, the future and present may need to be Quality Perceptions of Artistic Products linked). After, all participants rated their enjoyment (dependent variable) and bugs noticed (proposed mechanism). The same game was EXTENDED ABSTRACT less enjoyable (p < .001), and buggier (p < .001), when participants Artistic product creation can be fundamentally intended or were merely aware of a better future (indirect effect via buggyness: unintended from its outset. Presumably, this should not influence b = 0.42, 95% CI = [0.28, 0.56]). Unrelated improvement, however, how the products are evaluated by consumers since their objective affected neither enjoyment nor buggyness (contrasts vs. control: ps qualities are identical. However, research has documented a power- ≥ .273). ful influence of intentionality behind actions in shaping consumer Experiment 2 replicated these effects using a different stimu- preference and perceptions. The current research explores how qual- lus. 309 participants played the same novel virtual-reality “blaster” ity perceptions of artistic products can be influenced by the inten- game. Better-future participants enjoyed the game less, and found it tionality involved in their inception; in other words, whether the buggier, compared to control (ps ≤ .011; mediation, b = 0.32, 95% product was originally intended by the creator. We suggest that be- CI = [0.18, 0.52]). We also tested boundaries: Other participants cause knowledge of unintentional product creation elicits heightened were informed of a better future version, but additionally evaluated counterfactual thought about how the product may have never been a “Blast From The Past!” booklet reminding them of older blaster- created at all, consumers place a quality premium on unintentionally game technology. The effect was attenuated with converse reminders created artistic products. of the past, by reducing perceived bugs (contrast versus control: p = Prior literature largely indicates that intention in creation can .206, d = 0.18). heighten quality perceptions of a finished product. Intention and Experiment 3 replicated these effects using a third stimulus, and perceptions of effort are psychologically linked (Caruso, Waytz, and tested consequences. 487 participants tested the same novel 360-de- Epley 2010; Heider 1958; Malle 2010; Malle and Knobe 1997), and gree technology by which viewers can rotate the camera at any angle perceptions of effort have been shown to heighten quality percep- while a video plays. Better-future participants again reported a less tions of artistic products such as effortfully created paintings and po- enjoyable, and buggier, experience (ps ≤ .005; mediation, b = 0.25, ems (Cho and Schwarz 2008; Kruger et al. 2004). This influence of 95% CI = [0.17, 0.36]). Moreover, all participants were informed perceived effort in product creation on perceptions of quality, and the they had to re-experience the same technology, but could buy out association between intention and effort (Caruso et al. 2010; Heider of this task. Better-future participants wagered double to opt out (p 1958; Malle 2010; Malle and Knobe 1997), suggests that consumers = .014). Their “tainted” experience—for which they were willing may place a quality premium on intentionally created artistic prod- to pay a real cost to avoid repeating—was purely a product of their ucts over unintentionally created artistic products. beliefs. Despite work showcasing such positive influences of inten- These experiments reveal a perverse effect of positive future tionality and effort on quality perceptions, unintentionality has been expectations: The same stimulus is less enjoyable merely when shown to heighten consumer preference for products in certain con- participants believe a better future version is in the works. This oc- texts. Newman and Bloom (2012) demonstrated that near duplicates curs because knowledge of better future is, in effect, self-fulfilling: of paintings were considered more valuable when they resembled an It spurs people to discover more problems in the present that must original piece of artwork by mere coincidence compared to an in- have needed fixing. tentional attempt to copy. Furthermore, when consumers perceive a These findings highlight the need for more research on how company as intentionally creating an environmentally friendly prod- future circumstances serve as a standard for present circumstances uct, they are less interested in purchasing it due to a belief that the (for a review, see Strahan & Wilson, 2006). We bridge these ideas company devoted resources away from other product attributes to with yet-disconnected literatures on expectancy effects, suggesting make it environmentally friendly (Newman, Gorlin, and Dhar 2014). new insights into when and why, exactly, positive expectations can Especially relevant to the current research, Reich, Kupor, and Smith backfire. As reviewed by Lee et al. (2006), labels affect experiences: (2017) demonstrate that consumers prefer hedonic products resulting Drinking a “vinegar” beer makes the same beer taste worse; eating a from a mistake in the creation process to otherwise identical prod- “fatty” dessert makes the same dessert taste more decadent; and so ucts made intentionally. We build upon these lines of research by on. However, past findings typically stop at the moment of choice. demonstrating that mere unintentionality in an artistic product’s in- As Lee et al. (2006) note: “A third question concerns how specific ception heightens quality perceptions of the final product. perceptual, attentional, and cognitive mechanisms mediate the effect We base our theorizing on the well-documented “intentionality of expectations on experience” (p. 1057). To this point, we provide bias”, according to which people have a tendency to assume that finer-grained evidence for why expectations change preferences: actions are intentional (Bègue et al. 2010; Rosset 2008; Rosset and Mere awareness of a better future changes attention and prompts Rottman 2014; Spunt, Meyer, and Lieberman 2015). Because un- people to actively search for flaws. intentional actions diverge from this assumption, and events con- Finally, future research should extend beyond technology. Peo- sidered unusual have been shown to inspire counterfactual thought ple with positive views of their future self may make wiser plans (Kahneman and Miller 1986), we propose that unintentional creation (Hershfield, 2011), but also discover more flaws in their present self inspires greater counterfactual thought than intentional creation. Re- and thus feel less happy at any one moment. Surrounded by so much search has demonstrated a variety of ways in which counterfactual 1034 / Cognitive Influences on Consumption Experiences thinking can benefit perception (Kray et al. 2010; Teigen and Glad The Effect of Google-Induced Confidence on 2011). For example, Kray et al. (2010) demonstrate that counterfac- Consumption Experiences tual thinking can actually increase the perceived meaningfulness of one’s own life experiences by leading to greater recognition of posi- EXTENDED ABSTRACT tive life events and perceptions that these events were fated. Further, People increasingly turn to the Internet for information on vir- research by Medvec, Madey, and Gilovich (1995) indicates that tually every topic. In consumer research, scholars have primarily counterfactuals can change the perception of objective qualities of focused on how the unparalleled access to information provided by an outcome. Medvec et al. (1995) found that bronze medalists in the the Internet carries consequences for decision processes or decisions Olympics were judged as looking happier than silver medalists be- themselves (e.g., Bhargave, Mantonakis, & White, 2016; Lynch & cause the counterfactual of not having won a medal at all was more Ariely, 2015). The current research shifts this focus to explore con- salient to bronze medalists, heightening perceptions of their outcome sequences to consumption experiences. Specifically, we propose that above those of the objectively superior silver medalists’. Similarly, confidence stemming from online search exerts a positive top-down we predict consumers will perceive an unintentionally created artis- influence on chosen consumption experiences, thereby enhancing tic product as higher quality than an otherwise identical intentionally subjective evaluations of objectively identical stimuli. created product because the counterfactual of an unintentionally cre- Our hypothesis builds on and contributes to two distinct streams ated product never having been created should be especially salient of prior research: one examining the metacognitive errors induced to them. by online search, and a second investigating the impact of expecta- Four studies examined the quality premium offered to unin- tions on experiences. A growing body of research finds that people tentionally created artistic products in contexts in which consum- misattribute online search results to their own memories; as a conse- ers actually experienced the products. Our first study demonstrated quence, search tools like Google may inflate people’s confidence in that consumers perceive an unintentionally created poem’s quality their own cognitive skills (Ward, 2013; Wegner & Ward, 2013). We as superior to that of an otherwise identical intentionally created extend this research by suggesting that when online search is used poem. Our second study demonstrated the robustness of the effect to access information related to a product category (such as when of unintentional artistic creation on heightened quality perceptions, consumers prepare for upcoming decisions or experiences), this extending it to the domain of visual art. Our third study identified phenomenon causes consumers to be more confident in their own that heightened quality perceptions of unintentionally created artis- decision-making abilities. We further suggest that Google-induced tic products are mediated by increased counterfactual thinking about confidence can exert a positive top-down influence on evaluations how a product’s creation might never have occurred at all. Further, it of chosen options. Prior research on top-down effects shows that ruled out an alternative explanation that this phenomenon is driven expectation-inducing factors can cause subjective experiences with by a magical belief in divine intervention playing a role in unin- the same objective stimuli to conform to the expectation (e.g., Lee, tentionally created artistic products. Through a process by modera- Frederick, & Ariely, 2006). This type of effect is typically observed tion design in which participants evaluated a marketing slogan, our in response to information from an external source (e.g., telling fourth study showed that quality perceptions of intentionally created someone that movie will be funny, or that a beer has been spiked artistic products rise to the level of quality perceptions of uninten- with a weird ingredient). We extend this line of research by arguing tionally created artistic products when consumers are encouraged to that confidence in one’s own decision skills functions as an internal consider how an intentionally created product might never have been source of positive expectations. Thus, factors that affect confidence created at all. — including Google-induced metacognitive errors — may improve This research is the first to reveal a quality premium given to subjective experiences by creating self-fulfilling expectations re- unintentionally created artistic products, identifying greater counter- garding the quality of chosen options. factual thought about how such products might never have been cre- Briefly, we propose that Internet search improves subjective ated at all as a hereto unrecognized antecedent of consumer quality experiences through a process mediated by confidence in decision perceptions. Consequently, it builds on previous work on benefits skills. Study 1 (N = 465, MTurk) tested the first part of this process. of unintentionality in product creation (Newman and Bloom 2012; Participants were randomly assigned to complete a cinema-related Newman et al. 2014; Reich et al. 2017) by illustrating in multiple trivia challenge (sample question, “What was the first fully comput- artistic domains that mere unintentionality in the inception of an er-animated movie ever?”) using Google vs. not using Google. In the artistic product’s creation colors how the quality of the product is Google condition, they verified their answers using Google; in the no perceived. Google condition, they received the same answers their counterparts These results could benefit artists by making them aware of what would find on Google but without engaging in online search. Follow- information they should include and emphasize about the creation of ing this manipulation, we measured participants’ confidence in their their artwork in advertisements in order to heighten perceptions of abilities to choose movies using a validated scale of confidence in their artwork’s quality. Additionally, organizations based around the consideration-set formation skills (Bearden, Hardesty, & Rose, 2001 appreciation of artwork such as museums could benefit from empha- – sample item: “I am confident about my ability to recognize options sizing unintentionality in the inception of their artworks’ creation worth considering when deciding to watch a movie”). We found that wherever possible. By positively coloring customers’ perceptions of online search increased confidence in decision abilities, M = 5.59 vs. artistic products, museums and marketing managers could heighten M = 5.29, t(461) = 3.17, p = .002. foot traffic and increase purchases of artistic products. Study 2 (N = 306, MTurk) tested the overarching hypothesis that online search enhances subjective experiences. After completing the same Google manipulation used in Study 1, participants chose a short film to watch as part of the study. Three options were provided, all of which were pretested to match on perceived quality. Film-choice was not different across conditions, χ2(2) = 1.53, p = .466, indicating that choices were not influenced by the manipula- Advances in Consumer Research (Volume 48) / 1035 tion. After watching the film of their choice, participants were asked Building on prior research on dynamic presentations induc- how much they enjoyed the film they watched. Despite the fact that ing greater mental involvement (Roggeveen et al. 2015) and this, in participants in both conditions watched the same films, those who turn, further increasing elaboration (e.g., Keller and McGill 1994), used Google reported greater film enjoyment, M = 5.22 vs. M = 4.83, we propose that sampling a future experience through VR (vs. static t(304) = 2.17, p = .031. images) enhances consumers’ involvement with the mediated envi- Study 3 (N = 329, MTurk) connects the results of studies 1 ronment and the extent to which consumers conjure up stories about and 2. After completing the trivia challenge (same as in the previ- the future experiences. These imagined stories are likely to entail ous studies) and before choosing the film, we measured participants’ conscious representations of the self-experiencing future consump- confidence in decision skills by asking them “how confident are you tion situations (Walter et al. 2007), positively influencing behavioral that you will be able to identify and choose the best film available intentions toward them. Importantly, we further propose a more nu- among the options provided?” (0 = Not confident at all, 100 = 100% anced view of the effect of VR—the extent to which the effect of VR confident). They then watched the film of their choice and reported is maximized vs. undermined (i.e., boundary conditions of VR effec- film enjoyment. Confidence in decision skills mediated the Google tiveness) depends on sensory richness of the target experience, indi- effect on enjoyment, b = 0.17, 95% bootstrap (10,000 simulations) vidual differences in imagery vividness, and the perspective-taking CI = [0.05, 0.36]. prompted during the sampling process. We conducted four empirical In Study 4 (N = 499, MTurk), we further examined the role studies across various experiential contexts (retail, hotel stay, tour) of confidence in decision skills using moderation. Theoretically, to test these ideas. confidence in one’s ability to choose should only enhance experi- A pilot study (N = 108) utilizing a retail content (NYC Nike ences that one chooses. We therefore added a “choice” manipula- flagship store in New York City) found that, compared to 360-de- tion to our study design. For participants in the choice condition, gree tour and static photo images, VR significantly increased mental the study design replicated that of Study 2. Participants in the no- involvement (but no difference between 360-degree tour and static choice condition, however, were randomly assigned to a short film photo images), subsequently enhancing future store visit intentions. (instead of choosing one). If the increased enjoyment in studies 2 and Building on the initial pilot results, Study 1 (N = 188) investigated 3 stemmed from self-fulfilling expectations regarding one’s ability the role of sensory richness embedded in the experience by utilizing to choose the best option, the subjective experience of participants the content marketing materials posted on Youtube by actual hotel in the no-choice condition — who do not have a chance to employ companies. With a 2 (sensory richness: luxury vs. economy-plus ho- their decision-making abilities — should not be affected by online tel experience) x 2 (medium: VR vs. photo images) between-subject search. We found a significant Google × Choice interaction,b = 0.51, design, we found that compared to photo images, sampling an ex- t(494) = 1.97, p = .049 (controlling for participants’ general liking of perience through VR heightened mental involvement only for the short films). When participants chose the film, Google increased film sensory rich experience (i.e., luxury hotel tour). No difference was enjoyment M = 5.26 vs. M = 4.82, t(494) = 2.41, p = .016. However, found between the two medium conditions when the target experi- when participants watched a random film, Google did not predict ence lacked high sensory richness (i.e., economy-plus hotel tour), enjoyment, M = 4.91 vs. M = 4.97, t(494) = 0.35, p = .723. rendering no advantage of VR over a more traditional medium of The current studies illuminate a process where Internet search browsing static photo images. We further found support for an indi- boosts decision-making confidence and, thus, ignites a self-fulfilling rect serial mediation whereby mental involvement and the ability to prophecy that results in improved subjective experiences. Market- imagine future consumption stories sequentially mediated the effect ers may benefit from this phenomenon by developing marketing of VR on behavioral intentions only when the experience was high communication that encourages consumers to learn about products on (vs. lacked) sensory richness. through Internet search. However, we acknowledge that to better un- Study 2 (N = 188) further demonstrates that the extent to which derstand how Google-induced confidence can be explored in mar- VR enhances consumers’ mental involvement and the ability to con- keting practice and how it affects consumers’ lives, future research jure up stories about the future consumption experience depends on should study the interplay of this boosted confidence with other “real individual differences in imagery vividness (VVIQ; Marks 1973). world” factors – e.g., distractions, stress – that might inhibit/inten- If VR boosts consumers’ mental involvement of the experience and sify its effects. elaboration on future consumption events, this facilitative effect of VR should be more pronounced for individuals with low imagery Glimpse into the Future Experience: When is Virtual vividness. By providing a sampling experience of a gondola ride in Reality Sampling Beneficial? Venice, Italy, we found that the enhancing effect of VR occurred for those who are low on imagery vividness, but not for those who are EXTENDED ABSTRACT high on imagery vividness. Those who are high on imagery vivid- Rapid adoption of virtual technologies is opening up new op- ness demonstrated a high level of mental involvement and the ability portunities, enabling consumers to experience products and destina- to imagine consumption stories on their own even with static photos. tions and acquire information in an immersive manner. Although an Study 3 (N = 225) demonstrates that the perspective-taking ad- increasing number of companies are designing immersive VR con- opted during the experience also matters in maximizing the effect of tents to engage consumers across various touchpoints of customer VR. We worked with a media content production company to create journey, they do so without clear understanding of when and how the content that took predominantly a first-person vs. third-person VR technology would actually benefit consumer experience and ex- perspective, filming a zipline local activity in Maui, Hawaii for the ert its intended positive effect. Given the technological investment study stimuli. We found that the immersive content that took more and financial costs required for creating and distributing the immer- of a first-person perspective (i.e., experiencer), as opposed to a third- sive marketing content, companies need to be better informed about person perspective (i.e., observer), facilitates the mental involve- both the theoretical underpinnings of VR technology from the con- ment and conjuring up of future consumption stories, influencing sumer psychology perspective and its practical implications. behavioral intentions. 1036 / Cognitive Influences on Consumption Experiences

Across all three main studies, we find support for an indirect ef- Keller, Punam Anand, and Ann L. McGill (1994), “Differences in fect whereby mental involvement and the ability to evoke future con- the relative influence of product attributes under alternative sumption stories sequentially mediate the effect of VR on behavioral processing conditions: Attribute importance versus attribute intentions. Taken together, VR’s facilitating effect is limited when ease of imagability,” Journal of Consumer Psychology, 3 (1), the target experience lacks sensory richness, when consumers tak- 29-49. ing a virtual tour are high in imagery vividness, and when the third- Kray, Laura J., Linda G. George, Katie A. Liljenquist, Adam person perspective is promoted in the sampled experience. These D. Galinsky, Philip E. Tetlock, and Neal J. Roese (2010), findings not only provide insights into the mechanism through which “From What Might Have Been to What Must Have Been: the facilitative effect of VR manifests but also provides practical im- Counterfactual Thinking Creates Meaning,” Journal of plications for designing marketing content for the maximal effect. Personality and Social Psychology, 98 (1), 106–18. Kruger, Justin, Derrick Wirtz, Leaf Van Boven, and T. William REFERENCES Altermatt (2004), “The Effort Heuristic,”Journal of Bless, H., & Schwarz, N. (2010). Mental construal and the Experimental Social Psychology, 40 (1), 91-8. emergence of assimilation and contrast effects: The inclusion/ Lee, L., Frederick, S., & Ariely, D. (2006). Try it, you’ll like it: The exclusion model. In M. P. Zanna (Ed.), Advances in influence of Experimental Social Psychology, Vol. 42. (pp. 319–373). San expectation, consumption, and revelation on preferences for Diego, CA: Elsevier Academic Press. beer. Psychological Science, 17, 1054-1058. Bearden, W., Hardesty, D., & Rose, R. (2001). Consumer Lynch, J., & Ariely, D. (2015). Wine Online: Search Costs and Self-Confidence: Refinements in Conceptualization and Competition on Price, Quality, and Distribution. Marketing Measurement. Journal of Consumer Research, 28(1), 121-134. Science, 19(1), 83-103. Bègue, Laurent, Brad J. Bushman, Peter R. Giancola, Baptiste Malle, Bertram F. (2010), Intentional Action in Folk Psychology, A Subra, and Evelyn Rosset (2010), “There is No Such Thing as Companion to the Philosophy of Action, 357-65 an Accident, Especially When People are Drunk,” Personality Malle, Bertram F., and Joshua Knobe (1997), “The Folk Concept of and Social Psychology Bulletin, 36 (10), 1301-04. Intentionality,” Journal of Experimental Social Psychology, 33 Bhargave, R., Mantonakis, A., & White, K. (2016). The Cue-of- (2), 101-21. the-Cloud Effect: When Reminders of Online Information Marks, David F. (1973), “Visual imagery differences in the recall of Availability Increase Purchase Intentions and Choice. Journal pictures,” British journal of Psychology, 64 (1), 17-24. of Marketing Research, 53(5), 699-711. Medvec, Victoria Husted, Scott F. Madey, and Thomas Gilovich Caruso, Eugene M., Adam Waytz, and Nicholas Epley (2010), “The (1995), “When Less is More: Counterfactual Thinking Intentional Mind and the Hot Hand: Perceiving Intentions and Satisfaction Among Olympic Medalists,” Journal of Makes Streaks Seem Likely to Continue,” Cognition, 116 (1), Personality and Social Psychology, 69 (4), 603-10. 149-53. Newman, George E., and Paul Bloom (2012), “Art and Carver, C. S., & Scheier, M. F. (2014). Dispositional optimism. Authenticity: The Importance of Originals in Judgments of Trends in Cognitive Sciences, 18, 293-299. Value,” Journal of Experimental Psychology: General, 141 Cho, Hyejeung, and Norbert Schwarz (2008), “Of Great Art (3), 558-69. and Untalented Artists: Effort Information and the Flexible Newman, George E., Margarita Gorlin, and Ravi Dhar (2014), Construction of Judgmental Heuristics,” Journal of Consumer “When Going Green Backfires: How Firm Intentions Shape Psychology, 18 (3), 205–11. the Evaluation of Socially Beneficial Product Enhancements,” Ersner-Hershfield, Hal, Adam D. Galinsky, Laura J. Kray, and Journal of Consumer Research, 41 (3), 823-39. Brayden G. King (2010), “Company, Country, Connections: Pinker, S. (2018). Enlightenment Now: The Case for Reason, Counterfactual Origins Increase Organizational Commitment, Science, Humanism, and Progress. New York, NY: Viking Patriotism, and Social Investment,” Psychological Science, 21 Press. (10), 1479-86. Reich, Taly, Daniella M. Kupor, and Rosanna K. Smith (2017), Gershoff, A. D., Kivetz, R., & Keinan, A. (2012). Consumer “Made by Mistake: When Mistakes Increase Product response to versioning: How brands’ production methods Preference,” Journal of Consumer Research, 44 (5), 1085-103. affect perceptions of unfairness.Journal of Consumer Roggeveen, Anne L., Dhruv Grewal, Claudia Townsend, and Ram Research, 39, 382-398. Krishnan (2015), “The impact of dynamic presentation format Grice, H. P. (1975). Logic and conversation. In P. Cole and J. L. on consumer preferences for hedonic products and services,” Morgan (Eds.), Syntax and Semantics 3: Speech Acts. New Journal of Marketing, 79 (6), 34-49. York, NY: Academic Press. Rosset, Evelyn (2008), “It’s No Accident: Our Bias for Intentional Heider, Fritz (1958). The Psychology of Interpersonal Relations, Explanations,” Cognition, 108 (3), 771-80. New York: Wiley. Rosset, Evelyn, and Joshua Rottman (2014), “The Big ‘Whoops!’ Hershfield, H. E. (2011). Future self-continuity: How conceptions in the Study of Intentional Behavior: An Appeal for a New of the future self transforms intertemporal choice. Annals of Framework in Understanding Human Actions,” Journal of the New York Academy of Sciences, 1235, 30-43. Cognition and Culture, 14 (1-2), 27-39. Kahneman, Daniel, and Dale T. Miller (1986), “Norm Theory: Spunt, Robert P., Meghan L. Meyer, and Matthew D. Lieberman Comparing Reality to its Alternatives,” Psychological Review, (2015), “The Default Mode of Human Brain Function Primes 93 (2), 136-53. the Intentional Stance,” Journal of Cognitive Neuroscience, 27 (6), 1116-124. Advances in Consumer Research (Volume 48) / 1037

Strahan, E. J., & Wilson, A. E. (2006). Temporal comparisons, Walters, Gabrielle, Beverley Sparks, and Carmel Herington, (2007), identity, and motivation: The relation between past, present, “The effectiveness of print advertising stimuli in evoking and possible future selves. In C. Dunkel & J. Kerpelman elaborate consumption visions for potential travelers,” Journal (Eds.), Possible selves: Theory, research, and applications (pp. of Travel Research, 46 (1), 24-34. 1-15). New York, NY: Nova Science Publishers. Ward, A. (2013). One with the Cloud: Why People Mistake Teigen, Karl Halvor, and Kristin A. Glad (2011), “‘It Could have the Internet’s Knowledge for Their Own. Unpublished been Much Worse’: From Travelers’ Accounts of Two Natural Dissertation. Hardvard University. Disasters,” Scandinavian Journal of Hospitality and Tourism, Wegner, D. M., & Ward, A. F. (2013). How Google is Changing 11 (3), 237-49. Your Brain. Scientific American, 309(6), 58-61. Experiences vs . Material Goods: What Drives Consumers to Perceive Experiences and Material Goods Differently? Chair: Pureum Kim, University of Arizona, USA

Paper #1: The Bidirectionality of Experiences and Happiness: how consumers express their consumption stories for experiential Happiness Leads People to Perceive Purchases as More and material goods. The authors demonstrate that consumers adopt Experiential than Material a more confident tone and use more emotional words when writing Hyewon Oh, University of Illinois at Urbana Champaign, USA online reviews about experiential purchases. Joseph K. Goodman, The Ohio State University, USA The four papers in this special session extend our knowledge of Incheol Choi, Seoul National University, South Korea the drivers of “experiential advantage” and the consequences of the differences between experiential and material goods for consumer Paper #2: Happiness from Unique Purchases behavior, such as acquiring and writing online reviews. Considering Charlene K. Chu, Chapman University, USA the increasing consumer demand in acquiring life experiences, we Suzanne B. Shu, University of California, Los Angeles, USA expect that this session will be of interest to a wide range of consum- Paper #3: Distance Matters: What Information do People Want er researchers, particularly those examining experiential consump- When Making Experiential and Material Purchases? tion, word of mouth, and consumer well-being. Iñigo Gallo, IESE Business School, Spain Tom Gilovich, Cornell University, USA The Bidirectionality of Experiences and Happiness: Paper #4: Why Are Reviews of Experiential Purchases Less Happiness Leads People to Perceive Purchases as More Credible? Experiential than Material Pureum Kim, University of Arizona, USA Anastasiya Pocheptsova Ghosh, University of Arizona, USA EXTENDED ABSTRACT Martin Reimann, University of Arizona, USA How consumers perceive the nature of purchases has significant impact on how they react to product appeals, and ultimately their SESSION OVERVIEW purchase decisions. One dimensions of purchases that prior research Consumers often prefer to acquire life experiences over materi- has identified is material versus experiential (Gilovich, Kumar, and al goods, because they derive greater happiness from such purchases Jampol 2015). In particular, prior research has demonstrated that (Van Boven and Gilovich 2003). Much of the follow-up work on this consumers derive more satisfaction and happiness when they con- notion of experiential consumption has focused on demonstrating sume experiences compared to material goods (Nicolao, Irwin, and that experiential consumption contributes to better post-consumption Goodman 2009; Van Boven and Gilovich 2003), suggesting that con- outcomes and consumer preferences (Van Boven and Gilovich 2003). sumers should be spending more of their discretionary income on However, less is known about the mechanisms that drive consum- experiences instead of material goods to increase happiness (Dunn, ers to perceive experiential and material purchases differently and Gilbert, and Wilson 2011). In this research, we investigate whether how such differences affect consumer behavior. In this session, we part of this puzzle may be backwards: Instead of experiences lead- address this gap in the literature, by examining the following ques- ing to more happiness, could happiness lead to more experiences? tions: what are the antecedents and drivers of previously identified In other words, do happy consumers view their purchases as more “experiential advantage?” And how do consumers seek and express experiential and less material? information related to experiential vs. material goods? As we demon- Many relationships in psychology are bidirectional (e.g., power, strate, the questions investigated are important for many marketing choice, and control, Inesi et al. 2011) and related research hints at stakeholders, including consumers and managers. the possibility that happy people may view their purchases more The first two papers of the session explore the antecedence of experientially. Positive affect facilitates cognitive flexibility (Isen the “experiential advantage.” In the first paper, Oh, Goodman, and 1987) and broadens thought-action repertoires (Fredrickson 1998; Choi demonstrate that the causal direction of the experiential advan- Fredrickson and Branigan 2005) such that it broadens the range of tage can be reversed by manipulating consumers’ happiness: happier potential actions people are prepared to do. Given that experiences consumers perceive their purchases as experiential. Further, they find are “doing” and thinking about what to do rather than to have (Van that the effect of happiness on experiential perception is eliminated Boven and Gilovich 2003) and that there is the association between when consumers evaluate the purchase with a concrete mindset. In positive affect and broadened thoughts toward action, we expect that the second paper, Chu and Shu identify uniqueness of purchase as happy participants (both measured and manipulated) will perceive one of the driving forces of the “experiential advantage.” They show their purchases as more experiential than material compared to those that consumers find experiential purchase to be more self-defining less happy. Further, to examine our process, we test whether ma- and promote a greater sense of personal uniqueness. nipulating one’s mental representation would moderate the effect of The next two papers turn to how the differences in perceiving happiness on experiential perceptions. Specifically, we expect that experiential and material purchases affect consumer behavior. In the the effect of happiness will be attenuated when a concrete mental third paper, Gallo and Gilovich investigate which information sourc- representation is induced. es consumers are more likely to rely on depending on whether the Studies 1 and 2 test our main hypothesis with chronic happi- purchase is material or experiential. They find that when evaluating ness, examining whether consumers high on subjective well-being experiential purchases, consumers rely more on reviews from those (SWB) perceive purchases as more experiential (vs. material) than who have a closer social connection, but for material purchases, those low on SWB. Study 1 examined consumers’ own purchases consumers are more likely to rely on reviews from aggregate on- using the experience sampling (Csikszentmihalyi and Larson 1987). line reviews. In the fourth paper, Kim, Ghosh, and Reimann examine We measured SWB by creating an index based on satisfaction with

Advances in Consumer Research 1038 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1039 life and PANAS and messaging participants five times a day by a Happiness from Unique Purchases smartphone notification for 7 days. Participants also identified their most important purchase for the day and rated how experiential or EXTENDED ABSTRACT material they perceived their purchase (sliding scale 0 =definitely Consumer research to date in the area of subjective well-being material to 100 = definitely experiential). An HLM model (control- has focused on the experiential advantage, the finding that people ling for demographics and the purchase category), found a positive derive greater enduring happiness from discretionary spending on relationship between SWB and experiential-material perceptions (γ experiences than material purchases (Nicolao, Irwin, and Goodman = 0.23, t(206) = 1.99, p = .048). In other words, the happier the re- 2009; Van Boven and Gilovich 2003). While a number of reasons un- spondent was, the more experiential he or she perceived their pur- derlying the experiential advantage have been proposed and exam- chases. ined, little work has been done on how the uniqueness of experiential Study 2 examined purchases made by others. Participants rated purchases may be contributing to this sense of happiness. 37 different purchase items [varying from more tangible (e.g., wal- According to (Maslow, 1954), uniqueness is foundational to let), less tangible (e.g., a movie ticket), and were ambiguous (e.g., and even definitional of self-actualization, the highest level of hu- television)] in terms of the extent to which each purchase was experi- man need fulfillment at which individuals realize their full potential. ential or material on a 9-point scale (1=definitely material, 5=equally Recent research further suggests that believing oneself to be unique material and experiential, 9=definitely experiential; counterbal- is positively associated with subjective well-being. Individuals with anced). Consistent with study 1, there was a positive relationship a higher sense of personal uniqueness demonstrate higher levels of between SWB and material-experiential rating (r = .24, p = .001). optimism, hope, and resilience, as well as greater life satisfaction In other words, the happier participants were, the more experiential (Şimşek and Yalınçetin 2010). As possessions and experiences come they thought a purchase was in general. to be incorporated into our extended self (Belk 1988), our purchases Study 3 provides causal evidence by manipulating positive may also contribute to our sense of uniqueness and therefore en- (happy), negative (sad), and neutral (control) affect. Participants first hance our subjective well-being. Despite the richness in the market- wrote about three recent events (either happy, sad, or neutral, de- ing literature documenting the pursuit of uniqueness through con- pending on condition) in their life in a very detailed manner. Then sumption, there is little on the subject directly examining the link participants rated 14 purchases as more experiential or more mate- between unique consumption and happiness. rial. As expected, happy participants perceived purchases as more In this research, we directly examine the role of uniqueness in experiential (M = 4.87) compared to those in the sad (M = 4.56; the experiential advantage. In a series of five studies, we show that t(356) = 2.31, p = .022) and control conditions (M = 4.58; t(356) = more unique purchases lead to greater happiness. Situating our inves- 2.15, p = .032). No difference was found between the control and sad tigation within the existing literature on the experiential advantage, conditions (p = .87). we directly examine whether experiential purchases are thought of as Study 4a tested our mechanism by manipulating mental repre- more unique than material purchases and whether the greater unique- sentation in a 2(affect: happy vs. neutral) × 2(representation: con- ness of experiential purchases compared to material purchases concrete vs. abstract) between subjects design. In the first part, 287 par- tributes to the experiential advantage. An exploration into potential ticipants listed three ways of spending with the goal of advancing explanations for why people derive greater happiness from unique their happiness and enjoyment in life ($10-$1,000). In the second purchases reveals that more unique purchases are also more self- part, participants wrote about either a happy event or an event from defining and contribute to a greater sense of personal uniqueness. yesterday (depending on condition). The concrete condition focused Study 1 confirmed that experiences are generally regarded as on the ‘how’ aspect of the event, whereas those in the abstract con- more unique than material purchases; 50 participants in MTurk were dition focused on the ‘why’ aspect (Vallacher and Wegner 1987). given brief definitions of material and experiential purchases, de- Results showed the predicted affect by mindset interaction F(1,283) scribed as in Van Boven and Gilovich (2003). They were then asked = 5.67, p = .018). In the abstract condition, happy participants still to briefly describe a recent purchase of $100 or more and made rated their purchase more experientially (M = 5.89) than control (M with the intention of advancing happiness and enjoyment, a prompt = 4.91; t(283) = 3.02, p = .003). In the concrete condition, however, also adapted from the initial study. Compared to material purchas- the impact of happiness on experiential perceptions was eliminates, experiential purchases were rated as significantly more unique ed. There was no difference in experiential perception between the (M=5.26 vs. M=3.83, t=2.751, p=.008), but not significantly more happy (M = 4.95) and control condition (M = 5.05; t(283) = 0.31, p important (M=5.68 vs. M=4.83, t=1.892, p=.065) or more expensive = .75). Study 4b measured subjective well-being (instead of manipu- (M=$426.53 vs. M=$451.54, t=.117, p=.907). lating happiness) and replicated this interaction, providing further Studies 2 and 3 replicate and extend the original investigation evidence of this process. by Van Boven and Gilovich (2003) on the experiential advantage, In sum, these results suggest that happy consumers view their and demonstrate that the greater uniqueness of experiences medi- purchases as more experiential than their less happy counterparts. By ates greater happiness from spending on experiences compared to examining the other side of experiential advantage, the current find- material goods. In Study 2 (n=100), we survey individuals about ei- ings help consumer researchers better understand the potential bi- ther experiential or material purchases and find a significant indirect directionality between happiness and experiential purchases. While effect for the mediating role of uniqueness on experiential goods’ this research does not rule out an experiential advantage (i.e., that perceived contribution to happiness (indirect effect=.1803, standard experiential purchases can lead consumers to feel more happiness), it error=.1076, 95% CI [.0161, .4441]). Study 3 (n=500) further tests does show that the reverse is true too: Happiness can lead consumers the mechanism by examining the role of uniqueness in driving self- to feel that their purchases are more experiential. Further, the results definition; we find in a serial mediation that the indirect effectof have important implications for how marketers may need to adjust purchase type on happiness in life is significant through uniqueness their appeals based on consumers’ perceptions of a purchase. increasing self-definition (indirect effect=.1288, SE=.0396, 95% CI [.0609, .2147]). 1040 / Experiences vs. Material Goods: What Drives Consumers to Perceive Experiences and Material Goods Differently?

Finally, in Studies 4 and 5, we look at how this relationship EP to be more of a reflection of their true self than their MP (Carter between uniqueness and self-definition helps to explain why spend- & Gilovich, 2012), and so they should be more interested in consulting on experiences affords greater happiness than spending on ma- ing information sources that can take into account their idiosyncratic terial objects. In study 4, we explicitly manipulate the uniqueness preferences. In contrast, when evaluating MP, consumers will rely of the purchase reported by participants to further examine whether more on information sources that are seen as more objective, such as purchase uniqueness drives purchase happiness. A sample of 424 aggregate online reviews or the opinions of experts. MTurkers were instructed to describe a recent material or experiential In Study 1, participants considered the 26 Amazon categories purchase that was unique or similar to other purchases. We find that from Dai et al. 2020. Half of them rated how useful they would find experiential purchases produce greater happiness in life compared Amazon reviews when making a decision about a purchase in that to material purchases when the purchases are not unique (M=7.25 category; the other half rated how useful they would find the opin- vs. M=6.64; t=2.510, p=.013), but that this experiential advantage ions of close others in making such a decision. As expected, we rep- disappears when purchases are unique (M=7.41 vs. M=7.19, t=.967, licate Dai et al.’s result for the Amazon reviews (i.e., they are rated p=.335). Replicating Study 3, self-definition mediates the effect of as less helpful for more experiential categories) but find the opposite purchase uniqueness on happiness in life. Study 5 (n=215) uses a pattern for information from close friends and family (i.e., such indifferent study design in which participants described any recent pur- formation is seen as more helpful for more experiential categories). chase, and we find that the indirect effect of experiential purchase In Study 2, participants read a scenario in which they were type on happiness in life is significant through personal uniqueness encouraged to imagine they were considering making a number of increasing self-definition (indirect effect=.0272, SE=.0119, 95% CI purchases (12 in total, half material and half experiential) and that [.0074, .0538]). they were looking for advice. They had access to two sources of Taken together, our results show that the benefits of experiential information: (a) standard on-line summary reviews by previous cus- purchases on happiness come primarily through the fact that expe- tomers, and (b) a new smartphone app that solicits input from people riential purchases are regarded as more unique than material pur- in your own social network who have had experience with the purchases, and this uniqueness increases self-definition, which in turn chase in question. For each of the purchases, participants were asked leads to higher happiness. These findings connect previously dispa- to choose which of the two sources of information they would like to rate streams of research on happiness and subjective well-being in consult. As expected, participants chose to access input from people psychology and consumer research and provide new insights into the in their social network significantly more frequently when consider- role of uniqueness in consumer behavior. ing the six EP than when considering the six MP. In Study 3, participants read the same scenario from Study 2. Distance Matters: What Information do People Want They were asked to imagine that, before making a decision, they When Making Experiential and Material Purchases? could get advice from someone experienced with the item: “Sup- pose there are 12 people you can consult. Six of these people are EXTENDED ABSTRACT not people who are close to you, but everyone would agree that they When consumers consider making an experiential purchase, do are very similar to you. The other six people are the opposite: They they resort to more or less word of mouth, compared to when they are people who are close to you, but you wouldn’t say that you and consider making a material purchase? There is previous research they are very similar.” To encourage participants to consider each in support for both. On the one hand, the evaluation of experien- purchase and source of advice individually, we limited to six the tial purchases (EP) – such as trips, restaurants, and gym member- number of times they could use either source. In line with our theory, ships – is more uncertain than the evaluation of material purchases participants preferred to receive input from close others (as opposed (MP)– such as computers, clothes, and jewelry – and this uncertainty to similar others) significantly more often when considering the six is frequently solved by turning to other consumers who have already EP than when considering the six MP. undergone that experience (Nelson 1970). Nevertheless, there is also In study 4, participants were reminded that “before making any evidence that people rely less on consumer reviews when making kind of purchase, we frequently access one or several sources of in- EP than when making material purchases (Dai et al. 2020), because formation.” They were asked to rate the extent to which they would they believe that assessments of experiences are based less on the use each of four sources for 12 possible purchases: WOM from close purchase’s objective quality, which makes other consumers’ reviews others, online summary reviews by previous customers, expert re- less helpful. This apparent contradiction is important to resolve given views, and purchase description provided by the seller. As expected, the ever-increasing volume of experiential consumption in consum- participants claimed to use WOM coming from close others more ers’ lives, and the increasing use of customer reviews in consumers’ often for EP than for MP. On the other hand, participants claimed decision-making. the opposite regarding the use of the other three information sources: We distinguish two different types of customer reviews. One they claimed to use online reviews, information from experts, and type, accessible online (e.g., Amazon, Yelp, TripAdvisor), offers product description coming from the seller more often for MP than the assessments of a large number of unknown others. The other for EP. involves WOM assessments from a smaller number of well-known In Study 5, participants were asked to consider the purchase others (friends, family members). The former is likely to be seen as of each of four goods (two material, two experiential). Either a especially informative about what a potential purchase is like “objec- friend or an expert (varying between subjects) randomly endorsed tively” because it is thought to reflect the wisdom of crowds. In con- each product. Consistent with our hypothesis, participants had bet- trast, WOM recommendations from close others are likely to be seen ter attitudes towards experiential goods endorsed by a friend (versus as more informative about how a potential purchase would be evalu- those endorsed by an expert), and towards MP endorsed by an expert ated by us subjectively—what we’re likely to think of it. We predict (versus those endorsed by a friend). Study 6 confirmed this same that when evaluating EP, consumers will rely more on reviews from result at the attribute level. One same purchase (smartphone) was sources that are closer to the self, such as WOM from friends and described through its two most important attributes, one experien- family members. Past research has shown that people consider their tial, one material (established at a pretest): user friendliness, and the Advances in Consumer Research (Volume 48) / 1041 ratio of screen size versus phone weight. We had four conditions, tarian purchase (Mmaterial = 29.08 vs. Mexperience = 59.05; F(1, 220) = depending on whether each attribute was endorsed by an expert or 94.50, p < .001) and used more positive emotional words (Mmaterial a friend. The phone that was rated higher was the one for which = 2.80 vs. Mexperience = 2.98; F(1, 213) = 10.70, p = .001). We found an expert endorsed the material attribute, and a friend endorsed the no significant effect of discounting or interactions on either of the experiential attribute. variables. Study 3 used a 2 (purchase: experiential vs. material) x 2 (think Why Are Reviews of Experiential Purchases Less about the readers: before vs. after) between-subjects design. Partici- Credible? pants were instructed to write the review about their recent dining experience (vs. kitchen appliance purchase) of the same price, ei- EXTENDED ABSTRACT ther before or after answering questions about how other consum- Consumers derive more happiness from spending money on ac- ers would evaluate their reviews. Again, participants wrote about quiring experiences than material purchases (Van Boven and Gilov- their dining experience in a more confident tone (Mmaterial = 29.92 ich 2003). Greater happiness with experiences is often attributed to vs. Mexperience = 54.44; F(1, 299) = 74.97, p < .001) and used more additional utility from memories of the experiences and ability talk positive emotional words (Mexperience = 2.92 vs. Mmaterial = 2.64; F(1, to others about them (Kumar and Gilovich 2015; Zauberman, Rat- 299) = 22.28, p < .001) compared to kitchen appliances. Out of three ner, and Kim 2009). Indeed, experiences have a greater conversa- measures of how others would evaluate their reviews (usefulness, tional value as compared to material purchases (Bastos and Brucks reliance, and objective quality), participants in the experiential con- 2017). We examine how consumers write about their consumption dition expected objective quality to be perceived lower, when par- experiences on online review platforms. ticipants answered questions before as compared to after writing the When consumers write reviews, their goal is to share their con- review(Mbefore = 6.86 vs. Mafter = 7.28; F(1, 299) = 4.44, p = .036). sumption experience online to persuade others by providing helpful Incentives are often employed to encourage consumers to take information (Berger 2014). However, recent research demonstrate a more professional approach to write more helpful online reviews that consumer reviews are often discounted, especially the reviews (Stephen et al. 2012). Thus, studies 4 and 5 tested monetary and that might appear more extreme in tone and emotions because they status incentives, respectively, as a moderator, to see if incentives are unexpected and considered relatively unhelpful due to subjectiv- could influence consumers to change their writing style. Both studies ity (Dai et al. 2020; Rocklage and Fazio 2020). Thus, if the review had a 2(purchase: experiential vs. material) x 2(incentives: yes vs. writers are aware of readers’ discounting, they should be using less no) between-subjects design. Participants in the incentive condition emotional words and extreme tone. Instead, we predict that for expe- were told that they would receive a $50 gift card and an Elite (Vine) riential purchases, review writers do exactly the opposite. This hap- status for monetary and status incentives, respectively. For monetary pens because consumers tend to view their experience more posi- incentives, we replicate the effect of purchase type on confidence in tively (Van Boven and Gilovich 2003), and they tend to use more tone (Mmaterial = 33.05 vs. Mexperience = 51.13; F(1, 274) = 43.15, p < positive emotional words as the valence of the experience increases .001) and positive emotions (Mmaterial = 2.80 vs. Mexperience = 2.97; F(1, (Rocklage and Fazio 2020). Further, when consumers focus on emo- 273) = 10.47, p = .001), but we found no effect of incentives on these tion in justifying their choice, they believe their choice reflects their variables or the interaction (ps > .2). The results were the same for true selves (Carter and Gilovich 2012) and are more confident about status incentive (confidence in tone: Mmaterial = 29.18 vs. Mexperience = their choice (Maglio and Reich 2019). 51.83; F(1, 167) = 46.27, p < .001; positive emotions: Mmaterial = 2.78 Study 1 established initial evidence that consumers write online vs. Mexperience = 2.99; F(1, 160) = 8.14, p = .005). Again, there was no reviews for material and experiential purchases differently. Two hun- effect of incentives on these variables or the interaction ps( > .14). dred most recent online reviews for dining experience from ten most Our findings contribute to the literature in two ways. First, we popular local restaurants for Yelp and ten more popular utilitarian add to the literature on how differences in consumers’ processing of products for Amazon were collected. These reviews served a proxy the experiential and material consumption manifest in consequen- for experiential and material purchases, respectfully. They were ana- tial behavior, such as writing reviews. Second, our findings give in- lyzed with Linguistic Inquiry and Word Count (LIWC, Pennebaker sights to the marketers on developing strategies to guide consumers et al. 2015) and Evaluative Lexicon (EL, Rocklage and Fazio 2015). to write reviews that are less likely to be discounted by the readers. Clout variable (degree of confidence) from LIWC and Extremity variable (extent of positive emotion) from EL were selected to assess REFERENCES differences in writing style. We found that compared to reviews on Bastos, Wilson and Merrie Brucks (2017), “How and Why Amazon, consumers adopt a more confident tone for reviews on Yelp Conversational Value Leads to Happiness for Experiential and

(Mmaterial = 41.15 vs. Mexperience = 60.39; F(1, 199) = 19.89, p < .001), Material Purchases,” Journal of Consumer Research, 44(3), controlling for actual review valence (number of stars) and length of 598–612. review. There was no difference in use of positive emotional words, Belk, Russell W. (1988), “Possessions and the Extended Self,” when controlling for length and valence of review (p = .274). Journal of Consumer Research, 15(2), 139–68. In the next set of studies, we replicate the effect and further at- Berger, Jonah (2014), “Word of Mouth and Interpersonal tempt to change consumers’ writing style by either drawing attention Communication: A Review and Directions for Future to discounting of reviews by readers (studies 2-3) or by incentivizing Research,” Journal of Consumer Psychology, 24(4), 586–607. them to write more convincing reviews (studies 4-5). Study 2 used Carter, Travis J. and Thomas Gilovich (2012), “I Am What I Do, a 2 (purchase: experiential vs. material) x 2 (discounting message: Not What I Have: The Differential Centrality of Experiential present vs. none) between-subjects design. Participants were asked and Material Purchases to the Self,” Journal of Personality to write a short review about their latest dining (material) purchase. and Social Psychology, 102(6), 1304–17. In the discount condition, participants were reminded that reviews Csikszentmihalyi, Mihaly and Reed Larson (1987), “Validity and are often discounted. Replicating Study 1, participants used a more Reliability of the Experience-Sampling Method,” The Journal confident tone when writing about their dining experiences vs. utili- of Nervous and Mental Disease, 175 (9), 526–36. 1042 / Experiences vs. Material Goods: What Drives Consumers to Perceive Experiences and Material Goods Differently?

Dai, Hengchen, Cindy Chan, and Cassie Mogilner (2020), “People Rocklage, Matthew D. and Russell H. Fazio (2020), “The Rely Less on Consumer Reviews for Experiential than Enhancing Versus Backfiring Effects of Positive Emotion Material Purchases,” Journal of Consumer Research, 46(6), in Consumer Reviews,” Journal of Marketing Research, 1052–75. 0022243719892594. Dunn, Elizabeth W., Daniel T. Gilbert, and Timothy D. Wilson Rocklage, Matthew D. and Russell H. Fazio (2015), “The (2011), “If Money Doesn’t Make You Happy, Then You Evaluative Lexicon: Adjective Use as a Means of Assessing Probably Aren’t Spending It Right,” Journal of Consumer and Distinguishing Attitude Valence, Extremity, and Psychology, 21 (2), 115–25. Emotionality,” Journal of Experimental Social Psychology, 56, Fredrickson, Barbara L. (1998), “What Good are Positive 214–27. Emotions?” Review of General Psychology, 2 (3), 300–19. Schwarz, Norbert and Gerald L. Clore (1983), “Mood, Gilovich, Thomas, Amit Kumar, and Lily Jampol (2015), “A Misattribution, and Judgments of Well-Being: Informative Wonderful Life: Experiential Consumption and the Pursuit of and Directive Functions of Affective States,”Journal of Happiness,” Journal of Consumer Psychology, 25 (1), 152–65. Personality and Social Psychology, 45 (5), 513–23. Inesi, M. Ena, Simona Botti, David Dubois, Derek D. Rucker, and Şimşek, Ömer Faruk and Berna Yalınçetin (2010), “I Feel Adam D. Galinsky (2011), “Power and Choice: Their Dynamic Unique, Therefore I Am: The Development and Preliminary Interplay in Quenching the Thirst for Personal Control,” Validation of the Personal Sense of Uniqueness (PSU) Scale,” Psychological Science, 22 (8), 1042–8. Personality and Individual Differences, 49(6), 576–81. Kumar, Amit and Thomas Gilovich (2015), “Some ‘Thing’ to Stephen, Andrew T., Yakov Bart, Christilene du Plessis, and Dilney Talk About? Differential Story Utility From Experiential Gonçalves (2012), “Does Paying for Online Product Reviews and Material Purchases,” Personality and Social Psychology Pay Off? The Effects of Monetary Incentives on Consumers’ Bulletin, 41(10), 1320–31. Product Evaluations,” Vancouver, Canada: University of Maglio, Sam J. and Taly Reich (2019), “Feeling Certain: Gut Pittsburgh, http://d-scholarship-dev.library.pitt.edu/18124/. Choice, the True Self, and Attitude Certainty,” Emotion, 19(5), Vallacher, Robin R. and Daniel M. Wegner (1987), “What Do 876–88. People Think They’re Doing? Action Identification and Maslow, Abraham H. (1954), “Motivation and Personality Harper Human Behavior,” Psychological Review, 94 (1), 3–15. and Row,” New York, NY. Van Boven, Leaf and Thomas Gilovich (2003), “To Do or to Have? Nelson, Phillip (1970), “Information and Consumer Behavior,” That Is the Question,” Journal of Personality and Social Journal of Political Economy, 78(2), 311–29. Psychology, 85(6), 1193–1202. Nicolao, Leonardo, Julie R. Irwin, and Joseph K. Goodman Zauberman, Gal, Rebecca K. Ratner, and B. Kyu Kim (2009), (2009), “Happiness for Sale: Do Experiential Purchases Make “Memories as Assets: Strategic Memory Protection in Choice Consumers Happier than Material Purchases?,” Journal of over Time,” Journal of Consumer Research, 35(5), 715–28. Consumer Research, 36(2), 188–98. Pennebaker, James W., Roger J. Booth, Ryan L. Boyd, and Martha E. Francis (2015), Linguistic Inquiry and Word Count: LIWC2015, Austin, TX: Pennebaker Conglomerates, www. LIWC.net. Small Actions, Big Difference: Consumer Research for A Better World Chairs: Aparna Labroo, Northwestern University, USA Shilpa Madan, Virginia Tech, USA

Paper #1: Subjective Age and the Greater Good perception and over-generalization literatures, they suggest theoreti- Jen H. Park, Stanford University, USA cally rich and practically useful interventions to persuade consumers Szu-chi Huang, Stanford University, USA to buy ugly produce, thereby reducing food waste. Wu, Malter, and Daniella Kupor, Boston University, USA Johar suggest that product anthropomorphism may be a low-cost way to encourage consumers to recycle their possessions. They sug- Paper #2: Finding “I” in Activism: Seeking Agency after gest increased empathy for human like products as the underlying Personal Goal Failures mechanism while identifying important boundary conditions under Sarena Su, Northwestern University, USA which anthropomorphism backfires as an effective way to encourage Aparna Labroo, Northwestern University, USA recycling. Paper #3: Does Appearance Reveal Character? Lay Theories As social, environmental, political, and demographic changes of Physiognomy Influence Consumers’ Willingness to Buy transform and disrupt the way consumers think about and live their Imperfect Produce lives, these insights aspire to promote actions that enhance the long- Shilpa Madan, Virginia Tech, USA term well-being of humankind and the environment. In this session, Krishna Savani, Nanyang Technological University, Singapore we aspire to encourage a lively discussion on the ways in which the Gita V. Johar, Columbia University, USA field of consumer research can come together to provide theory driv- Paper #4: When Products are People: The Impact of en, practically sound solutions for the most pressing concerns of the Anthropomorphism on Recycling twenty first century. Alisa Wu, Columbia University, USA Maayan Malter, Columbia University, USA Subjective Age and the Greater Good Gita V. Johar, Columbia University, USA EXTENDED ABSTRACT SESSION OVERVIEW Charitable organizations often struggle to solicit sufficient con- Marketing, with its deep understanding of consumers, has the tributions to aid people in need. Substantial research reveals that in- unique ability to improve their lives and make this world a better creasing people’s felt responsibility towards others can boost their place. How can we, as consumer researchers, encourage people to prosocial contributions (e.g., Michel 2007; Wilson and Musick, make a difference? Tied together by this unifying theme, the four 1997). We uncover a new driver that can increase consumers’ felt papers in this session showcase novel theoretical insights to pro- responsibility towards others in need and increase their prosocial mote an array of pro-social and pro-environmental behaviors such contributions as a result. as conserving energy and recycling, reducing food waste, advocating The perception that older (vs. younger) adults have greater re- for social and environmental causes, and helping strangers in need. sponsibility for aiding society at large originates from a rich array Consistent with the conference theme “Rendezvous,” this session is of sources, including cultural narratives, public communications a melting pot of diverse theories, approaches, and methods (includ- (Gheaus 2016), and normative pressures (e.g., Browning 1975; Mc- ing field and/or behavioral data across all papers) to illuminate how Adams and Aubin, 1992). Drawing from this perception, we hypoth- consumer research can contribute to the greater good of the society. esize that when consumers encounter cues that make them feel sub- The session features insights that are of strong interest to a broad set jectively older, their felt responsibility for contributing to the greater of stakeholders — including policy makers, businesses, not for profit good of society increases, which in turn heightens their contributions. organizations, communities, and consumers. Studies 1A-1C found that objectively younger reference points The first two papers present novel theoretical insights toin- increased consumers’ subjective age. Participants in Study 1A were crease pro-social/environmental behaviors and advocacy. Park, randomly assigned to a Younger Cue condition (imagined interact- Huang, and Kupor set the stage by leveraging consumers’ subjec- ing with a younger individual), an Older Cue condition (imagined tive age to encourage donations of time and money to charitable and interacting with an older individual), or a Baseline condition (did not environmental causes, across lab and field studies. They find that complete this interaction task). Participants then indicated how old increasing people’s subjective age, i.e., feeling older, enhances vol- they felt on a 100-point scale (1: extremely young; 100: extremely untary behavior by heightening felt responsibility for others. Su and old). Analysis revealed that participants in the Younger Cue con- Labroo uncover a novel antecedent for consumer activism, i.e., advo- dition felt older (M = 49.74, SD = 22.98) than participants in both cacy for critical social issues such as discriminatory gender, race, or the Older Cue condition (M = 38.90, SD = 24.44; p = .001) and the environmental policies. Drawing from goal theory, they suggest that Baseline condition (M = 39.19, SD = 18.94; p = .001), the latter of personal goal failures undermine consumers’ agency, leading them which did not differ (p = .93). Studies 1B-1C replicated this phenom- to engage in activist behaviors by affiliating with likeminded others enon across two additional interventions of subjective age: consider- thereby restoring personal agency. ing one’s younger or older self (Study 1B) and viewing the ages of While the first two papers tackle broad social and environmen- younger or older others (Study 1C). tal issues from a consumer perspective, the next two papers focus on Study 2 found that increasing consumers’ subjective age in- specific pro-environmental actions of “reduce, [reuse] and recycle” creased their aid to strangers in need. Participants were randomly by focusing on product characteristics. Madan, Savani, and Johar assigned to a Younger Cue (i.e., Feel Older) condition, an Older Cue take a novel approach to understand consumers’ aversion to ugly condition, or a Baseline condition; these conditions were manipu- produce —their lay theories about physiognomy, i.e., the idea that lated as in Study 1A. Then, all participants viewed an advertisement people’s appearance can reveal their character. Building on person- soliciting donations to aid the blind and indicated their willingness to

Advances in Consumer Research 1043 Volume 48, ©2020 1044 / Small Actions, Big Difference: Consumer Research for A Better World contribute to this cause on a 1-7 scale. Analysis revealed that condi- Finding “I” in Activism: Seeking Agency after Personal tion impacted prosocial intentions (F(2, 297) = 7.22, p = .001): Par- Goal Failures ticipants in the Younger Cue (i.e., Feel Older) condition were more willing to support the cause (M = 6.04, SD = 2.04) than were partici- EXTENDED ABSTRACT pants in the Older Cue condition (M = 4.98, SD = 2.31; p = .001) and Activism is on the rise—and firms increasingly are not immune the Baseline condition (M = 5.05, SD = 2.24; p = .002); the latter two to activist action, as evidenced by a multitude of protests against conditions did not differ (p = .83). This effect persisted when control- brands including Chick Fil A, Uber, McDonalds, Starbucks, etc. for ling for participants’ chronological age, and participants’ chronologi- discriminatory gender, wage, or religious/racial policies. Activism cal age did not predict their prosocial behavior. typically arises in response to societal failures and injustice, and as a Studies 3-4 tested our prediction that feeling older increases way for people to join likeminded others to bring about changes for contributions by heightening felt responsibility for others and ruled the communal good. Activists are often seen as not directly benefit- out an alternative account of perception of time left. Participants ting from the change they bring about, and when they do benefit from were randomly assigned to one of the three conditions described in the change, the change is assumed to benefit the greater good and Study 2 and indicated their willingness to volunteer their time for community at large more. Indeed, engaging in activism is consid- a pro-environmental cause on a 1-7 scale. Then, they completed a ered risky. To overcome these personal costs, activists are assumed six-item scale assessing their felt responsibility for others’ welfare to have a high degree of communal orientation and a concern for (e.g., “I feel responsible for future generations”; Morselli and Pas- social justice (Hoffman 1989; Omoto, Snyder, and Hackett 2010). sini, 2015; α = .90) and a ten-item scale capturing their perception of Research shows consumers are driven by their prosocial motivations the amount of time remaining (e.g., “I have the sense time is running to engage in activism, address social-political failures, and advance out”; Carstensen and Lang, 1997; α = .71). Participants in the Young- the welfare of disadvantaged groups in society (Horberg, Oveis, and er Cue (i.e., Feel Older) condition were more willing to contribute Keltner 2011; Johnson 2009). While there is no doubt that prosocial to the prosocial cause (M = 4.56, SD = 1.77) than those in the Older orientations influence people’s willingness to engage in activism, we Cue condition (M = 3.76, SD = 1.73; p = .001) and the Baseline shed light on a novel antecedent that is not prosocial, but instead condition (M = 3.83, SD = 1.72; p = .003); the latter two conditions potentially selfish, in nature: we demonstrate that perceived personal did not differ (p = .78). This result persisted when controlling for goal failures can also increase activism. participants’ chronological age (b = .80, SE = .25, t(294) = 3.25, p Consumers who face personal goal failures typically look to = .001), which did not predict prosocial behavior (b = .03, SE = .03, understand the causes of their failures. The causes could pertain to t(294) = 1.24, p = .22). Importantly, a mediation analysis revealed internal factors, things the consumer did or did not do, or to external that felt responsibility mediated the effect of subjective age on proso- factors that contributed to the failure (Heider 1958). Consumers are cial behavior (95% CI: .1022 to .4149); a separate mediation analysis known to spontaneously make external attributions for their failures showed that time perception mediated the effect of chronological age to protect their sense of self (Jellison and Green 1981; Jones and Har- on prosocial behavior (95% CI: -.0105 to -.0002). In other words, ris 1967). For instance, drivers causing an accident are more likely feeling subjectively older increased intentions to act for the greater to report a light suddenly turned than to admit they were driving too good by heightening felt responsibility for others; in contrast, being fast; students performing poorly are more likely to say the professor chronologically older was associated with reduced perceived time was a tough grader or the course was too difficult than to admit they remaining, which in turn was associated with reduced prosocial be- did not try enough or seek assistance in time. We posit that making havior. Using a similar design, Study 4 found that this phenomenon external attributions could make consumers infer a lack of external persisted regardless of whether subjective age was shifted through an support, and they therefore have less personal agency—the ability to interpersonal or intrapersonal age cue. take action, be effective, influence their own life—because of others. Studies 5-6 examined that—because older subjective age in- As a result, they may seek to affiliate with like-minded others, as a creases prosocial behavior by increasing felt responsibility for dis- way to reestablish their personal agency. Activism, by offering the tant others—older subjective age increases giving to distant others opportunity to affiliate with like-minded others, could be seen by but not to close others. Specifically, Study 6 tested this phenomenon consumers as a way to garner personal agency. As a result, consum- in the field. We organized a thank-you card event on campus and ers who consider their personal failures may experience heightened invited campus visitors to give thanks to their family members (i.e., need to affiliate with likeminded others to reestablish personal agen- close others) and/or university staff (i.e., distant others) by writing cy, which will increase engagement in activism, especially activism thank-you cards. Participants’ subjective age was manipulated by that could offer a higher boost to personal agency. varying the average age information about prior participants. As ex- Across six studies, we show support for our hypotheses that pected, and consistent with Study 5, subjective age condition did failure to progress on one’s goal leads to an increase in activism. not impact the number of cards written to close others (MYounger Cue = In study 1, we examined the relationship between perceived failure

.73, SDYounger Cue = .97; MOlder Cue = .82, SDOlder Cue = .96; p = .58), but on career goals and political activism using real-world data from a participants wrote more cards to distant others when they felt older large cross-national survey (23 countries; N = 44,387). We show that

(MYounger Cue = 1.08, SDYounger Cue = 1.30; MOlder Cue = .62, SDOlder Cue = .87; expectations of unemployment in the next 12 months correlate posi- p = .01). These results persisted when controlling for participants’ tively with participation in public demonstrations (b = .05, SE = .02, chronological age. z = 2.00, p = .046) and online political activism (b = .08, SE = .02, z In sum, we find that increasing people’s subjective age height- = 5.11, p < .001), controlling for income, and for communal reasons ens their prosocial contributions through increased felt responsibility for engaging in activism, including mistrust in the system, political for others. Our findings provide a novel insight into how a classic orientation, etc. Study 2 (N = 182) established the causal link be- demographic variable (i.e., chronological age) can become a mallea- tween personal goal failure and political activism using experimental ble construct (i.e., subjective age), which can be leveraged to nudge methods, providing evidence that regardless of goal type, consumers people toward actions that aid others in need. who considered chronic or situational goal failures (M = 2.31, SD = .76) showed higher intentions of engaging in political activism (M = Advances in Consumer Research (Volume 48) / 1045

2.03, SD = .87; p = .030) . Directing their thoughts to goal success argue that consumers tend to overgeneralize and apply this lay belief reduced this tendency. In study 3, we replicated these findings in regarding people to judging produce with atypical appearance. a field study run at an Auto Expo and at an upmarket cafeteria (N Six studies provide evidence for this idea, rule out several alter- = 205). Consumers who considered failures to own a preferred car native explanations, test boundary conditions, and propose multiple (M = .060, SD = .072) were more willing to engage in animal rights interventions for retailers to persuade consumers to buy imperfect activism (M = - .15, SD = .073; p = .047). Notably, animal rights produce. activism cannot address their lack of a preferred vehicle but engag- Study 1 manipulated the appearance reveals character lay theo- ing in such activism can afford affiliation with likeminded others and ry and found that reading an article explaining that people’s appear- therefore a perception of reestablishing personal agency. Study 4 ance can reveal their character lowered consumers’ willingness to then demonstrated, as we predicted, that this effect only arises when pay for imperfect produce (WTPAppearanceRevealsCharacter =3.51, SD = 1.43, people make external attributions for their goal failure, thus estab- WTPAppearanceDoesNOTRevealCharacter =3.96, SD=1.66, t(211)=2.08, p=.038). lishing external (vs. internal) attribution as an important boundary The inference of lower quality (poor taste, texture, and nutrition) condition for our effect (N = 302,p = .017). In study 5, we examined mediated the effect (B=-.19, SE=.091, 95%CI [-.38, -.026]. This in- our full process model: the mediating role of need for affiliation with ference of poor quality was the only significant mediator even after like-minded others and a resulting need to reestablish one’s agency including several other explanations such as poor self-perception, through activism after the experience of personal goal failures, but self-esteem, disgust, and difficulty to use as competing mediators in only when activism can reestablish their agency (N = 299, p < .001). the mediation model. Lastly, study 6 explored the consequences of engaging in activism Study 2 replicated the above effect by measuring consumers’ (N = 400). appearance reveals character lay theory and adding a typical pro- Using multiple methods—observational data, experimental duce control condition. We found a significant indirect effect of the study, field settings—different personal goals (career, any listed lay theory on willingness to buy imperfect produce through the per- spontaneously, affording luxury cars)—different causes (political ac- ception that imperfect produce might be poor in taste, texture, and tivism, environmental activism, animal rights, free college for all)— nutrition (B=-.18, SE=.09, 95% CI [-.37, -.02]) while ruling out ad- across different populations (low-income MTurkers, middle/high- ditional concerns that imperfect produce is not natural or is geneti- income executives, European respondents from 23 countries)—we cally modified. showed a robust link between perceived failures to progress on Given our hypothesis that over-generalization from people to personal goals and activism. Importantly, our hypotheses held even produce drives this effect, Study 3 tested the idea that if consumers when controlling for a variety of factors that prior literatures suggest were made aware that over-generalizing from people to produce is having an influence on activist intent—communal motives, political illogical and irrational behavior, they would be less likely to use this ideology, objective income, highest level of education, subjective lay theory to judge imperfect produce. We measured the lay theory as experience of frustration, and general trust in the political system. in Study 2 and randomly assigned participants either to an interven- Thus, while prosocial motivations and concerns for a just society tion condition (where they read a short paragraph about how illogi- certainly matter, as previously shown, we show that the desire to cal it is to perceive human like attributes in objects) or the control advance one’s personal goals is also an important pathway to activ- condition (where they read a short paragraph about memory). The ism independent of prosocial influences. We therefore contribute to interaction between lay theory and condition was significant (B=- the current literature on activism by providing new insights into the 1.084, SE=.51, t(143)=2.13,p=.035). In the control condition, the importance of perceived personal goal failures in prompting people more participants believed that individuals’ appearance reveals their to engage in activism as a way to enhance their personal agency. character, the less willing they were to buy imperfect produce (B=- (1000 words) .95, SE = .42, t(71)=2.28, p=.026). More importantly, in the intervention condition, their lay theory was not related to their willingness to Reducing Food Waste: Lay Theories of Physiognomy buy imperfect produce (B=.13, SE=.301, t(72)=.2, p=.66). Shape Willingness to Buy Imperfect Produce Studies 4 and 5 tested if explicit anthropomorphization (a communication strategy used by ugly produce startups), improves EXTENDED ABSTRACT or reduces consumers’ willingness to purchase ugly produce. We The global population is projected to touch 9 billion by 2050 found that in this context, anthropomorphization backfires. Spe- making minimizing food waste critical to preventing large scale food cifically, consumers who believed that appearance reveals- charac insecurity. Food that gets rejected due to atypical or imperfect ap- ter (+1 SD) were not swayed by anthropomorphized depictions of pearance contributes up to 40% of total food waste. In this research, imperfect fruits and vegetables. However, consumers who believe we identify a novel lay theory that influences consumers’ willingness that appearance does not reveal character became less likely to buy to buy imperfect produce. We find that the more consumers believe imperfect produce when it was anthropomorphized indicating that in the idea of physiognomy, or that people’s outer appearance re- although these consumers did not infer poor quality from atypical veals their inner character, the less likely they are to buy imperfect looks, anthropomorphization shifted these participants’ attention to produce. the produce’s weird looks and reduced willingness to buy (Wen Wan Across cultures, people have believed that a person’s face is the et al. 2017). window to their soul, a reflection of what lies beneath the surface. Finally, two field studies on Facebook™ tested another action- People unfailingly infer others’ personality, intelligence, social skills, able intervention to persuade consumers to consider imperfect pro- and mental health from their appearances in split seconds (Berry and duce—by decoupling appearance from taste, texture, and nutrition Brownlow 1989; Willis and Todorov 2006). These judgments, such by creating an appealing and healthy-looking soup from imperfect as inferring personality traits from appearance, are surprisingly ro- produce. In discussion with an ugly produce start-up, we sought to bust and reliable (Berry and Wero 1993). The lay belief that people’s test if the intervention was effective at both increasing engagement outer appearance reveals their inner character suggests that what is (likes, comments, and shares) and link clicks (to go to an external on the outside is an accurate indicator of what lies on the inside. We website to learn more about imperfect produce). Studies 7 and 8 1046 / Small Actions, Big Difference: Consumer Research for A Better World

(N=34,602) found that the intervention ad was significantly more ef- in the lab, participants were randomly assigned to two conditions fective in generating active engagement among the target audience – anthropomorphism condition or the control condition. In both (152.9% more engagement) and also more persuasive in leading conditions, participants were given the same cookie to taste. An- them to go to the website to learn more (38.2% more clicks) at a thropomorphism was manipulated through an instruction sheet. In significantly lower cost. the anthropomorphism (vs. control) condition, participants read the The findings identify a novel lay belief that influences consum- instruction using first-person (vs. third-person) voice. After reading ers’ willingness to buy imperfect produce and multiple strategies the instructions and tasting the cookie, participants answered several for mitigating this bias. This research leverages a unique theoretical questions including the aesthetic feature, the sweetness, and their lik- perspective, at the intersection of lay theories, over-generalization, ing of the cookie. Before they left the lab, participants were told that and anthropomorphism, to create interventions that contribute to the since the bakery is new, everything should be kept confidential, so emerging, and critically important stream of work on food waste. they needed to discard the experimental materials before leaving the lab. There were two bins by the door, trash and recycling. Partici- When Products are People: The Impact of pants disposed of the instruction sheet, any remaining cookie into the Anthropomorphism on Recycling bins before leaving the lab. We found that in the anthropomorphism condition, 96.8% of the participants recycled the information sheet EXTENDED ABSTRACT compared with 89.40% did in the control condition (). Moreover, US consumers discard most materials (76%) after using them we found that six participants tore the instruction sheet into pieces only once (White, MacDonnell and Dahl, 2011); thus, recycling be- in the control condition, whereas no participant did the same thing havior is a critical step to creating a sustainable environment. This in the anthropomorphism condition, indicating that people wanted research investigates a specific type of product characteristic, an- to minimize potentially hurting the humanlike object when they dis- thropomorphism, and its effect on recycling. Anthropomorphism is posed of it. guided by the same social cognitive mechanisms that enable people Experiment 2 aimed to a) replicate the findings in experiment to think about the minds of other people (Epley, 2018). We argue that 1 and b) examine when anthropomorphism backfires. In this experi- when it comes to disposal decisions (trashing vs. recycling), people ment, we used a smoothie-maker as our stimulus and manipulated feel like they are discarding a person when they have to dispose a anthropomorphism by describing the smoothie-maker from a first- humanlike product. Because they feel empathy for this human-like person perspective. We manipulated disassembling instruction by product, they will be more likely to recycle it in order to avoid hurt- telling half of the participants that “Since the smoothie maker consists ing it. However, if recycling requires consumers to disassemble the of several parts, you must carefully disassemble it into different parts product into different parts (e.g., plastic, paper, metal), it feels like before disposing these parts into different recycling bins.” They then dismembering “a person” and thus consumers will be more likely to answered our dependent variable measures – “How likely are you trash the product as a whole to avoid hurting it. to trash the smoothie maker?” and “How likely are you to recycle Anthropomorphism can be operationalized by making a prod- the smoothie maker?” on a seven-point scale. The other half of the uct look humanlike (perceptual) or by giving the product a first-per- participants received no disassembly instructions and just answered son humanlike voice (conceptual). We believe that consumers will the dependent measures. We used the difference between the two only perceive the product to feel humanlike (not just look humanlike) measures as our dependent variable. Results showed a significant for conceptual anthropomorphism thus inducing empathy toward the interaction effect between anthropomorphism and disassembling (). product. Therefore, only conceptual anthropomorphism should in- When there was no disassembling instruction, participants recycled crease recycling. more when the product was humanlike (); whereas when the disas- To test our hypotheses, we conducted four studies. A pretest sembling instruction was present, participants recycled less when the examined various types of manipulations of anthropomorphism to product was humanlike (). The effect could not be explained by ef- identify which manipulations consistently increase recycling. Ex- fort because in the control condition, disassembling instruction made periment 1 tested our proposed main effect in a lab study with a real participants recycle more (). behavioral dependent measure of recycling. Experiments 2 and 3 ex- Experiment 3 replicated the effect that when people had to dis- plored our proposed interaction effect between anthropomorphism assemble a product, anthropomorphism backfires. We used the same and disassembling on recycling. stimuli as Experiment 2 and all participants read the instruction that Pretest. We designed 22 pairs of product pictures, with one if they wanted to recycle the smoothie maker, they must disassemble option anthropomorphized and the other non-anthropomorphized. it into different parts. Results showed a significant decreasing ef- We manipulated perceptual and conceptual anthropomorphism - fect of anthropomorphism on recycling (). When the disassembling the perceptual-anthropomorphism condition displayed products instruction was present, participants recycled less when the product with visual humanlike features (face or body shape), whereas the was humanlike () than when the product was not humanlike (). conceptual-anthropomorphism condition displayed products with In sum, we find that people prefer to minimize hurting the an- the same humanlike physical features accompanied by a first-person thropomorphized product – either by recycling it (and possibly giv- perspective description. Participants were randomly provided with ing it the potential to be resurrected) or by trashing it as a whole one of the pictures in each pair. Then participants were asked to drag rather than dismembering it. Our findings suggest that this effect oc- and drop each object into one of the two bins – trash or recycling. curs because consumers have empathy for the humanlike product, Results showed that conceptual anthropomorphism consistently in- but further research is required to test this theory. creased recycling behavior more than perceptual anthropomorphism (). Since conceptual anthropomorphism worked better in the pretest, REFERENCES we used it in the following experiments. Berry, Diane S. and Sheila Brownlow (1989), “Were the Experiment 1. We tested the hypothesis that anthropomorphism Physiognomists Right? Personality Correlates of Facial increases recycling. Participants were told that a newly opened local Babyishness,” Personality and Social Psychology Bulletin, bakery wanted customers to evaluate their cookies. Upon arriving 15(2), 266-79. Advances in Consumer Research (Volume 48) / 1047

Berry, Diane S. and Julia L. F. Wero (1993), “Accuracy in Face Jones, Edward E., and Victor A. Harris (1967), “The attribution of Perception: A View from Ecological Psychology,” Journal of attitudes,” Journal of Experimental Social Psychology, 3(1), Personality, 61(4), 497-520. 1-24. Browning, D. S. (1975). Generative man: Psychoanalytic McAdams, Dan P., and E. D. de St Aubin (1992), “A theory of perspectives. New York: Dell. generativity and its assessment through self-report, behavioral Carstensen, Laura L. and Frieder Lang (1997), “Measurement of acts, and narrative themes in autobiography,” Journal of time orientation in diverse population,” Unpublished data, Personality and Social Psychology, 62(6), 1003. Stanford University, Stanford, CA. Michel, Lacie M. (2007), “Personal responsibility and volunteering Epley, Nicholas (2018), “A Mind like Mine: the Exceptionally after a natural disaster: The case of Hurricane Katrina,” Ordinary Underpinnings of Anthropomorphism,” Journal of Sociological Spectrum, 27 (6), 633-652. the Association for Consumer Research, 3(4), 591-598. Morselli, D. and Passini, S. (2015), “Measuring prosocial attitudes Gheaus, Anca (2016), “The right to parent and duties concerning for future generations: The Social Generativity Scale,” Journal future generations. Journal of Political Philosophy,” 24(4), of Adult Development, 22 (3), 173-182. 487-508. Omoto, Allen M., Mark Snyder, and Justin D. Hackett (2010), Heider, Fritz (1958), The psychology of interpersonal relations. “Personality and motivational antecedents of activism and New York: Wiley civic engagement,” Journal of Personality, 78(6), 1703-1734. Hoffman, Martin L. (1989), “Empathy and prosocial activism,” Wen Wan, Echo, Rocky Peng Chen, and Liyin Jin (2017), “Judging Social and moral values: Individual and societal perspectives, a Book by its Cover? The Effect of Anthropomorphism on 65-85. Product Attribute Processing and Consumer Preference,” Horberg, Elizabeth J., Christopher Oveis, and Dacher Keltner Journal of Consumer Research, 43(6), 1008-30. (2011), “Emotions as moral amplifiers: An appraisal tendency White, Katherine, Rhiannon MacDonnell, and Darren W. Dahl approach to the influences of distinct emotions upon moral (2011), “It’s the mind-set that matters: The role of construal judgment,” Emotion Review, 3(3), 237-244. level and message framing in influencing consumer efficacy Jellison, Jerald M., and Jane Green (1981), “A self-presentation and conservation behaviors,” Journal of Marketing Research, approach to the fundamental attribution error: The norm of 48 (3), 472-485. internality,” Journal of Personality and Social Psychology, Willis, Janine and Alexander Todorov (2006), “First impressions: 40(4), 643. Making up Your Mind After a 100-ms Exposure to a Face,” Johnson, Devon (2009), “Anger about crime and support for Psychological Science, 17 (7), 592-98. punitive criminal justice policies,” Punishment & Society, Wilson, J. and Musick, M. (1997), “Who cares? Toward an 11(1), 51-66. integrated theory of volunteer work,” American Sociological Review, 62, 694-713. Coolness: Identifying What Makes People, Products, and Brands Cool Chairs: Caleb Warren, University of Arizona, USA Martin Reimann, University of Arizona, USA

Paper #1: Consumer Interpretations of Product Coolness Across Consumer Interpretations of Product Coolness Across Three Cultures Three Cultures Gratiana Pol, Hyperthesis, LLC, USA Eden Yin, University of Cambridge, UK EXTENDED ABSTRACT Gerard Tellis, University of Southern California, USA Deeply rooted in U.S. history and culture, the term and notion Paper #2: Why Are Autonomous Brands Cool? The Role of of ‘coolness’ have gained popularity in many other cultures (Belk Value Authenticity and Brand Biography et al. 2010; Van den Berg and Behrer 2011), with marketers around Alessandro Biraglia, University of Leeds, UK the world often willing to go to great lengths to figure out what ap- J. Joṣko Brakus, University of Leeds, UK peals to consumers, and how to design and sell “the next cool thing” (Belk et al. 2010; Kerner and Pressman 2007). Yet despite the global Paper #3: What Makes People Cool? importance of the coolness construct, prior research on coolness has Todd Pezzuti, University of Adolfo Ibanez, Chile been surprisingly scant, particularly in a cross-cultural context. This Caleb Warren, University of Arizona, USA lack of research can be partially attributed to coolness being a socially constructed characteristic, which is somewhat elusive (Nan- SESSION OVERVIEW carrow et al, 2002; Tapp and Bird 2008) and highly subjective (Mac- Consumers want to be cool because coolness brings them admi- Adams 2001; O’Donnell and Wardlow 2000). In the present paper, ration and esteem (Belk et al. 2010; Heath and Potter 2004; Warren we set out to understand the different ways in which U.S.-based—or, 2010). Merely being in the presence of a cool product is enough to more broadly, English-speaking—consumers intuitively interpret the make consumers feel better about themselves (Quartz and Asp 2013). meaning of product coolness, and to what extent these interpretations Recognizing this, marketing professionals spend a lot of time and replicate to other cultures. We collect data from a survey in three ma- money trying to figure out how to create cool products and brands jor cultures that employ the term ‘cool’ in everyday language—the (Gladwell 1997; Kerner and Pressman 2007; Southgate 2003) and U.S. (along with a pilot study in the U.K.), Germany and China. Our marketing researchers attempt to understand what may distinguish research focuses on coolness in the context of a product (as opposed cool product designs from designs that seem either funny or normal to a brand or person). (Warren and Reimann 2019). And for good reason. Just as becom- Study 1 is a pilot examination of how consumers interpret the ing a cool person can help consumers feel better about themselves meaning of product coolness. We use student samples from two (Quartz and Asp 2013), offering a cool product or brand can help English-speaking cultures: 207 respondents from the U.S. and 177 firms feel better about their bottom line (Im, Bhat, and Lee 2015; respondents from the U.K. Using open-ended questions—which Kim, Shin, and Park 2015; Shin and Biocca 2018). we then code into categories—participants indicated how they de- Thus, upon closer inspection, it seems wise for consumers, fine product coolness, and what they believe determines whether a firms, and especially consumer researchers to better understand how product is cool. The results reveal a remarkably robust pattern across to become cool. All of the papers in this special session attempt to the two samples. Consistent with previous findings from coolness answer this question, each using a different approach to address a research (e.g., Warren 2010), respondents largely fail to articulate a different aspect of how to become cool. Pol, Yin, and Tellis inves- clear definition of coolness, but consistently resort to using specific tigate how coolness varies across cultures by gathering data in the product attributes (such as usefulness, aesthetic design, uniqueness, US, UK, Germany, and China, but shift their focus to how products, etc.) as a basis for defining coolness. When it comes to the more gen- rather than people, become cool. Pol et al. also identify two inter- eral determinants of coolness, consumers interpret a product’s cool- pretations of coolness, one relatively personal and the other more ness as emerging when the product provides some form of intrinsic, social, develop a set of measures to assess these two interpretations, social, subjective, and/or marketing-driven value. Because our in- and identify cross-cultural commonalities and variations in these investigation appears to have good ecological validity, we decide to terpretations. Biraglia and Brakus shift the focus from products to continue examining both the coolness-driving product attributes and brands, by investigating why and when it is cool to for a brand to the more general interpretations of coolness in tandem. be autonomous. They find that autonomy is only cool under certain Using the insights generated in Study 1, in Study 2 we develop conditions, including when a brand comes from a disadvantaged and validate a series of scale measures meant to capture the inter- background, the audience is relatively independent, and the product pretations of product coolness, and the coolness-driving product at- is self-expressive. Pezzuti and Warren investigate the characteristics tributes. Given the very strong similarity between the U.S. and U.K. that make people cool using a new methodology that distinguishes samples, this study focuses solely on the U.S. as a representative of cool traits from traits that seem socially desirable or good. They addi- English-speaking cultures. We include 136 undergraduate students tionally explore how the characteristics associated with cool people from a large US university, and a more demographically represen- vary across a variety of cultural variables. tative sample of 248 participants recruited via Amazon Mturk. For By taking different approaches to a similar question, the three both the coolness-driving product attributes and the interpretations papers advance our wisdom and understanding of how people, prod- of coolness, we generate a pool of scale items derived from the exact ucts, brands, and designs become cool. Collectively, they dispel some words provided by participants in Study 1. Respondents are asked to of the ineffability that has long characterized the concept of coolness, indicate, on a scale from 1 to 7, the extent to which they believe each and help researchers, marketers, and designers alike become wiser item describes cool products. about how to capture and create coolness. For the coolness-driving product attributes, we use EFA to arrive at a three-factor solution that is stable across the two samples,

Advances in Consumer Research 1048 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1049 and which includes functional attributes (i.e., usefulness and ease- are often offered in the companies’ narratives about the background of-use), symbolic attributes (i.e., exclusivity and aesthetics), and of their founders (e.g., Steve Jobs “home brewing” computers in a creativity. For the interpretations of product coolness, a two-factor garage). These narratives are persuasive because of the underdog ef- solution emerges. Specifically, respondents view a product as cool fect -- individuals root for and prefer a disadvantaged entity over (a) if it closely meets their needs and preferences (i.e., a personal an advantaged one (Paharia et al., 2011). We claim that consumers interpretation that represents a combination of the subjective and perceive a highly autonomous brand as cooler when the company intrinsic value identified in Study 1, and contains items such as “it highlights a disadvantaged rather than an advantaged biography. possesses outstanding attributes” and “it makes me happy when I use We argue that value authenticity (i.e. the capacity of sticking with it.”), or (b) if it receives the social approval of others (i.e., a social and remaining true to inner values and beliefs; Newman and Smith, interpretation, which represents a broad combination of the social 2016) mediates this relationship. In other words, if a brand acts in and the marketing-driven value identified in Study 1, and contains congruence with its proclaimed values at all stages—production, items such as “it starts a popular trend,” “it is perceived as cool by promotion, or customer care—, then consumers perceive the brand most people,” and “it is associated with a cool brand”). as authentic and consequently as cool. We believe that this is the Study 3 examines to what extent U.S. consumers’ interpreta- case because acting autonomously to succeed is more aligned with tions of product coolness replicates to Germany and China. We con- an underdog rather than with a topdog biography. struct an integrative predictive framework that links the coolness- We further demonstrate how this mediating effect is stronger for driving product attributes to the two interpretations of coolness, and individuals with an independent rather than with an interdependent to consumers’ desirability perceptions vis-à-vis cool products. Using self-construal (Markus and Kitayama, 1989). Moreover, we show Structural Equation Modeling, we find that the framework provides a how the link between high autonomy and being an underdog matters satisfactory fit to the data across all cultures (χ2/df= 2.33, CFI = .89, especially for the coolness in self-expressive than in functional prod- GFI = .82, RMSEA = .05). The two coolness interpretations replicate uct categories. Finally, we extend our findings to political marketing cross-culturally in terms of both factorial structure and importance, by demonstrating how the effect holds also when a politician is a with consumers across cultures agreeing more strongly with the highly autonomous underdog. personal interpretation of coolness–and hence rejecting the notion We conducted four experiments to test the hypotheses. In that they see coolness purely as a socially-determined characteris- Experiment 1 (n = 215) we presented respondents in a laboratory tic. Nevertheless, while U.S. consumers view the two interpretations setting with a scenario about a fictitious coffee company. We maas fully independent facets of coolness (r = -.04, p > .05), Chinese nipulated the background (disadvantaged vs. advantaged) and the consumers perceive them as strongly inter-related (r = .65, p < .001). positioning of the company (high autonomy vs. low autonomy). We The coolness-driving product attributes also vary cross-culturally, found that the interaction between a disadvantaged background and with symbolic attributes being the “cooler” ones in China (Msymbolic a highly autonomous positioning significantly affects the perception

= 5.56, Mfunctional = 5.25, F(1, 140) = 9.33, p < .05), functional ones of coolness (MDisadvantaged_high_autonomy = 5.38, MAdvantaged_ high_autonomy = 3.81; being the “cooler” ones in Germany (Mfunctional = 5.22, Msymbolic = 4.85, MDisadvantaged_low_autonomy = 4.41, MAdvantaged_ low_autonomy = 4.05; F(1, 214) = F(1, 135) = 7.65, p < .05), and neither of them dominating in the 12.01, p < .001). We also found that the interaction was fully medi- U.S. These findings suggest that Chinese consumers view coolness ated by value authenticity (B =.0631, SE = .0862, t = .73, NS, LLCI more strongly related to social and symbolic factors than do U.S. and = –.1069, ULCI = .2330; bias-corrected confidence interval = 95%; German consumers, which may also explain why Chinese consum- number of bootstrapping resamples = 10,000). ers perceive cool products as particularly high in desirability (MChina In Experiment 2 (N= 317 on MTurk) participants rated the cool-

= 5.59, MGermany = 4.94, MUS = 4.75, F(1, 522) = 17.04, p < .001). ness of a fictitious apparel brand in a 2 (autonomy: high vs. low) x Together, these findings support the notion that, despite gravi- 3 (background: disadvantaged vs. advantaged vs. control) between- tating towards a rather personal interpretation of coolness, different subjects design. The results confirmed a significant interaction effect cultures perceive this interpretation, as well as the product attributes of the high autonomy with the disadvantaged background F(2, 316) that ultimately factor into it, in a manner that is subject to cultural = 8.09, p < .001. Next, we tested a moderated mediation (Model 8; influences. We use several perspectives on cross-cultural differences Hayes, 2013) using the interaction term as the IV, value authenticity to shed light on how the general replicability of our model across as the mediator, coolness as the DV, and independent self- construal cultures can be reconciled with cultural variations. as the moderating factor. Results showed that value authenticity fully mediated the direct effect of the interaction effect on coolness and Why Are Autonomous Brands Cool? The Role of Value that consumers’ level of independence moderated the relationship Authenticity and Brand Biography between high autonomy, disadvantaged biography, and coolness. In Experiment 3 (N = 442) we examined the relationship between the EXTENDED ABSTRACT disadvantaged company background and the perception of coolness Coolness has attracted the interest of scholars from different in different product categories (i.e. self-expressive vs. functional). disciplines: semiotics (Danesi, 1994), psychology (Dar Nimrod et We designed a 2 (autonomy: high vs. low) x 2 (background disad- al., 2012), and cultural studies (Hebdige and Potter, 2008). Con- vantaged vs. advantaged) x 2 (category: functional vs. self-expres- sumer researchers have linked coolness to autonomous brands (i.e. sive) between subject design. As predicted, the a high autonomous brands that diverge from the norm; Warren and Campbell, 2014) or underdog is perceived as cool especially when the company operates to the extent by which an endorser expresses or conceals emotions in a self-expressive category (MUnderdog_ High_Autonomy_Self-expressive = 5.49, (Warren et al., 2018). In this paper we extend this research by ad- MTopdog_ High_Autonomy_Self-expressive = 3.26, MUnderdog_ High_Autonomy_Functional = dressing an overarching research question: when and why is high 4.48, MTopdog_ High_Autonomy_Functional = 3.89, MUnderdog_ Low_Autonomy_Self-expressive autonomy cool? = 4.28, MTopdog_ Low_Autonomy_Self-expressive = 3.28, MUnderdog_ LowAutonomy_Functional We contribute to the literature on coolness by linking highly = 4.37, MTopdog_ Low_Autonomy_Functional = 3.88, F(1, 441) = 4.32, p <.05). Fi- autonomous behaviors to the disadvantaged financial and social nally, in Experiment 4 we show how the link between disadvantaged backgrounds a brand can present in its biography. Such backgrounds background and high autonomy works also in a political marketing 1050 / Coolness: Identifying What Makes People, Products, and Brands Cool context. MTurk participants (N = 207) rated a politician with a dis- (characteristic type: cool vs. good) x 2 (characteristic level: present advantaged background and a highly autonomous slogan as more vs. absent) ANOVAS with each of the 15 characteristics as depen- cool than his advantaged counterpart (MDisadvantaged_high_autonomy = 4.57, dent variables. These analyses tested whether a characteristic was as-

MAdvantaged_ high_autonomy = 3.42, MDisadvantaged_low_autonomy = 3.51, MAdvantaged_ sociated with both coolness and goodness (indicated by a main effect low_autonomy = 3.77; F(1, 206) = 17.06, p < .001). of “characteristic level”) or more (or less) associated with coolness This research extends the literature on coolness, demonstrating than goodness (indicated by an interaction). how consumers’ evaluation of advertising clues (such as a highly Consistent with the literature (Dar Nimrod et al. 2012; Horton autonomous positioning) can be considered cooler under certain et al. 2012), most of the characteristics associated with coolness were conditions (i.e disadvantaged company background; a high level of also associated with goodness. Participants rated both cool people consumer independence; for self-expressive more than functional and good people as being more adventurous, capable, benevolent, categories) due to the mediating role of value authenticity. On a man- warm, conscientious, agreeable, and open, and less neurotic than not- agerial level, this research demonstrates that the use of a disadvan- cool and not-good people, respectively. Interestingly, however, sig- taged brand biography in presenting the company enhances the level nificant interactions revealed that the following characteristics were of coolness in different contexts (groceries, apparel, and politics). more strongly associated with coolness than goodness: adventurous, open, and (less) neurotic. In contrast, being benevolent warm, con- What Makes People Cool? scientious, and agreeable, were more strongly associated with goodness than coolness. Additionally, we identified several characteristics EXTENDED ABSTRACT that were associated with cool people but not good people, including People want to be cool because being cool brings them the being autonomous, hedonistic, powerful, and extroverted. admiration of others and makes them feel better about themselves Interestingly, most of the characteristics that American partici- (Heath and Potter 2004; Quartz and Asp 2015). In turn, people em- pants more strongly associated with coolness than goodness, includ- ulate and imitate the behaviors of cool people in an attempt to be ing being autonomous, adventurous, open, hedonistic, powerful, and seen as cool (Runyan, Noh, and Mosier 2012; Warren and Campbell non-neurotic, signal that the person has the ability or motivation to 2014). Recognizing the power of coolness in the marketplace, mar- do what they want rather than follow the expectations or desires of keters pour money into consultants, such as cool hunters and trend others (Warren and Campbell 2014). The other characteristic that is spotters, who can ostensibly help them identify and develop cool perceived to be cool but not good was extroversion, which suggests products (Gladwell 1997; Southgate 2003). Consequently, identify- that cool people not only have the freedom to do what they want, ing the characteristics and traits that distinguish cool people from they are not loners. They remain highly interactive and connected uncool people is an important task for marketers and consumer re- with others, despite having the ability to do their own thing. searchers. Our second study (N = 335, USA, MTurk) examined whether Initial research finds that cool people possess a variety of de- measured cultural differences, including Hofstede’s dimensions (col- sirable characteristics (Dar-Nimrod et al. 2012; Horton et al. 2012; lectivism, uncertainty avoidance, power distance, and masculinity; Warren, Pezzuti, and Koley 2018). However, if coolness is a distinct Yoo et al. 2011), materialism (Richins and Dawson 1992), and tight- trait, rather than merely another way of saying that someone is desir- ness-looseness (at the state level; Harrington and Gelfand 2014), able or good, then there should be a set of characteristics that dis- influence which characteristics participants associate with cool and tinguish cool people both from people perceived to be not cool and good people. The study replicated the findings from study 1. Inter- from people perceived to be good. estingly, it also showed that the characteristics that distinguish cool Our research had three objectives. One, to identify the person- people from not-cool people remained fairly stable across the mea- ality traits and characteristics that differentiate cool people from sured cultural differences. For example, even though participants uncool people. Two, to determine whether these characteristics are from relatively tight states (e.g., Mississippi, Oklahoma) perceived merely desirable (i.e., they also differentiate good people from bad autonomy as being less good than participants from loose states (e.g., people) or whether they distinctly apply to cool people. Three, ex- Oregon, California), all participants perceived autonomy as being amine whether the characteristics that are perceived to be cool vary cool. Similarly, even though participants who scored higher on un- across cultures. certainty avoidance perceived conformity as being more good than To distinguish the characteristics that are perceived to be cool participants who scored lower on uncertainty avoidance, participants from characteristics perceived to be good, we designed three studies did not associate coolness with conformity, regardless of their score in which we asked participants from the USA and Chile to nominate a on this measure. non-famous person who they know and like but who they also either Our third study, which recruited undergraduate students from consider good, not good, cool, or not cool using a 2 (characteristic Chile (N = 241; survey in Spanish) and the USA (N = 178, survey type: cool or good) by 2 (characteristic level: present vs. absent) be- in English), replicated the results in a cross-cultural sample. To ex- tween-subjects design. Participants evaluated this person on 15 char- amine whether the results are similar in Eastern cultures, we are in acteristics from the Portrait Values Questionnaire (autonomy, hedo- the process of recruiting participants from India for a fourth study. nism, excitement seeking, power, achievement, conformity, warmth In sum, the results of three studies provided consistent evidence [caring for friends and family], benevolence [appreciation, tolerance, that the characteristics that make people seem cool are not the same and protection for the welfare of all people and nature], concern for as the characteristics that make people seem good. In three studies security, and traditionalism [respect, commitment and acceptance of in two culturally distinct countries, autonomy, adventurousness, cultural traditions; Schwartz et al. 2001) and the Big Five personality openness, hedonism, power, a lack of neuroticism, and extroversion traits (openness, conscientiousness, extroversion, agreeableness, and differentiated cool people from uncool people more than they dif- neuroticism; Gosling, Rentfrow and Swann 2003). ferentiated good people from not-good people. Moreover, although Our first study recruited USA participants (N = 287) on MTurk. the characteristics that are seen as good vary depending on cultural To determine which characteristics are most useful for differentiat- differences like tightness-looseness and uncertainty avoidance, the ing people that are cool from people that are uncool, we estimated 2 Advances in Consumer Research (Volume 48) / 1051 characteristics that make people seem cool appear to be relatively Pieters, Rik (2017), “Meaningful Mediation Analysis: Plausible stable across cultures. Causal Inference and Informative Communicatio,” Journal of Consumer Research, 44, 692–716. REFERENCES Quartz, Stephen and Anette Asp (2015), “Cool: How the Brain’s Belk, Russell. W., Kelly Tian, and Heli Paavola (2010), Hidden Quest for Cool Drives our Economy and Shapes Our “Consuming Cool: Behind the Unemotional Mask,” in World,” Farrar, Straus and Giroux. Research in Consumer Behavior, ed. Russell W. Belk, Emerald Richins, Marsha L. and Scott Dawson (1992), “A Consumer Values Group Publishing Limited, 183–208. Orientation for Materialism and its Measurement: Scale Danesi, Marcel (1994), Cool: The Signs and Meanings of Development and Validation,” Journal of Consumer Research Adolescence, University of Toronto Press. 19, 303–316. Dar-Nimrod, Ilan, I. G. Hansen, T. Proulx, and D. R. Lehman Runyan, R. C., Noh, M., and Mosier, J. (2013). What is cool? (2012), “Coolness: An Empirical Investigation,” Journal of Operationalizing the Construct in an Apparel Context. Journal Individual Differences, 33, 175–185. of Fashion Marketing and Management: An International Gladwell, Malcolm (1997), “The Coolhunt,” The New Yorker Journal, 17, 322–340. (March 17), 78-87. Runyan, Rodney C., Mijeong Noh, and Jon Mosier (2013), Gosling, Samuel D., Peter J. Rentfrow, and William B. Swann Jr. “What Is Cool? Operationalizing the Construct in an Apparel (2003), “A Very Brief Measure of the Big-Five Personality Context,” Journal of Fashion Marketing and Management, 17, Domains,” Journal of Research in Personality, 37, 504–528. 322–340. Harrington, Jesse R. and Michele J. Gelfand (2014), “Tightness– Schwartz, Shalom H., Gila Melech, Arielle Lehmann, Steven Looseness Across the 50 United States,” Proceedings of the Burgess, Mari Harris, and Vicki Owens (2001), “Extending the National Academy of Sciences, 111, 7990–7995. Cross-Cultural Validity of the Theory of Basic Human Values Heath, Joseph and Andrew Potter (2004), “Nation of Rebels: Why with a Different Method of Measurement,”Journal of Cross- Counterculture Became Consumer Culture,” HarperCollins. Cultural Psychology, 32, 519–542. Hebdige, Dick and Andrew Potter (2008), “A Critical Reframing Shin, Donghee, and Frank Biocca (2018), “Exploring immersive of Subcultural Cool and Consumption,” in European experience in journalism,” New Media & Society, 20(8), Advances in Consumer Research, ed. Stefania Borghini, Mary 2800–2823. Ann McGrath, and Cele Otnes, Association for Consumer Southgate, Nick (2003), “Coolhunting, Account Planning and Research, 527–528. the Ancient Cool of Aristotle,” Marketing Intelligence & Horton, Matthew, Janet C. Read, Daniel Fitton, Linda Little, and Planning, 21 (7), 453-461. Nicola Toth (2012), “Too Cool at School: Understanding Cool Spencer, Steven J., Mark P. Zanna, and Geoffrey T. Fong (2005), Teenagers,” PsychNology Journal, 10, 73–91. “Establishing a Causal Chain: Why Experiments are Often Im, Subin, Subodh Bhat, and Yikuan Lee (2015), “Consumer More Effective than Mediational Analyses in Examining Perceptions of Product Creativity, Coolness, Value and Psychological Processes.” Journal of Personality and Social Attitude,” Journal of Business Research, 68 (1), 166–72. Psychology 89, 845–851. Kerner, Noah and Gene Pressman (2007), “Chasing Cool: Standing Tapp, Alan and Sara Bird (2008), “Social Marketing and the out in Today’s Cluttered Meaning of Cool,” Social Marketing Quarterly, 14, 18–29. Marketplace,” Atria. Van Den Bergh, Joeri and Mattias Behrer (2011), “How Cool Kim, Ki Joon, Dong-Hee Shin, and Eunil Park (2015), “Can Brands Stay Hot: Branding to Generation Y,” Kogan Page coolness predict technology adoption? Effects of perceived Publishers. coolness on user acceptance of smartphones with curved Warren, Caleb and Margaret C. Campbell (2014), “What Makes screens,” Cyberpsychology, Behavior, and Social Networking, Things Cool? How Autonomy Influences Perceived Coolness,” 18 (9), 528-533. Journal of Consumer Research, 41, 543–563. MacAdams, Lewis (2001), Birth of The Cool: Beat, Bebop, and the Warren, Caleb (2010), “What Makes Things Cool and Why American Avant Garde. Free Press, New York. Marketers Should Care,” Proquest Umi Dissertation Markus, Hazel R. and Shinobu Kitayama (1991), “Culture and the Publishing. Self: Implications for Cognition, Emotion, and Motivation,” Warren, Caleb, Todd Pezzuti, and Shruti Koley (2018), “Is Being Psychological Review, 98, 224–53. Emotionally Inexpressive Cool?” Journal of Consumer Nancarrow, Clive, Pamela Nancarrow, and Julie Page (2002), Psychology, 28, 560–577. “An Analysis of the Concept of Cool and Its Marketing Warren, Caleb and Martin Reimann (2019), “Crazy-funny-cool Implications,” Journal of Consumer Behavior, 1, 311–322. theory: Diverging reactions to unusual product designs,” Newman, George E. and Rosanna K. Smith (2016), “Kinds of Journal of the Association of Consumer Research, 4 (4), 409- Authenticity,” Philosophy Compass, 11 (10), 609–618. 421. O’Donnell, Kathleen, A. and Daniel L. Wardlow (2000), “A Theory Yoo, Boonghee, Naveen Donthu, and Tomasz Lenartowicz (2011), of the Origins of Coolness,” Advances in Consumer Research, “Measuring Hofstede’s Five Dimensions of Cultural Values 27, 13–18. at the Individual Level: Development and Validation of Paharia, Neeru, Anat Keinan, Jill Avery, and Juliet B. Schor (2011), CVSCALE,” Journal of International Consumer Marketing, “The Underdog Effect: The Marketing of Disadvantage 23, 193–210. and Determination through Brand Biography,” Journal of Consumer Research, 37, 775–790. Consumer Consequences of Robotic Replacement Chairs: Emanuel de Bellis, University of St. Gallen, Switzerland Bernd Schmitt, Columbia University, USA

Paper #1: Consumer Preferences for Human Versus Robotic on adaptive algorithms are perceived as more creative whereas those Labor: The Role of Symbolic Consumption based on pre-programmed algorithms are perceived as more predict- Stefano Puntoni, Erasmus University Rotterdam, The able, showing that also the “inner values” of robotic technologies Netherlands influence their adoption by consumers. Armin Granulo, Technical University of Munich, Germany Taken together, this session is intended to deepen our knowl- Christoph Fuchs, Technical University of Munich, Germany edge on the psychological consequences of robotic replacement from a consumer perspective. The papers utilize multi-method approach- Paper #2: This Robot Doesn’t Judge Me – Service Robots and es, combining data from online, lab, and field studies to explore the the Choice of Embarrassing Products robustness and generalizability of the proposed effects. This session Jenny van Doorn, University of Groningen, The Netherlands speaks to ACR’s call for special sessions on the impact of technology Jana Holthoewer, University of Groningen, The Netherlands on consumption by highlighting the diverse consequences of replac- Paper #3: How Robotic Customer Service Impacts Consumers’ ing humans by robotic technologies. This session should not only be Evaluation of the Firm of interest to scholars working on new technologies, AI, and robotics Noah Castelo, University of Alberta, Canada but also to anybody interested in symbolic consumption, emotion, Johannes Boegershausen, University of Amsterdam, The and creativity. Netherlands Christian Hildebrand, University of St. Gallen, Switzerland Consumer Preferences for Human Versus Robotic Labor: Alex Henkel, Open University, The Netherlands The Role of Symbolic Consumption Paper #4: Inner Values Also Count: Algorithm Types Drive Consumer Adoption of Robotic Technologies EXTENDED ABSTRACT Melanie Clegg, University of Lucerne, Switzerland Advances in robotics, automation, and artificial intelligence Reto Hofstetter, University of Lucerne, Switzerland increasingly enable firms to replace human labor with modern tech- Emanuel de Bellis, University of St. Gallen, Switzerland nology, fundamentally transforming how goods and services are pro- Bernd Schmitt, Columbia University, USA duced. The consequences of these developments for the demand for human labor are hotly debated in both academia and popular press. SESSION OVERVIEW Within the social sciences, the debate has been most intense within Advances in artificial intelligence (AI) and robotics enable economics with a focus on when, that is, for which tasks, human firms to replace humans with modern technology, leading to a radi- labor can potentially be replaced by robotic labor (e.g., Autor, Levy, cal transformation in the production and consumption of goods and and Murnane 2003). Despite the relevance of these developments to services. In the service domain, robots are used to communicate consumers—robots and algorithms are transforming consumer-firm and interact with consumers in a wide variety of contexts, including interactions in many industries—consumer researchers have so far stores, restaurants, and hotels. By 2025, it is predicted that 95% of all provided little insight into how consumers react to today’s robotic customer interactions will be powered by some form of AI or robot- revolution (cf. Granulo, Fuchs, and Puntoni 2019). ics. While research has started to shed light on this transformation In this research, we examine consumer acceptance of human (Castelo, Bos, and Lehmann 2019; de Bellis and Johar 2020; Granu- versus robotic labor—and how this acceptance depends on product lo, Fuchs, and Puntoni 2019; Mende et al. 2019; Schmitt 2019), we characteristics. Specifically, this research investigates for which spe- are only beginning to understand its multifaceted consequences for cific products, services, and product features consumers are more consumers. The four papers in this session aim to improve this unlikely to favor human over robotic labor. In seven studies using dif- derstanding by answering the following key question: What are the ferent products, designs, and samples, we predict and demonstrate psychological consequences of robotic replacement for consumers? that consumers value human (vs. robotic) labor more for products The first paper by Puntoni, Granulo, and Fuchs examines with higher symbolic value (i.e., where expressing something about for which products consumers prefer robotic versus human labor. It one’s beliefs and personality is of central importance; e.g., Belk shows that consumers value human (vs. robotic) labor more for prod- 1988; Reed et al. 2012). We theorize that human (vs. robotic) labor ucts with higher symbolic value, as it lends products unique quali- creates value by lending products unique qualities in the eyes of con- ties in the eyes of consumers. The second paper by Van Doorn and sumers. Consumers should value these unique qualities (e.g., Brock Holthoewer examines for which products consumers prefer robotic 1968; Tian and McKenzie 2001) more for products with higher sym- versus human service providers. It shows that consumers are more bolic value, resulting in greater preferences for human (vs. robotic) accepting of a robotic service provider when acquiring embarrass- labor in the case of products with higher symbolic value. We provide ing products because they feel less judged by a robot than by a hu- support for this theoretical framework by documenting (i) greater man. The third paper by Castelo et al . examines the downstream consumer preference for human (vs. robotic) labor in more symbolic consequences of robotic versus human service providers for firms. It consumption contexts, (ii) that this effect is mediated by product shows that consumers perceive firms that use robotic service provid- uniqueness, controlling for various other alternative accounts (e.g., ers as less customer-centric, which in turn decreases service satis- love, product importance, and product quality), (iii) that the effect is faction and willingness to recommend the firm to others. Whereas moderated by consumers’ need for uniqueness, and (iv) that it can be the first three papers contrast robots with humans, the fourth paper observed even when there is no direct (physical) contact in the pro- by Clegg et al . examines different types of algorithms that are the duction process (differentiating the effect from work on handmade backbone of robotic technologies. It shows that technologies based products; Fuchs et al. 2015).

Advances in Consumer Research 1052 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1053 To test our main prediction that consumers favor human (vs. ro- This Robot Doesn’t Judge Me – Service Robots and the botic) labor more in symbolic consumption contexts, we conducted Choice of Embarrassing Products five studies focusing on the symbolic value of different products, services, consumption contexts, and product components. In study 1 EXTENDED ABSTRACT (N = 72, students), we asked participants to rate a series of products During the last years, there has been significant progress in ser- on symbolic value or on relative preference for human labor. Prefer- vice automation, robots, and artificial intelligence (AI; Kahn et al. ence for human (vs. robotic) production was highest for product cat- 2013). Service robots have started to communicate and interact with egories whose symbolic value was in the upper third (e.g., jewelry or customers in a frontline service setting. With a 61% sales increase suits), and lowest for products whose symbolic value was in the low- of service robots for professional use from 2017 to 2018 (Industrial er third (e.g., toolsets or car brakes; 20 product categories in total). In Federation of Robotics 2019), it is estimated that 95% of all cus- study 2A (N = 144, students), we find that consumers prefer human tomer interactions will be powered by AI by 2025 (Servion Global (vs. robotic) labor more when the symbolic value of a service is high Solutions 2018). Service robots distinguish themselves from previ- (i.e., getting a tattoo) than when the symbolic value of a service is ous generations of technology and automation because they often low (i.e., getting a tattoo removed; Mhigh = 5.03 vs. Mlow = 4.03, t(142) go beyond a merely functional role and engage their users on a so- = 3.65, p < .0001).1 In study 3A (N = 322, MTurk), we document that cial level. The degree to which technology makes consumers feel consumers are more likely to purchase products designed by a hu- the presence of another social entity is known as automated social man (vs. algorithm)—but produced by a machine—when they have presence (Van Doorn et al. 2017). Yet, previous literature shows that more symbolic consumption goals (i.e., purchasing a printed poster service robots are not always easily accepted. On the contrary, they to decorate; Mhuman = 4.39 vs. Malgorithm = 2.30) than when they have can trigger negative feelings and compensatory behaviors (Mende less symbolic consumption goals (i.e., purchasing a printed poster to et al. 2019), and cause more discomfort than dealing with a human educate; Mhuman = 4.30 vs. Malgorithm = 3.25; F(1, 318) = 5.20, p < .05). employee (Čaić et al. 2019). Given that service robots are on the And in study 4A (N = 124, MTurk), we document that consumers rise, it is imperative to identify circumstances where the negative choose human (vs. robotic) production for more symbolic product effects documented in previous literature are attenuated, or can even components (i.e., the frame of eyeglasses; 59% vs. 41%, Z = 2.00, p be avoided altogether. In this paper, we investigate whether service < .05) but not for less symbolic product components (i.e., the lenses robots can help overcome consumer reluctance to acquire embar- of eyeglasses; 23% vs. 77%, Z = 6.01, p < .0001, = 35.2, p < .0001). rassing products. To test our prediction that uniqueness can explain why consum- Social identity theory posits that unwanted negative evaluations ers prefer human (vs. robotic) labor more in symbolic consumption from audiences can cause a threat to one’s social identity (Edelmann contexts, we conducted two additional studies. Specifically, we test- 1987; Miller and Leary 1992) and influence patronage decisions ed whether this relative preference is moderated by individuals’ need (Moore et al. 2006; Grace 2009). In embarrassing service situa- for uniqueness, and whether it is mediated by product uniqueness. tions, consumers endanger their health, engage in risky behavior, In study 3B (N = 402, MTurk), we find that consumers value human and avoid seeking medical care in order to avoid social judgment (vs. robotic) labor more strongly for products with higher symbolic (Helweg-Larsen and Collins 1994; Lund-Nielsen et al. 2011; Kiefe value when they have a higher need for uniqueness (b = 0.724, (394) et al. 1998). We argue that robots may reduce this feeling and thus = 2.08, p < .05; 95% CI: [0.04, 1.41]). Specifically, the interaction make it easier for consumers to, for example, collect medication that between production mode and symbolic value was non-significant they experience as embarrassing. Hence, we extend the concept of for individuals low in need for uniqueness (F(1, 133) = 0.57, p = social presence, that has shown to influence embarrassment in pur- .451) and highly significant for respondents high in need for unique- chase situations (Argo et al. 2005; Dahl et al. 2001), to automated so- ness (F(1, 123) = 8.78, p < .01). In study 4B (N = 201, Prolific), cial presence, therewith taking into account novel ways of purchase we find that consumers value human (vs. robotic) more for products interactions that come with the development of new technology. with higher symbolic value because they want more unique products We furthermore explore the role of anthropomorphism in shap- (indirect effect: b = 0.222, SE = 0.09, Z = 2.42, p < .05; 95% CI: ing how consumers react to service robots. Anthropomorphization [0.08, 0.45]). means imbuing the “behavior of nonhuman agents with human-like In summary, our research demonstrates that human labor cre- characteristics, motivations, intentions, or emotions” (Epley, Waytz, ates additional economic value in consumption contexts where and Cacioppo 2007, p. 864). Although humanizing robots is the ul- symbolism plays a central role. By doing so, we complement the timate goal in robotics (Rubin 2003), there is a certain risk that con- supply-side perspective in the economics literature (e.g., for which sumers feel uncomfortable in dealing with very human-like robots. tasks human labor can and cannot be replaced by firms) with a de- This phenomenon is known as the “uncanny valley” (Mori, MacDor- mand-side perspective (e.g., for which products human labor can and man, and Kageki 2012). We argue that very human-like robots might cannot create value for consumers). Moreover, we demonstrate that increase consumers’ apprehension of judgments by anthropomor- one important driver behind our findings is that human (vs. robotic) phizing this attribute to the robot, leading to a less positive consumer labor lends products unique qualities in the eyes of consumers. Our evaluation (Miller 1995). Thus, consumers may prefer less anthro- findings suggest that the extent to which human labor will continue pomorphic robotic service providers in an embarrassing situation. playing a role in production processes in the years to come should In three studies, we find that consumers are indeed more accept- not only depend on cost and efficiency, but also on the inherent char- ing of robotic service providers in embarrassing situations. In study acteristics of products. 1 (N=136, 47 males) involving real choice behavior participants watched a video featuring a human vs. robotic ambassador promoting a new health initiative and offering a free giveaway. The giveaway was either a pack of condoms (high embarrassment) or hand sanitizer (low embarrassment). The percentage of participants taking 1 In study 2B (N = 192, MTurk), we replicate this effect in a 2(symbolic: the pack of condoms rather than refusing it rose from 38% for the high vs. low) × 2(labor: human vs. robotic labor) between-participants design (F(1, 188) = 6.40, p < .05). 1054 / Consumer Consequences of Robotic Replacement human ambassador to 61% for the robotic ambassador ((1)=3.353, We propose that encountering service robots activates a lay p=.067). theory that automation is motivated by profit maximization at the ex- Study 2 is a 2 (service provider: human vs. robot) × 2 (em- pense of the customer experience. Since firms have finite resources, barrassment product: low vs. high) between-subjects Prolific study. consumers rely on a “zero-sum heuristic” (Chernev 2007), such that Respondents (N=395, 174 males) imagine buying antibiotics for investments into service robots may be perceived as coming at the an ear infection (low embarrassment) or a sexually transmitted expense of a worse customer experience. disease (high embarrassment) from a human vs. robot at their lo- Firms using service robots should therefore be seen as less cus- cal pharmacy. The results reveal a main effect of service provider tomer-centric than firms that do not, which should in turn decrease (F(1,394)=27.209, p<.001) and a two-way interaction between ser- consumers’ service satisfaction and willingness to recommend the vice provider and product embarrassment (F(1,394)=5.066, p=.025). firm to others. Participants acquiring non-embarrassing products had higher loyalty These effects should be attenuated when the firm’s perceived intentions when there was a human (M=5.04) rather than a robotic motivation for using the automation is to overcome current short- service provider (M=3.93; F(1,391)=27.943, p<.001). Yet, the prefer- comings in customer service. ence for a human service provider was less pronounced when buying Study 1: In a field study over four days, we varied whether 109 embarrassing products (Mhuman=4.65, Mrobot=4.21; F(1,391)=4.387, customers in a café placed their order with a Pepper robot or with p=.037). This indicates that product embarrassment attenuates the a human. After ordering, customers rated their agreement with the negative effect of a robot. As expected, social judgment acts as a me- statements “I would recommend this coffee bar to a friend or col- diator in that consumers feel less judged by a robot when acquiring league” and “I am satisfied with my service experience today” (α = embarrassing products, which again drives their loyalty intentions .92). These firm outcomes were worse in the robot condition (M =

(Mrobot=3.81, Mhuman=4.38; F(1,391)=8.684, p=.003, 95%CI: [0.020, 5.34) than in the human condition (M = 5.81, F(1,107) = 4.89, p = 0.206]). .029). The difference remained significant controlling for technology Study 3 (N=795, 376 males) extends the previous study and affinity, age, gender, and frequency of visiting the café. explores the impact of the level of anthropomorphism of the robot Study 2: 238 Prolific participants engaged in an online chat with (machine-like, human-like, very human-like). When acquiring non- a chatbot but were told that they were interacting either with a bot or embarrassing products, participants express higher loyalty inten- with a human. This keeps the interaction identical across conditions. tions when served by a human (M=5.10) than by any robot (Mma- were told that we were testing new customer service platforms for a chine-like=4.24, Mhuman-like=4.37, Mveryhuman-like=3.93; F(3,787)=10,733, telecommunications company and that they would interact with one p<.001). Buying embarrassing products attenuates this consistent of the company’s representatives. After the interaction, we measured preference for the human service provider. Particularly, participants service satisfaction and recommendation willingness as in Study 1. are equally likely to be loyal to the pharmacy when served by a We also measured perceived customer centricity of the firm (Habel human service provider (Mhuman=4.60) or a more anthropomorphic et al. 2019). robot (Mhuman-like=4.45, p=.490; Mveryhuman-like=4.37, p=.297). Counter- The firm seemed less customer-centric when participants being our expectations, lowest loyalty is expressed towards the least lieved they were chatting with a bot than with a human (MChatbot = anthropomorphic robot (Mmachine-like =4.14; p=.034), although a mod- 4.75, MHuman = 5.28; t(236) = 2.66, p = .008), and consumers had erated mediation analysis reveals that consumers feel less socially lower satisfaction and willingness to recommend the firm (MChatbot judged by it (M=3.75; 95%CI: [0.003, 0.130]) compared to a hu- = 4.87, MHuman = 5.38; t(236) = 2.47, p = .014). A mediation model man (M=4.54) when acquiring embarrassing products. Consumers with bootstrapped estimates (5000 iterations) confirmed the positive feel also less judged by a human-like robot (M=3.76; 95%CI: [0.003, effect of customer centricity on firm outcomes (β = .79, p < .001), 0.126]), yet the very human-like robot is perceived as equally judg- switching off the main effect of condition (β = -.09, p = .49), and an ing as the human (M=4.03; 95%CI: [-0.001, 0.093]). indirect effect excluding zero (-.42, 95% CI = -.71, -.11), indicating This paper contributes to literature by showing an impor- full mediation. tant boundary condition to earlier work cautioning that consumers Study 3: We used a platform called Chatplat, which allows are reluctant to accept robotic service providers (Čaić et al. 2019; researchers to connect participants to engage with text-based chat Mende et al. 2019). Importantly, we find that consumers are more with each other. We recruited 1076 participants from Prolific and as- accepting of robotic service providers when acquiring embarrassing signed them to either the “employee” role or the “customer” role. All products. As expected, social judgment mediates the relationship in participants were told we were working with a telecommunications that consumers feel less judged by a robot than by a human. Given company to test new customer service platforms and were given a that consumers feel less judged by a moderately anthropomorphic script to follow, which involved the customer asking about changing robot compared to a human, yet express equally high loyalty inten- a mobile phone plan and the employee asking for their customer ID tions when acquiring an embarrassing product, a moderately anthro- and providing the requested information. pomorphic robot seems a better choice when selling embarrassing Customers were further divided into three conditions. In the products than a machine-like or very human-like robot. human condition, they were told they would chat with a customer service employee from the company. In the chatbot condition, they How Robotic Customer Service Impacts Consumers’ were told they would chat with a chatbot. In the customer-centric Evaluation of the Firm chatbot condition, they were told that they would chat with a chatbot and that the company’s goal in using chatbots was to “reduce wait EXTENDED ABSTRACT times and improve service for customers.” This condition also stated Robots are being used to provide customer service online and that since introducing the chatbot, “the average wait time for cus- in stores. However, only 15% of firms that currently use the market- tomer to connect with an agent has decreased from over 5 minutes to leading Pepper service robot plan to renew their contracts (Whitton less than 10 seconds.” 2018). We use real customer-machine interactions to shed light on In reality, the “employee” with whom the “customers” chatted the challenges facing the use of service robots. was always another Prolific participant. Our interest was in com- Advances in Consumer Research (Volume 48) / 1055 paring participants’ experience across the three customer conditions. Morewedge 2019) while highlighting additional barriers to consum- After the interaction, participants again reported service satisfaction er adoption of autonomous technologies (de Bellis and Johar 2020). and recommendation likelihood. Based on computer science literature we distinguish two funda- We first cleaned the data by conducting several exclusions. We mentally different types of algorithms: adaptive (whose parameters excluded the 541 participants in the “employee” role given our inter- can be changed by using external data) and pre-programmed (whose est in the “customers’” experiences. We excluded 141 participants parameters are fixed in-advance by a programmer). We predict that who reported having technical difficulties during the chat; these were adaptive (vs. pre-programmed) algorithms increase the intention to inevitable since participants matched with each other in real time to use a technology (usage intention) because they are perceived as conduct the chat and thus some participants had to wait too long for more creative (H1A, H1B). We further predict that this effect re- their chat partner and could not complete the chat in a timely manner. verses when the tasks of a technology require only a low degree of Finally, we excluded 72 participants who did not correctly remem- creativity and specific outcomes are valued (H2). In such cases, a ber whether they had chatted with a human employee or a chatbot pre-programmed algorithm increases usage intention, because it is and 111 participants who did not believe that they had chatted with perceived as more predictable (H2-H3). a human or chatbot despite what they had been told. Final sample The pilot study applied a Turing test paradigm (Turing sizes per condition were 55 in the human condition, 86 in the regular 1950/2009) to test whether algorithm type influences consumer per- chatbot condition, and 83 in the customer-centric chatbot condition. ceptions independently from a technology’s output. According to the Firm outcomes were best in the human condition (M = 8.47), Turing test, a machine is considered intelligent if its communication worse in the regular chatbot condition (M = 6.99, t(139) = 3.96, p < behavior is indistinguishable from human behavior. In our version,

.001), and also worse in the customer-centric chatbot condition but we let participants (US-MTurkers, N=192, Mage=34.14, 46% female) to a lesser degree (M = 7.56, t(135) = 2.41, p = .016). The differ- chat with an advanced chat bot. Though all participants chatted with ence between the two chatbot conditions was marginally significant the same bot (thus holding output constant), we experimentally ma- (t(167) = 1.75, p = .087). nipulated information about the bot’s implemented algorithm (pre- Taken together, these results show that firms using service ro- programmed vs. adaptive). We found that participants rated the bot bots to automate customer service are perceived as less customer- as more intelligent (Madaptive=4.10, Mpre-programmed=3.49, F(1, 190)=5.25, centric and that consumers are less satisfied with and less likely to p=.04) and more human-like (Madaptive=3.73, Mpre-programmed=2.80, F(1, recommend such firms to others, relative to firms using humans. 190)=16.34, p<.001) if they were told it used an adaptive (vs. pre- Furthermore, we show that firms can reduce these negative conse- programmed) algorithm. This indicates that not only the output but quences by highlighting a customer-centric motivation for using ro- also the implemented algorithm matters for consumers. bots. Additional completed studies show that using service robots for Study 1 formally tests how algorithm type shapes consum- less customer-centric tasks (i.e., food delivery vs. waiter) can also ers’ perception of and intention to use a technology (H1-H2). In a reduce these effects. lab study conducted at a large Northeastern university, participants

Our findings help explain why initial rollouts of service robots (N=207, Mage=24.55, 57% female) were informed about the algorithm have been lackluster and have clear implications for how firms can implemented in a voice assistant (adaptive vs. pre-programmed, improve outcomes while continuing to pursue the use of this tech- 2-cell between subjects), before indicating their intention to use it. nology. A one-way ANOVA revealed that usage intention was significantly increased if participants were informed that the voice assistant ap-

Inner Values Also Count: Algorithm Types Drive plies an adaptive (vs. pre-programmed) algorithm (Madaptive=4.70,

Consumer Adoption of Robotic Technologies Mpre-programmed=4.05, F(1, 205)=6.50, p=.01). This effect is mediated by perceived creativity of the algorithm (b=.61, 95%CI=[.34; .94]). EXTENDED ABSTRACT Note that we did not find a mediation via predictability (b=-.02, Algorithms are implemented in all robotic technologies. These 95%CI=[-.20; .14]) because a voice assistant is considered a creative technologies take care of our elderly (Bemelmans et al. 2012), serve technology. us food (Mende et al. 2019), and undertake domestic work (Rijsdijk Study 2 tests whether task creativity (i.e., the creativity required and Hultink 2009). One major obstacle to accept robotic technolo- by the tasks of a technology) moderates the effect of algorithm type gies (e.g., as co-workers or employees) is lacking transparency and on usage intention (H1-H3). We conducted an online study with US- informedness about their true technological underpinnings (Moreau, MTurkers (N=395, Mage=35.71, 48% female) and employed a 2×2 Lehmann, and Markman 2001; Rai 2020), since their “inner values” between-subjects design. Participants were informed about the algo- oftentimes remain opaque for consumers. Therefore, firms increas- rithm implemented in a cooking app (adaptive vs. pre-programmed), ingly start to communicate details about algorithms of technolo- before indicating their intention to use it. We additionally manipu- gies to their consumers (e.g., the app FitnessAI or Samsung’s smart lated the task for which the app should be used (high creativity: bak- fridge; FitnessAI 2019; Samsung 2018). However, no research so ing a completely novel cake vs. low creativity: baking a cake exactly far has addressed the question of how this information can influence by recipe). A two-way ANOVA revealed a significant interaction be- consumer behavior. tween both factors (F(1, 391)=14.61, p<.001). Follow-up contrasts This research examines whether and how information about the confirmed that usage intention is significantly higher for the adap- details of implemented algorithms can shape consumers’ technology tive (pre-programmed) algorithm when task creativity is high (low). adoption. It therewith complements former research that presumably Parallel moderated mediation analysis supported H2 and H3: The takes a materialistic perspective focusing either on the outer appear- mediation via perceived algorithm creativity on usage intention is ance (Kim, Schmitt, and Thalmann 2019; Mende et al. 2019), or the significantly higher when task creativity is high (index -of moder observable output or attributes of a technology (Mukherjee and Hoy- ated mediation: b=.40, 95%CI=[.17; .67]), while a mediation via per- er 2001). It further extends research on algorithm acceptance that ceived predictability is significant only when task creativity is low mainly measured lay perceptions of algorithms and AI as generic (index of moderated mediation: b=.59, 95%CI=[.30; .93]). concepts (Castelo, Bos, and Lehmann 2019; Longoni, Bonezzi, and 1056 / Consumer Consequences of Robotic Replacement

Studies 3A and 3B generalize the findings to a broader range Dahl, Darren W., Rajesh V. Manchanda, and Jennifer J. Argo of products (H1-H3). In study 3A, we let US-MTurkers (N=181, (2001), “Embarrassment in Consumer Purchase: The Roles

Mage=35.46, 52% female) evaluate either task creativity or predict- of Social Presence and Purchase Familiarity,” Journal of ability of 54 robotic products. A separate MTurk sample (N=131, Consumer Research, 28 (3), 473–481.

Mage=34.32, 48% female) received information about both algorithm de Bellis, Emanuel and Gita Venkataramani Johar (2020), types and indicated in a binary choice which they preferred for each “Autonomous Shopping Systems: Identifying and Overcoming of the 54 products. Analyses revealed a positive correlation between Barriers to Consumer Adoption,” Journal of Retailing, 96 (1), products’ task creativity and the ratio of participants voting for an 74–87. adaptive algorithm in the binary choice (r=.83, p<.001). Predict- Edelmann, Robert J. (1987), The Psychology of Embarrassment, ability ratings correlate much less with the ratio of adaptive choices John Wiley & Sons. (r=.23, p=.084). Corroborating these findings, we calculated the sen- Epley, Nicholas, Adam Waytz, and John T. Cacioppo timent of over 11,000 Twitter posts that discussed the 54 products (2007), “On Seeing Human: A Three-Factor Theory of in the context of AI (with AI being colloquially used for adaptive Anthropomorphism,” Psychological Review, 114 (4), 864–886. algorithms; Arel, Rose, and Karnowski 2010). Sentiment positively FitnessAI. (2019), “Get Stronger with AI,” https://www.getfitnessai. correlated with products’ creativity ratings (r=.29, p=.09), suggest- com/. ing that consumers perceive AI in creative products more positively. Fuchs, Christoph, Martin Schreier, and Stijn M. J. Van Osselaer In sum, we find that also the inner values of technologies in- (2015), “The Handmade Effect: What’s Love Got to Do with fluence consumers’ technology adoption. Interestingly, our findings It?” Journal of Marketing, 79 (2), 98–110. imply that consumers assign a degree of creativity to technologies Grace, Debra (2009), “An Examination of Consumer if they are appropriately informed about their capabilities (i.e., their Embarrassment and Repatronage Intentions in the Context adaptivity). This contradicts typical lay assumptions that creativity of Emotional Service Encounters,” Journal of Retailing and is a human-like trait which algorithms are incapable of due to their Consumer Services, 16 (1), 1–9. mechanical and analytical thinking (Huang and Rust 2018). Our Granulo, Armin, Christoph Fuchs, and Stefano Puntoni (2019), findings imply that, aside from increasing the human-like- appear “Psychological Reactions to Human versus Robotic Job ance of technologies, informing about their inner values can boost Replacement,” Nature Human Behavior, 3 (10), 1062–1069. their adoption. Habel, Johannes, Roland Kassemeier, Sascha Alavi, Philipp Haaf, Christian Schmitz, and Jan Wieseke (2019), “When REFERENCES Do Customers Perceive Customer Centricity? The Role of a Arel, Itamar, Derek C. Rose, and Thomas P. Karnowski (2010), Firm’s and Salespeople’s Customer Orientation,” Journal of “Deep Machine Learning–A New Frontier in Artificial Personal Selling and Sales Management, 39 (1), 1–19. Intelligence Research,” IEEE Computational Intelligence Helweg-Larsen, Marie and Barry E. Collins (1994), “The UCLA Magazine, 5 (4), 13–18. Multidimensional Condom Attitude Scale: Documenting the Argo, Jennifer J., Darren W. Dahl, and Rajesh V. Manchanda Complex Determinants of Condom Use in College Students,” (2005), “The Influence of a Mere Social Presence in a Retail Health Psychology, 13 (3), 224–237. Context,” Journal of Consumer Research, 32 (2), 207–212. Huang, Ming-Hui and Roland T. Rust (2018), “Artificial Autor, David H., Frank Levy, and Richard J. Murnane (2003), “The Intelligence in Service,” Journal of Service Research, 21 (2), Skill Content of Recent Technological Change: An Empirical 155–172. Exploration,” The Quarterly Journal of Economics, 118 (4), Industrial Federation for Robotics (2019), “Executive Summary 1279–1333. World Robotics 2019 Service Robots,” https://ifr.org/ Belk, Russell W. (1988), “Possessions and the Extended Self,” downloads/press2018/Executive_ Summary_WR_Service_ Journal of Consumer Research, 15 (2), 139–168. Robots_2019.pdf. Bemelmans, Roger, Gert Jan Gelderblom, Pieter Jonker, and Luc Kahn Jr, Peter H., Heather E. Gary, and Solace Shen (2013), De Witte (2012), “Socially Assistive Robots in Elderly Care: A “Children’s Social Relationships with Current and Near Future Systematic Review into Effects and Effectiveness,”Journal of Robots,” Child Development Perspectives, 7 (1), 32–37. the American Medical Directors Association, 13 (2), 114–120. Kiefe, Catarina I., Ellen Funkhouser, Mona N. Fouad, and Daniel Brock, Timothy C. (1968), “Implications of Commodity Theory for S. May (1998), “Chronic Disease as a Barrier to Breast and Value Change,” in Psychological Foundations of Attitudes (pp. Cervical Cancer Screening,” Journal of General Internal 243–275). Elsevier. Medicine, 13 (6), 357–365. Čaić, Martina, João Avelino, Dominik Mahr, Gaby Odekerken- Kim, Seo Young, Bernd H. Schmitt, and Nadia M. Thalmann Schröder, and Alexandre Bernardino (2019), “Robotic versus (2019), “Eliza in the Uncanny Valley: Anthropomorphizing Human Coaches for Active Aging: An Automated Social Consumer Robots Increases their Perceived Warmth but Presence Perspective,” International Journal of Social Decreases Liking,” Marketing Letters, 30 (1), 1–12. Robotics, 1–16. Longoni, Chiara, Andrea Bonezzi, and Carey K. Morewedge Castelo, Noah, Maarten W. Bos, and Donald R. Lehmann (2019), (2019), “Resistance to Medical Artificial Intelligence,”Journal “Task-Dependent Algorithm Aversion,” Journal of Marketing of Consumer Research, 46 (4), 629–650. Research, 56 (5), 809–825. Lund-Nielsen, Betina, Julie Midtgaard, Mikael Rørth, Finn Gottrup, Chernev, Alexander (2007), “Jack of All Trades or Master of One? and Lis Adamsen (2011), “An Avalanche of Ignoring–A Product Differentiation and Compensatory Reasoning in Qualitative Study of Health Care Avoidance in Women with Consumer Choice,” Journal of Consumer Research, 33 (4), Malignant Breast Cancer Wounds,” Cancer Nursing, 34 (4), 430–44. 277–285. Advances in Consumer Research (Volume 48) / 1057

Mende, Martin, Maura L. Scott, Jenny van Doorn, Dhruv Grewal, Rijsdijk, Serge A. and Erik Jan Hultink (2009), “How Today’s and Ilana Shanks (2019), “Service Robots Rising: How Consumers Perceive Tomorrow’s Smart Products,” Journal of Humanoid Robots Influence Service Experiences and Elicit Product Innovation Management, 26 (1), 24–42. Compensatory Consumer Responses,” Journal of Marketing Rubin, Charles T. (2003), “Artificial Intelligence and Human Research, 56 (4), 535–556. Nature,” The New Atlantis, (1), 88–100. Miller, Rowland S. and Mark R. Leary (1992), “Social Sources Samsung (2018). “‘All In’ on AI, Part 3: The Family Hub and Interactive Functions of Emotion: The Case of Recognizes You and Provides a Personalized Diet,” https:// Embarrassment,” in M. S. Clark (Ed.), Review of personality news.samsung.com/global/all-in-on-ai-part-3-the-family-hub- and social psychology, Vol. 14. Emotion and social behavior recognizes-you-and-provides-a-personalized-diet. (p. 202–221). Sage Publications, Inc. Servion Global Solutions (2018), “What Makes Emerging Miller, Rowland S. (1995), “On the Nature of Embarrassability: Technologies the Future of Customer Experience?” https:// Shyness, Social Evaluation, and Social Skill,” Journal of servion.com/blog/what-emerging-technologies-future- Personality, 63 (2), 315–339. customer-experience/. Moore, Sarah G., Darren W. Dahl, Gerald J. Gorn, and Charles B. Tian, Kelly Tepper and Karyn McKenzie (2001), “The Long-Term Weinberg (2006), “Coping with Condom Embarrassment,” Predictive Validity of the Consumers’ Need for Uniqueness Psychology, Health & Medicine, 11 (1), 70–79. Scale,” Journal of Consumer Psychology, 10 (3), 171–193. Moreau, C. Page, Donald R. Lehmann, and Arthur B. Markman Turing, Alan M. (1950/2009), “Computing Machinery and (2001), “Entrenched Knowledge Structures and Consumer Intelligence,” in Parsing the Turing Test, pp. 23-65. Springer, Response to New Products,” Journal of Marketing Research, Dordrecht. 38 (1), 14–29. Van Doorn, Jenny, Martin Mende, Stephanie M. Noble, John Mori, Masahiro, Karl F. MacDorman, and Norri Kageki (2012), Hulland, Amy L. Ostrom, Dhruv Grewal, and J. Andrew “The Uncanny Valley [From the Field],” IEEE Robotics & Petersen (2017), “Domo Arigato Mr. Roboto: Emergence Automation Magazine, 19 (2), 98–100. of Automated Social Presence in Organizational Frontlines Mukherjee, Ashesh and Wayne D. Hoyer (2001), “The Effect of and Customers’ Service Experiences,” Journal of Service Novel Attributes on Product Evaluation,” Journal of Consumer Research, 20 (1), 43–58. Research, 28 (3), 462–472. Whitton, Rian (2018), “Whiz Robot from SoftBank Arrives as Rai, Arun (2020), “Explainable AI: From Black Box to Glass Pepper Renewals Lag,” The Robot Report, https://www. Box,” Journal of the Academy of Marketing Science, 48 (1), therobotreport.com/renewals-softbank-pepper-robot-lag/. 137–141. Reed II, Americus, Mark R. Forehand, Stefano Puntoni, and Luk Warlop (2012), “Identity-Based Consumer Behavior,” International Journal of Research in Marketing, 29 (4), 310–321. Might I Recommend This Session? Insights on the Effectiveness of Modern Persuasion Agents Chairs: Jennifer D’Angelo, Texas Christian University, USA Francesca Valsesia, University of Washington, USA

Paper #1: Slanguage & Cultural Value: It’s Lit to Say It - How the entertainment-value of influencers to be of utmost importance Mainstream Slang Affect Online WOM to consumers. Moreover, they find the effectiveness of influencer- Shaheer Ahmed Rizvi, University of Alberta, Canada communicated (vs. brand-) messages is reduced when the persua- Sarah G Moore, University of Alberta, Canada sive intent of the message is made salient and consumers’ persuasion Paul Richard Messinger, University of Alberta, Canada knowledge is activated. Finally, Hall, Hyodo, and Kristofferson explore the conse- Paper #2: You Should Try It With This: Recommending quences of persuasion agents’ failed persuasion attempts. Across five Compatible Items Increases Perceptions of Expertise and studies, they demonstrate that recommenders reduce future choice Persuasiveness away from their recommended product if their advice is not followed Jennifer D’Angelo, Texas Christian University, USA (unfulfilled recommendation). This switching occurs when persua- Francesca Valsesia, University of Washington, USA sion agents make a recommendation with less-than-full confidence, Paper #3: Influencer Effectiveness: Contexts and Boundaries leading them to use others’ choices to make inferences about the Jared Joseph Watson, New York University, USA product they recommended. Lauren Grewal, Dartmouth College, USA All papers are at an advanced stage of completion with mul- Shoshana Segal, New York University, USA tiple studies completed. Taken together, they shed light on the effec- Paper #4: You Didn’t Take My (Uncertain) Advice? Examining tiveness of persuasion agents and on novel tactics that leverage the the Effects of Confidence and Recommendation Outcomes on changing landscape in communication and technology. We expect Recommender Preferences this session to be of interest to researchers studying persuasion and Matthew James Hall, Oregon State University, USA social influence, as well as those interested in the impact of new tech- Jamie D. Hyodo, University of Nebraska-Lincoln, USA nology on consumer behavior. Kirk Kristofferson, Ivey Business School, Canada Slanguage & Cultural Value: It’s Lit to Say It - How SESSION OVERVIEW Mainstream Slang Affect Online WOM Technological advancements and new societal trends have affected the way in which consumers engage with companies and with EXTENDED ABSTRACT each other, as well as who consumers turn to when seeking infor- In digitally-mediated exchanges, language is the sole basis mation about products and services. These changes have posed new of communication (Danet & Herring, 2007). Accordingly, prior challenges for marketers who need to adapt to novel means of reach- research has studied how consumers are influenced by nuances in ing, engaging, and ultimately persuading consumers—oftentimes word choice in online word-of-mouth (WOM; e.g., Moore 2015). foregoing direct communication in favor of enlisting social media Despite their ubiquity (Zazulak 2016), slang terms have received influencers or leveraging online WOM communication. This special little attention in this context. Yet up to 25% of tweets contain slang session addresses important theoretical and substantive questions (Kundi, Ahmad, Khan, and Asghar 2014), and slang is common on surrounding the effectiveness of modern influencers and persuasion other social media platforms (Facebook, Instagram) and online set- agents. In particular, this session addresses how the effectiveness of tings (Yelp), where consumers engage in WOM by recommending, a persuasion agent is conditional on the language used (paper 1) and evaluating, and discussing products and brands: “This got an #epic the use of expertise cues (paper 2), who the most effective agents of feel to it. #lit #EminemKamikaze”; “These new Jordans are lit af!” persuasion are (paper 3), and what happens if these agents fail to be Slang is an informal type of speech that is ephemeral and dy- persuasive (paper 4). namic (Gleason 1961), and it can infuse all parts of speech with rich Rizvi, Moore, and Messinger study the persuasive effective- connotations: nouns (beast), adjectives (on fleek), verbs (trending), ness of using slang in digitally-mediated WOM exchanges. Across acronyms (yolo), and abbreviations (bae; Ayto and Simpson 2010). four studies, they find the use of mainstream slang (but not subcul- While most extant research examines group dynamics related to tural or unknown slang) improves a persuasion agent’s effectiveness. slang use (Eble 2012), we focus on broadly popular slang terms that In particular, relative to non-slang equivalents (e.g., great), the cul- are not restricted to specific groups. Specifically, we explore slang tural value carried by slang terms (e.g., lit) elicits more favorable terms that emerge, gain mainstream popularity, and are used over the consumer responses in terms of how they evaluate WOM messages internet (Kundi et al. 2014). We propose that using mainstream slang and the products promoted in these messages. improves the effectiveness of online WOM. D’Angelo and Valsesia identify a novel cue persuasion agents Specifically, we argue that slang terms have a unique ability to can use to increase their perceptions of expertise and ultimately their influence consumers through the cultural value imbued in them. We persuasiveness. Across four studies, they find individuals who use introduce the cultural value model to explain how slang affects con- combinatory mentions (talk of how products go well together) are sumers. Building on McCracken’s (1986) model of meaning transfer, perceived as domain experts. For this reason, consumers form a more we theorize that slang terms are specialized words that users have en- positive attitude toward the products endorsed by these individuals. coded with value and meaning. Thus, internet users’ collective (and Watson, Grewal, and Segal investigate why social media influ- tacit) agreement to regard certain informal words as valuable results encers are effective and compare the effectiveness of influencer- vs. in the social construction of culturally valuable verbal artifacts, in the brand-communicated messages. Across three studies, they identify form of slang terms. These slang words are similar to other cultural artifacts which hold special value in the marketplace, due to broad Advances in Consumer Research 1058 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1059 agreement about their status (e.g., fashion, brands; Bikhchandani, thermore, this research speaks to theories of group membership and Hirshleifer, and Welch 1992; Tsai 2005). Because of this cultural slang by demonstrating that mainstream slang functions differently value, we predict that, relative to equivalent non-slang terms, slang from group/subcultural slang. Mainstream slang terms do not rely on terms used in online WOM will elicit more favorable responses from shared group identity; rather, they benefit from their culturally-val- consumers in terms of how they evaluate the WOM message and the ued, popular appeal. Practically, this work suggests that employing product. Furthermore, we expect that these effects will be mediated culturally valued slang can enhance message effectiveness, a finding by perceived cultural value. Below, we report four experiments that applicable to both consumers and firms. test these hypotheses. Study 1 examined the basic effect of slang on consumers. We re- You Should Try It With This: Recommending cruited 158 North American undergraduates to participate in a single Compatible Items Increases Perceptions of Expertise and factor, 2-level (Slang: present vs. control) between-subjects design. Persuasiveness The study introduced participants to a fictional brand of sneakers. They then read an online review which described the sneakers us- EXTENDED ABSTRACT ing either a slang word (fire) or a control word (trendy). Next, par- Decades of research on social influence suggest that some in- ticipants rated the review (was it well written, clearly understood), dividuals can disproportionally sway the preferences and actions of and the product (willingness to buy, product performance).One-way others (Katz and Lazarsfeld 1955; Merton 1968). Generally, individ- ANOVAs revealed significant effects of slang versus control for the uals who are perceived as experts tend to be more influential, as con- review (Mfire = 5.58, Mtrendy = 4.87; p = .02) and the product (Mfire = sumers are more likely to be persuaded by experts’ recommendations

5.72, Mtrendy = 4.99; p = .006), providing preliminary evidence for the (Alba and Hutchinson 1987; Gershoff, Broniarczyk, and West 2001). proposed effect. In recent years, retailers, particularly those based online, have Study 2 (N = 125 undergraduates) replicated and extended made significant investments in hiring individuals (e.g., stylists, study 1 by examining various slang words that we hypothesized designers, nutritionists) to consult consumers in what to purchase would vary in cultural value. We tested five words in descending or- across a breadth of product domains (Bobb 2019; Gorin 2018; How- der of cultural value: a currently popular slang term (lit), an equiva- land 2018). To what extent do consumers perceive these individuals lent non-slang adjective (great), a generic older slang term (cool), an as experts? Perceived expertise can depend on many factors, includ- emerging unknown slang term (blinding), and a made-up slang term ing the individual’s appearance and choices (Bellezza, Gino, and (antipel). The procedure and measures were similar to study 1. The Keinan 2014; Sela et al. 2019). We suggest another signal of exper- results revealed that the popular slang term outperformed the other tise is through the offering of opinions about compatibility among conditions in review (Mlit = 5.04, Mgreat = 4.32, Mcool = 3.97, Mblinding products (combinatory mentions).

= 3.30, Mantipel = 3.04; p < .001) and product evaluations (Mlit = 4.12, Why might combinatory mentions signal expertise? Typically,

Mgreat = 3.80, Mcool = 3.60, Mblinding = 3.92, Mantipel = 3.60; p = .037). complementary products are paired together by the company (Tellis Prior work has established that slang terms often originate in 1986; Karataş and Gürhan‐Canli 2020), alluding to the notion that groups, and are intimately linked with group language (Eble 2012). lay consumers may lack knowledge of which products are compat- We argue that slang can transcend its group status and attain main- ible. Indeed, consumers often experience difficulty assessing the out- stream appeal in the form of cultural value. However, since the come of products that are combined or used together (Petre, Sharp, identity of the review writer was not specified in previous studies, and Johnson 2006). A more nuanced understanding of product at- our effects may have been due to inferences of shared group mem- tributes and how they are related develops as expertise in a domain bership (i.e., similarity) between the reader and the writer (Naylor, grows (Alba and Hutchinson 1987). Hence, if an individual not only Lamberton, and Norton 2011). To address this issue, we manipu- understands the attributes of individual products, but also whether lated sender identity in study 3 (N = 243 undergraduates), using a 2 these attributes are compatible, consumers might view this individu- (Slang: current [lit] vs. unknown [blinding]) x 2 (Sender: in-group al as having expertise in the product domain. In sum, we suggest that vs. out-group) between-subjects design. Participants read a prod- an individual’s mere mention of two or more products as compatible uct review about a sweater and then completed measures similar to (i.e., go well together) can bolster perceptions that the individual is those in previous studies. The results revealed only a positive main an expert within a product domain, subsequently increasing consum- effect of slang (Mlit = 3.78, Mblinding = 3.18; p < .001). The main effect ers’ attitudes toward the products recommended by that individual. of sender identity and the interaction were insignificant, ruling out Four studies test our predictions, address an alternative explanation, group membership as an alternative explanation. and provide external validity for our findings. Study 4 (N = 80 undergraduates) examined the proposed under- Study 1 tests whether combinatory mentions improve attitudes lying process by measuring perceived cultural value. The study com- toward recommended products. Undergraduate participants (N=257) pared popular slang with outdated slang (i.e., previously cherished were randomly assigned to one of two combinatory mention condi- words that are no longer afforded cultural value). It was a single fac- tions (explicit vs. non-explicit). Participants received a subscription tor, 2-level (Slang: popular [dope] vs. outdated [gnarly]) between- box modeled after the real company SpiceBar. A note from Spice- subjects design, and used similar procedures and measures as before. Bar’s chef detailed the contents of the box: one focal “spice of the Additionally, prior to measuring evaluations, we measured cultural month” and three other spices. The chef recommended the same value (2 items, e.g., “Would you consider the choice of words as spices across conditions, but in the high (vs. low) condition, the chef culturally valued [cool]?”). The results showed that popular slang highlighted the compatibility among the spices. Participants then performed better than outdated slang (Mdope = 3.84, Mgnarly = 2.97; p < rated 1) their attitude toward the focal spice and 2) their collective

.001) and that cultural value mediated this effect (CI95%: -0.65, -0.03). attitude toward the other three spices (1=not at all positive; 9=very This research contributes to the literatures on slang and WOM. positive). It extends our understanding of how mainstream slang affects con- The focal spice was rated more positively in the explicit (vs. sumers’ product perceptions via cultural value; this renders slang a non-explicit) condition (Mexplicit=6.47 vs. Mnon-explicit=5.66; F(1, more effective communication tool than common vernacular. Fur- 255)=15.21, p<.001), as were the other three spices (Mexplicit=6.02 1060 / Might I Recommend This Session? Insights on the Effectiveness of Modern PersuasionAgents vs. Mnon-explicit=5.23; F(1, 255)=13.27, p<.001). Participants then en- Influencer Effectiveness: Contexts and Boundaries tered to actually win a subscription to either SpiceBar or another box. More participants in the explicit (vs. non-explicit) condition en- EXTENDED ABSTRACT tered to win the SpiceBar subscription (42% vs. 29%; Wald χ2=4.67, With the proliferation of social media, influencers are more im- p<.04). portant than ever. Brands now spend over one billion dollars annu- Study 2 tests whether consumers’ improved attitudes are driven ally on influencer marketing, where micro-influencers are thought to by increased perceptions of expertise. MTurk participants (N=266) be 6.7 times more cost efficient in generating engagement relative imagined shopping at a furniture store and overhearing an individual to their celebrity counterparts (Pierucci 2018). While brands have (i.e., the recommender) discuss dining chairs with a friend. In the traditionally employed a broadcasting strategy (e.g., national com- combinatory [substitutable] condition, the recommender said the mercials or print ads), they can now employ easily employ a nar- Logan chairs would go really well mixed in with [be a really good rowcasting strategy by selecting specific influencers to reach more alternative to] the Bacci chairs. Participants then rated their attitude homogenous populations. This begs the following questions: “why toward the Bacci chair (1=not at all positive; 9=very positive) and do consumers listen to social media influencers?” and “when should perceived interior design expertise of the recommender (1=not at all, brands outsource their messaging vs. sending the message directly 9=very much). to consumers?” In the combinatory (vs. substitutable) condition, the Bacci While the effectiveness of celebrity endorsers has been widely chair was rated more positively (Mcombinatory=6.90 vs. Msubstitutable=6.42; researched (for a review, see: Erdogan 1999), this literature can’t F(1, 264)=5.36, p<.03) and the recommender was perceived as fully explain the effectiveness of social media influencers today. In having greater expertise (Mcombinatory=7.06 vs. Msubstitutable=6.68; F(1, addition to credibility and liking (Friedman and Friedman 1979), we 264)=3.58, p=.06). Mediation analysis revealed the effect of the rec- posit that the desire for a parasocial relationship (Rubin and McHugh ommenders’ mention on attitudes was driven by perceived expertise 1987) plays a role with consumers today. Whereas influencers may

(bindirect=.18, SE=.10, CI95%: .0058, .3947). have traditionally existed through standard advertisements, consum- An alternative explanation for our findings may be that men- ers now seek out and “follow” influencers, whose marketing may tioning products are compatible implies versatility (i.e., the product seem more like an authentic message between friends than an ad- could have many uses). Study 3 addresses this alternative explana- vertisement. Thus, a consumer’s persuasion knowledge (Friestad tion. If consumers’ attitudes are driven by versatility, a recommend- and Wright 1994) might not be activated in the same fashion that it er who established expertise by providing a combinatory mention is with traditional advertisements. Promoting too many brands can should not improve consumers’ attitudes toward other subsequent- decrease an influencer’s effectiveness (Mowen and Brown 1981; ly recommended products where compatibility is not mentioned. Tripp, Jensen, and Carlson 1994), but there are likely to be many MTurk participants (N=500) read three Instagram posts from a food other ways by which influencer effectiveness is moderated. influencer. In each post, the influencer in the combinatory [control] In this research, our goal is first to understand the attributes that combinatory mention condition recommended a food item and men- characterize effective influencers. Then, we aim to determine the tioned [but did not mention] a compatible food item. Participants boundaries of said influencer effectiveness. In doing so, we aim to rated the influencer’s food expertise (1=not at all, 9=very much). contribute to both theory and practice regarding this emerging mar- Next, all participants viewed the influencer’s fourth post which rec- keting trend. In study 1, we survey participants to identify the mani- ommended a food product with no mention of a combatable item. fest attributes of effective influencers, in general. Study 2 then asks Participants then rated their attitude toward this product (1=not at all participants to identify an influencer they follow and evaluate the in- positive; 9=very positive). fluencer on the same attributes. Finally, in study 3, we investigate the In the combinatory (vs. control) condition, the influencer was relative effectiveness of a marketing message when it comes from perceived as more expert (Mcombinatory=7.26 vs. Mcontrol=6.01; F(1, the brand vs. an influencer to determine the relative value of employ- 498)=65.15, p<.001); the food product was also rated more posi- ing an influencer. Due to space constraints, we will only highlight tively (Mcombinatory=6.37 vs. Mcontrol=5.96; F(1, 498)=7.42, p<.007). select findings. Importantly, the effect of combinatory mention on attitudes toward Study 1 asked participants “what characteristics of influenc- the food product was driven by perceived expertise of the influencer ers are important to you” then featured a battery of 14 items on a

(bindirect=.73, SE=.10, CI95% [.5399, .9383]). 5-point scale. A factor analysis yielded a 3-factor solution with the Study 4 provides external validity for our findings using a measures of message value (8 items; α = .907; e.g., “inspiring” or Facebook split-ad field study. We ran advertisements for a skincare “helpful”), entertainment value (2 items; r = .622; i.e., “entertaining” website where website users provide skincare advice. In the explicit and “funny”), and physical value (4 items; α = .741; e.g., “attractive” combinatory mention ad, the ad quoted and depicted a skincare ad- or “similar”). Relative to physical value (M = 2.45), paired t-tests vice giver’s recommendation for which products go well together in indicated that consumers placed more importance on message value a skincare routine. In the non-explicit condition, the ad quoted and (M = 3.49; t(106) = 11.453; p < .001) and entertainment value (M = depicted the same product recommendations from a skincare advice 3.59; t(106) = 10.602; p < .001). There was no difference in the per- giver without explicitly stating that the products went well together. ceived importance of message value and entertainment value (t(106) A logistic regression predicting click-through rate revealed that the = -.944; p = .347). explicit combinatory ad lead to significantly more clicks (1.52%) Study 2 followed the same paradigm as study 1 with a few than the non-explicit ad (CTR=1.26%; Wald χ2=9.92, p<.002). exceptions. We used fewer items (six) and elicited responses on a Among those who clicked on the ad to visit the skincare website, 7-point scale. And importantly, in this study, we first asked partici- membership sign-up rates were greater among those who arrived via pants to indicate an influencer that they currently follow on Insta- the explicit combinatory ad (7.12%) relative to the non-explicit ad gram, and thus, asked them to what extent each attribute described (4.51%; Wald χ2=3.24, p=.07). their influencer rather than the importance of that attribute. In doing so, we wanted to assess whether the responses in study 1 were reflective of perceptions of influencers that consumers already follow. Advances in Consumer Research (Volume 48) / 1061

Paired samples t-test indicated that this was not the case. Relative to when they were unsure whether the other consumer would like their physical value (M = 5.04), paired t-tests indicated that consumers recommended product. placed less importance on message value (M = 4.12; t(79) = -3.386; Given recommendations are being provided with less-than-full p = .001), while there was no difference with entertainment value (M confidence, low-confidence recommenders may be prone to prefer- = 5.25; t(79) = .946; p = .347). Naturally, entertainment value was ence- or choice-influence based on recommendation feedback. The also more descriptive of their recalled influencer relative to message rise of social media as a recommendation platform, along with in- value (t(79) = -5.163; p < .001). Taken together, these studies sug- creased sharing of consumption choices, increases the likelihood gest that entertainment value seems to be of utmost importance while recommenders will discover whether their recommendations are consumers might oversell the importance of message value and unfulfilled/unfulfilled (Duan and Dholokia 2017). We propose those dersell the importance of physical value. recommending with less-than-full confidence, despite publicly com- Having established why consumers might follow influencers. mitting to the recommended product, will shift future product choice We now seek to understand the context-dependent value of influenc- away from their recommended option when their recommendation ers relative to brands. Study 3 adopts a 2x2 design in which frame is not fulfilled. We propose this will occur because low-confidence the social media post as coming from the brand or an influencer, and consumers perceive others’ divergent choices as indicative of differ- whether a promotion code is absent or present. Our theorizing would ing objective quality (Spiller and Belogolova 2013), thus inferring suggest that consumers generally prefer marketing messages from the unpurchased (recommended) product is of lower quality. Thus, influencers vs. brands due to the activated persuasion knowledge in we predict the interaction between recommendation fulfillment and the presence of a brand’s marketing attempt, but the presence of a confidence will be mediated by perceived product quality. promotion code leads consumers to identify the influencer’s mes- In Study 1a, undergraduates (n=91) tested three pens and se- sage as a marketing message, activating persuasion knowledge, and lected their favorite. Next, they individually interacted with a trained lowering their effectiveness. After viewing the post, we measured confederate who had information about the pens but did not test Blue Apron purchase intentions (3 items; α = .867). A 2x2 ANOVA them. During the scripted interaction, the confederate solicited a rec- on purchase intentions yielded a significant interaction (F(1,403) = ommendation on which pen to select, and the participant provided 10.757; p = .039). Neither main effect was significant (p’s > .10). his/her recommendation. After the recommendation, the confederate Planned contrasts demonstrate that in the absence of a promotion, indicated she planned to select the recommended pen (fulfilled) or an whether the brand or an influencer posts the content does not sig- alternative (unfulfilled). Participants returned to the lab and selected nificant impact intentions (Mbrand = 4.09; Minfluencer = 4.30; F(1,403) = a pen to keep. Participants were more likely to shift preference when .869; p > .35). However, the presence of a promotion led to a signifi- their recommendation was unfulfilled (36.5%), relative to fulfilled 2 cant decline for influencers (Mbrand = 4.32; Minfluencer = 3.88; F(1,403) (5.9%;χ Wald=9.88, p=.002). = 4.025; p = .045), suggesting the promotion made salient an ulterior In Study 1b, participants (n=301) read a scenario about recom- motive and undermined the influencer’s credibility. mending a bottle of wine to a wine-savvy friend, who either did or did Taken together, these studies begin to address the questions of not fulfill the recommendation. We also included a control condition “why do we follow influencers” and “what are the boundaries of in which participants did not learn the recommendation outcome. their influence?” In subsequent studies, we plan to investigate the Again, those whose recommendations were unfulfilled reported low- discrepancies between studies 1 and 2 by further exploring the role er intentions to repurchase the recommended wine (M=4.93; 7-point of an existing parasocial relationship with an influencer and creating scale), relative to those in the fulfilled (M=5.57; F(1,286)=13.47, a taxonomy of message characteristics (e.g., promo codes, behind- p<.001) and control (M=5.46;F(1,286)=9.26, p=.003) conditions. the-scenes content, etc.) that are better suited to come from the brand In Study 2a, participants (n=326) completed a 2(fulfilled/un- or the influencer. fulfilled recommendation) x 2(high/low confidence) study in which they read about recommending a bottle of wine to a friend. Con- You Didn’t Take My (Uncertain) Advice? Examining the fidence was manipulated via recommendee wine knowledge (Con- Effects of Confidence and Recommendation Outcomes on stant et al. 1994). A 2x2 ANOVA with product intentions as the de- Recommender Preferences pendent variable revealed the predicted interaction (F(1,322)=3.94, p=.048). Unfulfilled recommendations negatively affected future

EXTENDED ABSTRACT purchase intentions when confidence was low (MFulfilled=5.78, MUnfulfill

Although limited, research considering how providing word-of- ed=5.08;F(1,322)=8.23, p=.004), but not high (MFulfilled=5.40, MUnfulfille mouth influences thesharer often assumes that, i) those who recom- d=5.39;F(1,322)=.005, p=.942). Perceived product quality mediated mend do so with initially-high confidence, ii) others will (and do) the interaction (CI95:[.089,.680]). take the recommender’s advice, and iii) recommending entrenches Study 2b tested the moderating role of the presence of third product-related attitudes. However, by incorporating the bidirec- parties’ divergent/convergent recommendations. We propose rec- tional nature of the recommendation environment, we challenge ommendation confidence will be enhanced when supporting rec- these assumptions and demonstrate that consumers who recommend ommendations from others are present, thus attenuating our effect. a product can and do reduce future preference for that product when Undergraduates (n=127) were assigned to one of three conditions their recommendation is not followed (unfulfilled recommendation). (recommendation context: alone/converging/diverging) in which Research shows that confidence indeed predicts WOM engage- they recommended a restaurant to a friend on Facebook (stimuli ment (Brown 2005). However, because recommendations are fre- presented as mock Facebook Recommendation posts). In the alone quently solicited and provided in social settings, consumers often condition, the participant provided the only recommendation. In recommend with less-than-full confidence to fulfill perceived social the diverging condition, another friend also responded, but recom- obligations (Lui and Gal 2011) or avoid appearing unknowledgeable mended a different restaurant. In the supporting condition, one friend (Berger 2014). Confirming this assumption, a pilot study revealed recommended the same restaurant as the participant, while two oth- that 77.9% of participants (n=103) had made a recommendation ers recommended an alternative. All participant recommendations were unfulfilled. An ANOVA revealed a significant effect of recom- 1062 / Might I Recommend This Session? Insights on the Effectiveness of Modern PersuasionAgents mendation context on product intentions (F(1,124)=3.71, p=.027). Brown, Tom J., Thomas E. Barry, Peter A. Dacin, and Richard F. Consistent with our framework, unfulfilled recommendations re- Gunst (2005), Spreading the word: investigating antecedents duced future support of participants’ recommended product in the of consumers’ positive word-of-mouth intentions and lone/divergent conditions (M=5.23) compared to when a convergent behaviors in a retailing context, Journal of the Academy of recommendation was present (M=5.65;F(1,124)=6.23, p=.014). Marketing Science, 33(2), 123-138. No differences emerged between the lone and divergent conditions Buckless, Frank A., and Sue Pickard Ravenscroft (1990), Contrast

(MLone=5.13, MDivergent=5.33, p=.275). Importantly, the focal contrast coding: A refinement of ANOVA in behavioral analysis, was again mediated by perceived product quality (CI95=[.049,.446]). Accounting Review, 933-945. Others’ choices are seen as more diagnostic when product qual- Constant, David, Sara Kiesler, and Lee Sproull (1994), What’s mine ity is perceived to be objective (vertical differentiation; Spiller and is ours, or is it? a study of attitudes about information sharing, Belogolova 2016). Alternatively, when products are differentiated on Information Systems Research, 5(4), 400-421. more taste-based dimensions (horizontal differentiation), divergent Danet, Brenda, and Susan C. Herring (2007), The multilingual choices can be attributed to idiosyncratic preferences. Thus, we pre- Internet: Language, culture, and communication online, dict our focal effect will be attenuated when the product is perceived Oxford University Press on Demand. as horizontally (vs. vertically) differentiated. Study 3 employed the Duan, Jingye, and Dora R. Dholokia (2017), Posting purchases same design as Study 1a (confederate interaction) with one differ- on social media increases happiness: The mediating roles ence—manipulation of differentiation prior to participants’ initial of purchases’ impact on self and interpersonal relationships, choice. The vertical condition stated the pens were made with ma- Journal of Consumer Marketing, 34(5), 404-413. terials of varying quality, resulting in some pens being objectively Eble, Connie (2012), Slang and sociability: In-group language better than others—participants were asked to select the best pen. among college students. Chapel Hill, NC: University of North The horizontal condition stated the pens were made with materials Carolina Press. of similar quality, meaning observed differences resulted from dif- Erdogan, B. Z. (1999), Celebrity endorsement: A literature review, fering preferences—they were asked to select their preferred pen Journal of Marketing Management, 15(4), 291-314. (Spiller and Belogolova 2013). Undergraduates (n=163) tested the Friedman, H. H., and Friedman, L. (1979), Endorser effectiveness pens and recommended their choice to a confederate, who then did/ by product type, Journal of Advertising Research, 19(5), did not fulfill the recommendation. We planned to conduct a logistic 63–71. regression, but low switching rates in three conditions rendered this Friestad, M., and Wright, P. (1994), The persuasion knowledge inappropriate (Cox 1970). As such, to directly test our predictions, model: How people cope with persuasion attempts, Journal of we used simple contrasts to compare mean switching rates in the Consumer Research, 21(1), 1-31. vertical/unfulfilled condition versus the three other cells (Buckless Gleason Jr, Henry A. (1961), An Introduction to Descriptive and Ravenscroft 1990). As expected, those in the unfulfilled/vertical Linguistics (Revised Edition). condition changed their choice away from their recommended op- Gorin, Amy (2018), The Rise of the Supermarket Nutritionist, Food tion more often (10 of 40; M=25.00%) than participants in the other Network. Retrieved February 15, 2020 from https://www. three conditions (6 of 123; M=4.88%;F(1,161)=14.89, p<.001). All foodnetwork.com/healthyeats/news/2018/supermarket-rds. simple effect comparisons were also significant. Howland, Daphne (2018), Wayfair draws up interior design By integrating the bidirectional nature of solicited recommen- services, Retail Dive. Retrieved February 15, 2020 from dations and focusing on the consumption consequences of WOM https://www.retaildive.com/news/wayfair-draws-up-interior- sharers, this research challenges prior assumptions of WOM research design-services/530237/. to demonstrate possible deleterious effects of recommendation out- Karataş, Mustafa, and Zeynep Gürhan‐Canli (2020), When comes on subsequent recommender choice. We further demonstrate Consumers Prefer Bundles with Noncomplementary Items the crucial role that confidence and product differentiation play in to Bundles with Complementary Items: The Role of Mindset this understudied aspect of WOM behavior. Abstraction, Journal of Consumer Psychology, 30(1), 24-39. Katz, Elihu and Paul F. Lazarsfeld (1955), Personal influence. REFERENCES Glencoe, Ill: The FreePress Alba, J. W., & Hutchinson, J. W. (1987), Dimensions of consumer Kundi, Fazal Masud, Shakeel Ahmad, Aurangzeb Khan, and expertise, Journal of Consumer Research, 13(4), 411-454. Muhammad Zubair Asghar (2014), Detection and scoring of Ayto, John, and John Simpson (2010), Oxford dictionary of modern internet slangs for sentiment analysis using SentiWordNet, Life slang. Oxford University Press. Science Journal, 11(9), 66-72. Bellezza, S., Gino, F., & Keinan, A. (2014), The red sneakers Liu, Wendy, and David Gal (2011), Bringing Us Together or effect: Inferring status and competence from signals of Driving Us Apart: The Effect of Soliciting Consumer Input nonconformity, Journal of Consumer Research, 41(1), 35-54. on Consumers’ Propensity to Transact with an Organization, Berger, Jonah (2014), Word of mouth and interpersonal Journal of Consumer Research, 38(2), 242-259. communication: A review and directions for future research, McCracken, Grant (1986), Culture and consumption: A theoretical Journal of Consumer Psychology, 24(4), 586-607. account of the structure and movement of the cultural meaning Bikhchandani, Sushil, David Hirshleifer, and Ivo Welch (1992), of consumer goods, Journal of Consumer Research, 13(1), A theory of fads, fashion, custom, and cultural change as 71-84. informational cascades, Journal of Political Economy, 100(5), Merton Robert K. (1968), Patterns of influence: Local and 992-1026. cosmopolitan influentials, in social theory and social structure. Bobb, Brooke (2019), Amazon fashion launches a personal styling ed. Robert K. Merton, New York: Free Press, 441–74 service⁠—will customers actually use it?, Vogue, Retrieved Moore, Sarah G. (2015), Attitude predictability and helpfulness in January 22, 2020 from https://www.vogue.com/article/ online reviews: the role of explained actions and reactions, amazon-prime-personal-shopper. Journal of Consumer Research, 42(1), 30-44. Advances in Consumer Research (Volume 48) / 1063

Mowen, J. C., and Brown, S. W. (1981), on explaining and Spiller, Stephen A., and Lena Belogolova (2016), On consumer predicting the effectiveness of celebrity endorsers,ACR North beliefs about quality and taste, Journal of Consumer Research, American Advances. 43(6), 970-991. Naylor, Rebecca Walker, Cait Poynor Lamberton, and David Tellis, Gerard J. (1986), Beyond the many faces of price: an A. Norton (2011), Seeing ourselves in others: Reviewer integration of pricing strategies, Journal of Marketing, 50(4), ambiguity, egocentric anchoring, and persuasion, Journal of 146-160. Marketing Research, 48(3), 617-631. Tsai, Shu-pei (2005), Utility, cultural symbolism and emotion: a Petre, M., Sharp, H., & Johnson, J. (2006), Complexity through comprehensive model of brand purchase value, International combination: an account of knitwear design, Design Studies, Journal of Research in Marketing, 22(3), 277-291. York, NY: 27(2), 183-222. Holt, Rinehart and Winston. Pierucci, S. (2018). Why Micro-Influencer Marketing is Still ‘The Tripp, C., Jensen, T. D., & Carlson, L. (1994), The effects of Game’ in 2019, Medium, Retrieved February 21, 2020 from multiple product endorsements by celebrities on consumers’ https://medium.com/swlh/why-micro-influencer-marketing-is- attitudes and intentions, Journal of Consumer Research, 20(4), the-game-in-2018-fdeda0993c36. 535-547. Rubin, R. B., and McHugh, M. P. (1987). Development of Zazulak, Steffanie (2016), The rise in popularity of internet slang, parasocial interaction relationships. Retrieved August 18 2018 from https://www.english.com/blog/ english-internet-slang/. New Learnings in Financial Decision-Making: Influences on How Consumers Use and Obtain Financial Resources Chair: Wendy J. De La Rosa, Stanford University, USA

Paper #1: The Impact of Payment Frequency on Subjective present bias. Finally, in the fourth paper, Park, Cryder, and Gershon Wealth Perceptions and Discretionary Spending demonstrate the impact of relationship closeness on how consum- Wendy J. De La Rosa, Stanford University, USA ers decide to request debt repayment. Across a series of studies, the Stephanie M. Tully, Stanford University, USA authors show that consumers’ are more likely to request the payment in person (vs. using a bank enabled app like Venmo) when they per- Paper #2: Pennies Reframing of Savings Rates ceive the borrower to be a close friend. Stephen Shu, City University of London and Digital Nudging Together, these papers enhance our understanding of how con- Tech, UK sumers decide to use and obtain financial resources. This topical and Hal Hershfield, University of California Los Angeles, USA novel session sheds light on how unique interventions can help im- Shlomo Benartzi, University of California Los Angeles, USA pact consumers’ financial well-being. Richard Mason, Carnegie Mellon University and Voya Financial, USA The Impact of Payment Frequency on Subjective Wealth Paper #3: How Soon is Now? Present Bias and the Mental Perceptions and Discretionary Spending Accounting of Time Oleg Urminsky, University of Chicago, USA EXTENDED ABSTRACT Minkwang Jang, University of Chicago, USA An increasingly popular trend is for consumers to get paid more Paper #4: Fighting Fiscal Awkwardness: The (Dis)Advantages often, resulting in more frequent, yet smaller paychecks. For exam- of Digital Payment Methods on Peer-Debt Dynamics ple, beyond day laborers and tipped employees, Uber drivers, and Alexander Park, Washington University in St. Louis, Olin even Wal-Mart cashiers can now get paid on a daily basis. However, Business School, USA surprisingly little is known about whether and how payment frequen- Cynthia Cryder, Washington University in St. Louis, Olin cy impacts consumer behavior. Higher payment frequency has been Business School, USA shown to help consumers spread out their consumption throughout a Rachel Gershon, University of California San Diego, Rady period (e.g., Parsons and Van Wesep 2013; Stephens and Unayama School of Management, USA 2011). However, in the current work, we explore other potential consequences of payment frequency: consumers’ wealth perceptions and SESSION OVERVIEW their discretionary spending. Millions of consumers are struggling financially. According to If consumers focused only on the smaller number on their pay- the Consumer Financial Protection Bureau, 43% of Americans are check, it could lead consumers to have lower wealth perceptions unable to make ends meet (2017), and nearly 40% cannot cover an (Morewedge, Holtzman, and Epley 2007). However, we suggest that unexpected $400 expense (Federal Reserve Board, 2018). Thus, it is payment frequency is more likely to be experienced as a series of of great theoretical and societal importance to understand how con- segregated gains. Past research demonstrates that smaller, more fre- sumers make financial decisions. In this session, we take a holistic quent segregated gains lead to higher hedonic assessments than a approach and focus on understanding unique and underexplored fac- larger, less frequent, aggregated gain because hedonic assessments tors that impact how consumers use the resources they already have exhibit diminishing marginal returns (Thaler, 1985). We build on this (e.g., spending and saving) and factors that impact how consumers work and posit that subjective wealth perceptions, like but distinct acquire more financial resources (e.g., intertemporal choices and from hedonic assessments, also exhibit diminishing marginal returns. debt collection). Thus, more frequent (albeit smaller) paychecks should increase con- The first two papers focus on factors that impact how consumers sumers’ wealth perceptions. In short, consumers’ wealth perceptions use financial resources that they already have: payment frequency are formed not just from consumers’ objective wealth (or their objec- and saving frames. In the first paper, De La Rosa and Tully examine tive wealth relative to a benchmark), but also from the timing of their the impact of payment frequency on consumers’ spending decisions. payments. Moreover, we suggest that these increased wealth percep- The authors demonstrate that higher payment frequency increases tions also increase consumers’ willingness to spend on discretionary consumers’ perceptions of wealth, which in turn increases their dis- purchases. cretionary spending. In the second paper, Shu et al. demonstrate how In Study 1, we analyze the real-world spending behaviors of differing savings frames (save a penny of every dollar of salary vs. 27,000 consumers who were customers of a large US banking plat- save 1% of your salary) impact consumers’ perceptions of the afford- form. This data included consumers’ income and spending for each ability of saving. Through a series of experiments, the authors show day. Using a linear mixed-effects model, our analysis revealed a posi- that pennies framing increases savings rates by making saving seem tive and significant relationship between payment frequency and dis- more affordable. cretionary spending. Thus, higher payment frequency was associated The last two papers focus on factors that impact how consumers with more discretionary expenditures (β = 1.35, p < .001), even after obtain financial resources: budget periods and social ties. In the third controlling for consumers’ income (log-transformed). paper, Urminsky and Jang demonstrate that consumers’ perceptions In Study 2 (N = 405), we created a life simulation where par- of their budgeting periods significantly impacts their intertemporal ticipants earned income, incurred expenses (e.g., rent, utilities, phone decisions. Consumers choose the smaller, sooner option less often bill), and made discretionary spending decisions (e.g., whether to eat (e.g., become more patient) when the larger, later option is still without, whether to go to a concert, etc.) to examine the causal impact in (vs. outside of) their perceived budgeting period, which explains of payment frequency. We manipulated participants’ payment frequency while holding constant the total amount earned. Participants Advances in Consumer Research 1064 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1065 paid daily (vs. weekly) made more high-cost expenditures (Mdai- three survey-based, experimental studies which focuses on savings ly=9.47 vs. Mweekly= 8.48, p<.001) than those in the weekly condi- decisions which include the joint choice of an initial savings rate and tion. Moreover, those in the daily pay condition reported had higher an escalator rate. For example, if a person elects an initial, economi- wealth perceptions than those in the weekly pay condition (Mdai- cally equivalent savings rate of 3% and an escalator rate of 2%, then ly=2.86 vs. Mweekly= 2.24, p<.001). These wealth perceptions me- they will save 3% of their salary this year, 5% of their salary next diated changes in spending decisions (ab=0.27, 95% CI[0.12, 0.46]). year, 7% of their salary two years from now, and so forth. Study 3 was designed to disentangle differences in objective Study 1 (N=200) covers a low, introductory savings choice and wealth from subjective wealth perceptions. In Study 2, daily pay led may effectively be viewed as a practical boundary case where com- to a higher average daily balance than the weekly condition. Hence, panies would likely not implement initial savings rates and escalator in Study 3, we designed the life simulation such that participants rates any lower than 1% and 1%. We assess the difference in par- in the bi-weekly pay condition had a higher average daily balance ticipants’ selections in a between-subject design when faced with 1) than those in the daily pay condition (while maintaining the total an opportunity to either save 1% of salary, increasing by 1% every amount of income earned the same across conditions). Again, higher year thereafter or not save versus 2) an opportunity to either save 1 payment frequency led to more high-cost decisions than weekly pay penny per dollar of salary increasing by 1 penny per dollar of sal- (Mdaily=8.01 vs. Mbi-weekly=6.94, p<.001; Moreover, this effect ary every year thereafter or not save. We find evidence that setting was again mediated by participants’ wealth perceptions (ab=.12, rates this low, even for those among the lowest salaries, may be too 95% CI[0.01, 0.27]). low as evidenced by ceiling effects in participation rates (β=0.02, We next examined the introduction of a daily pay option by a p=0.709) and insignificant main effects in perceived affordability sizable gig-economy platform. We received earnings and spending (β=0.15, p=0.427), although a spotlight analysis indicates margin- data for 752 people working for this platform three months before ally significant simple effects of subjective numeracy on perceived and three months after the introduction of daily pay through a third- affordability. party app, accounting for over 427,000 transactions. Using a linear Study 2 (N=200) broadens the scope of inquiry by allowing par- mixed-effects model, we find that after the introduction of a daily ticipants to select higher savings rates within a range that effectively pay option, workers made more discretionary expenditures (β=4.02, spans the range of default rates of a large majority of defined con- p=.035) as compared to before the introduction of daily pay. tribution plans. Participants first make a choice to save at an initial We demonstrate that consumers’ form their wealth perceptions savings rate (limited to between 0% to 6% of salary) followed by an based not just on their objective wealth (or their objective wealth escalator rate (limited to between 0% to 3%) in a between-subject relative to some benchmark) but also from the timing of their pay. design with some seeing pennies versus percent framing. In this Across several pre-registered, controlled lab studies, and two analy- study, no differences were detected in the initial savings rate selected ses of real consumer spending behavior, the current research dem- (β=0.11, p=0.648), although participants selected significantly higher onstrated that higher payment frequency consumers’ wealth percep- escalator rates in the pennies condition (β=0.55, p=0.000). Note that tions and thus, their discretionary spending. Our research suggests a 55-basis point increase, if it were to be realized in the field, would that payment frequency impacts not just when consumers spend, but be quite substantial given the benefits of money compounding and also how much, and calls for future research to understand how other that the most common escalator rate in plans with auto-escalation is aspects of consumer spending that may be impacted by payment fre- 1%. Spotlight analysis indicates that subjective numeracy moderates quency. both choice of higher escalator rates and perceptions of affordability. Additionally, both structural equation modeling and Preacher-Hayes Pennies Reframing of Savings Rates mediation analyses indicate perceptions of affordability partially mediating initial savings rates and escalator rate choices. EXTENDED ABSTRACT Study 3 (N=401) broadens the scope of inquiry to include both The savings crisis in the United States affects a broad swath self-reported and objective psychological measures, the latter using of workers, and this disproportionately affects those with lower in- methodologies from Query Theory (Johnson et al., 2007). Further- comes, who may also have lower education levels, numeracy, and more, a limitation of Study 2 was also relaxed in that instead of re- financial literacy. For example, one of the most pervasive ways of quiring people to choose an initial savings rate and then an escalator choosing a savings rate is to identify a percent of salary that one rate (i.e., restricted process flow), participants could freely choose wants to contribute to their retirement account. However, the broader to select and change either rate in any order. In Study 3, we find literature indicates that how numerical and financial information is that people both save at both higher initial savings rates (β=0.50, presented affects people’s judgments and decisions. Examples in- p=0.004) and escalator rates (β=0.37, p<.001) in the pennies fram- clude the miles per gallon illusion versus gallons per mile (Larrick ing. We also find that people subjectively report higher affordability & Soll, 2008) where people underestimate the financial benefits of (β=0.41, p<.001) and understandability (β=0.21, p=0.043) with pen- replacing certain cars, and the temporal reframing of recurring sav- nies framing. Spotlight analyses indicate that financial literacy mod- ings decisions (Hershfield et al., 2020 forthcoming) where more erates savings choices, especially for those with lower financial liter- people participate in recurring savings programs when decisions are acy. Financial literacy also moderates self-reported affordability and framed as saving $5 per day versus $150 per month, despite being understandability. Structural equation modeling also shows evidence essentially the same choice from an economic perspective. In other for people choosing higher initial savings rates and higher escalators words, metrics matter. In our stream of research on pennies refram- with objective thoughts about affordability mediating choice. ing of savings rates, we hypothesize that eliciting savings rates using The studies above offer evidence that a small change of elicit- percentages may lead to less desirable outcomes for some subpopu- ing savings choices using pennies versus percent framing can impact lations. In a laboratory setting, we explore a novel, savings choice people’s perceptions of affordability and thus lead people to select architecture that is essentially not available in the market today by higher savings and savings escalator rates. Treatment effects seem framing savings decisions using pennies per dollar of salary, which to improve the outcomes for those with the lowest numeracy and avoids uses of percentages. This research stream currently includes financial literacy the most. In Study 3, those with the lowest financial 1066 / New Learnings in Financial Decision-Making: Influences on How Consumers Use and Obtain Financial Resources literacy selected initial rates that were more than 2 percentage points when thinking about budgeting. This measure allowed us to estimate higher in the pennies frame (relative to a sample mean of 2.9%), when the “current” budget period ends and when the “future” pe- and they also selected escalator rates that were more than 64 basis riod begins for each participant, relative to the date of the survey. points higher (relative to a sample mean of 0.82%). Research on the First, we observed substantial heterogeneity in the duration of the pennies reframing of savings rates is continuing with a survey-based current budget period, with the modal response (two weeks from to- study (N=270) covering a higher income participant pool within a day) taking up less than 20% of total responses, consistent with prior corporation and a field study (N = target of 1,000) covering the real research suggesting heterogeneity in people’s planning horizons for retirement choices of eligible employees across a number of retire- finances (e.g., Lynch, Netemeyer, Spiller, & Zammit, 2009). ment plans (study launched in the field in November 2019 with data Importantly, our analysis on intertemporal choices showed that expected in June 2020). people were significantly less likely to choose the larger-later option when the options occurred in different subjective periods, i.e., the How Soon is Now? Present Bias and the Mental smaller-sooner option in their current budget period and the larger- Accounting of Time later option in the future period (β=-0.235, p<.001). Whether the sooner option is promised today did not explain present-bias and EXTENDED ABSTRACT was a weaker predictor of choice (β=0.02, p=.07), which is further Consumers often face situations where they have to choose be- confirmed by model comparison analyses. While participants with tween a smaller but immediate outcome and a larger but delayed more savings had longer current periods, our results did not change outcome. Research on intertemporal choice has found that people after controlling for demographics, including the reported amount of are often more impatient for temporally proximate decisions (e.g., savings. In an additional study, we replicated the same finding using $40 today vs. $50 in 1 month) than for future decisions (e.g., $40 longer delays (N=289). in 12 months vs. $50 in 13 months), contradicting the standard eco- Next, we further replicated our findings in more specific bud- nomic model that dictates constant degree of patience over time (for geting and consumption contexts. In study 3 (N =234), we asked a review, see Urminsky & Zauberman, 2016). Present bias explains participants to make a series of intertemporal choices with outcomes this inconsistency by suggesting that rewards lose more value when earmarked to specific consumer budget categories (i.e., credits for delayed from “now” or the present (Loewenstein & Prelec, 1992). grocery shopping or credits for utility payment). We then again However, prior literature has not clearly established what counts as asked participants about their budget period specifically for the cor- the present. responding spending categories. On average, those who answered Some research has shown that people often perceive time as about utility payments reported having longer current budget periods discrete, which can impact their task initiation (Hennecke & Con- than those who answered about grocery shopping (p<.001). Repli- verse, 2017; Tu & Soman, 2014). Relating to financial decisions, cating the prior results again, we found that people were more impa- how people divide time differently was shown to affect their estimate tient when the smaller-sooner and larger-later credits were promised of future budget (Ülkümen, Thomas, & Morwitz, 2008) and their in different budgeting periods for the given spending category (β=- mental tracking of costs (Soster, Monga, & Bearden, 2010). 0.07, p<.001), and relatively insensitive to whether one of the op- In the current research, we build on the notion of discrete per- tions can be received “today” or not (β=0.009, p>.05). We found no ception of time as budget periods to fill in the gap in the literature significant differences in these effects between the contexts. on intertemporal choice regarding present bias. In seven studies, we Finally, in study 4 (N=532), we tested for a causal effect by demonstrate that people’s choices are not consistent with treating providing participants with hypothetical budget periods—either two now as the current day, and that how people subjectively categorize weeks or six weeks. For choices in which the smaller-sooner option time into binary periods, as either current or future financial period, was within two weeks and the larger-later option was after two weeks can provide a better explanation of preference reversals in intertem- (but before six), participants given the two-week current period were poral choice. more impatient on average (p<.001), but vice versa when the options In studies 1A and 1B (total N=917), we first tested when pres- spanned across current and future periods only in the six-week cur- ent-biased preference reversals emerge. One possible heuristic is to rent period condition (p=.002). This did not hold for other choices in define the current day to be the present and simply any time later to which period-spanning did not differ by condition p( s>.05). be the future. Under such an assumption, present bias would predict Our results suggest that budget periods can explain present-bi- people to demonstrate relatively drastic preference shift toward the ased preference: people are more willing to accept delays that occur larger-later option between a choice “today” (e.g., $30 today or $50 in the same period but are reluctant to delay rewards from the current in 1 month) and the same choice any time in the future (e.g., $30 in to future financial period. 1 month or $50 in 2 months), but less so between choices already in the future. Contrary to this prediction, delaying choice by either two Fighting Fiscal Awkwardness: The (Dis)Advantages of weeks or one month by adding a relatively short “front-end delay” Digital Payment Methods on Peer-Debt Dynamics (i.e., how much the same choice is delayed into the future), did not result in a significant increase in choices of the larger-later option EXTENDED ABSTRACT (ps>.05). Statistically significant differences only emerged when the People often lend and borrow money from peers. Although such front-end delay was three months or longer (ps<.01), which sug- financial transactions may appear superficially mundane, debt repay- gests people are not simply “today-biased.” In an additional study ment among acquainted individuals is socially complex. The discus- (N=615), we found that the hyperbolic discounting model also can- sion of money among peers can be uncomfortable (Krueger, 1986; not fully account for how preferences change with front-end delays. Trachtman, 1999; Rose & Orr, 2007), and accordingly, requesting In study 2 (N=262), we tested if how people think of budget and repaying money can evoke awkwardness. Because of this dis- periods can better explain present bias. After making a series of in- comfort, consumers may turn to digital payment methods such as tertemporal choices, participants were asked to categorize different Venmo to avoid fiscal confrontations. However, the use of digital future times as being part of the current or future financial period payment methods for peer-debt requests may also increase percep- Advances in Consumer Research (Volume 48) / 1067 tions of impersonality (Walther, 1995; Schroeder et al., 2017). Thus, Study 4 (N = 503) addressed the mechanisms underlying the the use of digital payment methods may act as a double-edged sword relationship between social distance and choice of request method. for peer debt requests, where it can potentially reduce fiscal confron- We investigated whether anticipated discomfort and impersonality tation, but also make the interaction feel less personal. between social distances are associated with Venmo versus in-per- In four studies, we investigate how consumers adjust the way son requests. Participants imagined requesting money from either that they request money from peers based on their relationship with a close-friend, distant-acquaintance, or extremely distant-acquain- the requestee. Specifically, we examine how the strength of social tance and chose to request the money in person or using Venmo. ties affects the choice of peer debt repayment methods. We find that Then, participants rated how uncomfortable and impersonal it would consumers are less likely to request money via digital payment meth- feel to request the money. Consistent with previous studies, fewer ods with their stronger social ties (i.e., close friends) compared to people chose to request via Venmo with a close-friend (25%) com- their weaker social ties (i.e., acquaintances). Mediation results in- pared to distant-acquaintances (42%; p < .001) or extremely distant- dicate that, with close friends, fewer consumers choose to request acquaintance (45%; p < .001). Conversely, more people chose to via Venmo because digital payment methods seem impersonal. For request in person with a close-friend (75%) than with a distant- weaker social ties, however, people attempt to avoid fiscal confron- acquaintance (58%) or an extremely-distant-acquaintance (55%; p tation, thus, more people choose to request money via Venmo. In < .001). Mediation results showed that anticipated discomfort from sum, consumers selectively use digital payment methods to match requesting in-person (indirect effect: 95% CI [-.07, .00]) and percep- relationship priorities. tions of impersonality using Venmo (indirect effect: 95% CI [.07, In Study 1 (N = 300), participants recalled a real instance where .00]) mediated the relationship between the strength of social dis- they requested money from someone they know. They described the tance and choice of request method. situation, indicated how they requested the money (i.e., in-person, Our findings suggest that people request money differently de- phone call, text message, electronic payment method, email, other), pending on the strength of their social ties with the requestee. With rated how they felt requesting the money, and rated their social close friends, people are less likely to use digital payment methods ties with the requestee. Results showed that people who requested because that makes the transaction feel less personal; instead, peo- the money via electronic methods felt more impersonal than those ple request in person with those with whom they share close social who requested in person (ps < .05). Moreover, correlational results bonds. At the same time, with weaker social ties (i.e., acquaintanc- showed that the weaker the social ties with the requestee, the greater es), people are more likely to use digital payment methods to avoid the discomfort people felt when requesting the money in person (p < fiscal confrontation and discomfort. Taken together, we see that con- .01). In short, we find that different request methods (e.g., in-person sumers adjust their approach to peer financial interactions based on versus electronic payment methods) elicit different socio-emotional the needs of the social relationship at hand. consequences depending on the social distance with the requestee. In Study 2, a between-subjects experiment (N = 300), par- REFERENCES ticipants imagined requesting money that was owed from either a Consumer Financial Protection Bureau. (2017). Financial close-friend or an acquaintance. Participants indicated if they would well-being in America. Retrieved from https://files. request the money either (1) in-person, (2) via Venmo, or (3) they consumerfinance.gov/f/documents/201709_cfpb_financial- would choose not to ask. For close-friends, participants were less well-being-in-America.pdf likely to request money using Venmo relative to requesting from an Federal Reserve Board. (2018). Report on the Economic Well-Being acquaintance (31% vs. 54%; p < .01). However, people were direc- of U.S. Households in 2017. Retrieved from https://www. tionally more likely to request the money in person with a close- federalreserve.gov/publications/files/2017-report-economic- friend versus an acquaintance (52% vs. 41%, p = .09). That is, peo- well-being-us-households-201805.pdf ple were less likely to use digital payment methods with stronger Hennecke, M., & Converse, B. A. (2017). Next week, next month, (vs. weaker) social ties, while more likely to request in person with next year: How perceived temporal boundaries affect initiation stronger (vs. weaker) social ties. expectations. Social Psychological and Personality Science, Study 3 (N = 503) replicated this basic effect and added a Ven- 8(8), 918-926. dor condition to test how a strictly transactional relationship with Krueger, D. W. (1986). Money, success, and success phobia. The the requestee affects the choice of request method. In addition, we last taboo: Money as a symbol and reality in psychotherapy added a new method of request, a standard bank app, to see how a and psychoanalysis, 3, 16. purely transactional, yet still digital, payment method differs from Larrick, R. P., & Soll, J. B. (2008). The MPG illusion. Science. Venmo and in-person requests. All participants read a scenario where 320(5883):1593-1594. a close-friend, acquaintance, or vendor owed them money (random- Loewenstein, G., & Prelec, D. (1992). Anomalies in intertemporal ly assigned). Participants were asked whether they would request the choice: Evidence and an interpretation. The Quarterly Journal money either (1) in person, (2) via Venmo, (3) via a Bank app, or (4) of Economics, 107(2), 573-597. they would not ask. Replicating Study 2, people were less likely to Lynch Jr, J. G., Netemeyer, R. G., Spiller, S. A., & Zammit, A. request the money that was owed via Venmo when requesting from a (2009). A generalizable scale of propensity to plan: the long close-friend (19%) versus an acquaintance (35%; p < .01) or a ven- and the short of planning for time and for money. Journal of dor (30%; p < .05). However, people were directionally more likely Consumer Research, 37(1), 108-128. to request the money in person with a close-friend (53%) versus an Morewedge, C. K., Holtzman, L., & Epley, N. (2007). Unfixed acquaintance (43%) or a vendor (43%). Moreover, people were sub- resources: Perceived costs, consumption, and the accessible stantially more likely to request the money using a standard bank account effect.Journal of Consumer Research, 34(4), 459-467. app with a vendor (25%) versus a close-friend (5%; p < .001) and an Parsons, C. A., & Van Wesep, E. D. (2013). The timing of pay. acquaintance (5%; p < .001). That is, people were more likely to use Journal of Financial Economics, 109(2), 373-397. a purely transactional payment app to request money from a strictly transactional relationship (i.e., with the vendor). 1068 / New Learnings in Financial Decision-Making: Influences on How Consumers Use and Obtain Financial Resources

Rose, G. M., & Orr, L. M. (2007). Measuring and exploring Trachtman, R. (1999). The money taboo: Its effects in everyday symbolic money meanings. Psychology & Marketing, 24(9), life and in the practice of psychotherapy. Clinical Social Work 743-761. Journal, 27(3), 275-288. Schroeder, J., Kardas, M., & Epley, N. (2017). The humanizing Tu, Y., & Soman, D. (2014). The categorization of time and its voice: Speech reveals, and text conceals, a more thoughtful impact on task initiation. Journal of Consumer Research, mind in the midst of disagreement. Psychological Science, 41(3), 810-822. 28(12), 1745-1762. Ülkümen, G., Thomas, M., & Morwitz, V. G. (2008). Will I spend Soster, R. L., Monga, A., & Bearden, W. O. (2010). Tracking costs more in 12 months or a year? The effect of ease of estimation of time and money: How accounting periods affect mental and confidence on budget estimates.Journal of Consumer accounting. Journal of Consumer Research, 37(4), 712-721. Research, 35(2), 245-256. Stephens, M., & Unayama, T. (2011). The consumption response Urminsky, O., & Zauberman, G. (2015). The psychology of to seasonal income: Evidence from Japanese public pension intertemporal preferences. The Wiley Blackwell handbook of benefits. American Economic Journal: Applied Economics, judgment and decision making, 2, 141-181. 3(4), 86-118. Walther, J. B. (1995). Relational aspects of computer-mediated Thaler, R. (1985). Mental accounting and consumer choice. communication: Experimental observations over time. Marketing science, 4(3), 199-214. Organization Science, 6(2), 186-203. Say It Like You Mean It: Revealing Market Narratives through Text Analysis Chair: Ignacio Luri, DePaul University, USA

Paper #1: Analyzing Brand Backstories: Combining Textual they borrow from a range of theories including narratology, concep- and Multi-sensory Data tual metaphor theory, rituals and identity projects, or diffusion the- Cristel Russell, Pepperdine University, USA ory. This session will facilitate scholarly discussion of the complex interplay between markets and discourses. A big theme in it guiding Paper #2: Rituals in Rap: Remaking the Road to Riches the session and discussion with the audience will be the performative Tonya Bradford, University of California Irvine, USA aspect of discourses or, in other words, how industries, firms, con- Courtney Jemison, University of California Irvine, USA sumers “fake it till they make it”, slowly creating a reality (a back- Paper #3: Metaphorical Framings in Market Discourse story, public consensus, wealth, a practice) by telling it. Jointly, we Ignacio Luri, DePaul University, USA anticipate sparking conversation around questions such as: 1) how Hope Schau, University of Arizona, USA does a historical view of a specific market discourse (i.e. longitudi- Bikram Ghosh, University of Arizona, USA nal, etymological, heritage or origin stories) inform understanding Paper #4: Written in the Sand: Tracing Practice Diffusion of current discourses? 2) how can industries, firms and consumers Across Time and Space utilize these frameworks to strategically tell their stories? We ex- Melissa Archpru Akaka, University of Denver, USA pect lively audience participation and involvement based around the Hope Schau, University of Arizona, USA timeliness and relevance of the contexts and methods, as well as the eclectic theoretical perspectives. SESSION OVERVIEW In this session, we present four empirical papers that, in addition Analyzing Brand Backstories: Combining Textual and to sharing text analysis as a methodological tool, explore the power Multi-sensory Data of narratives. The contexts studied range from market perceptions of debt to brand backstories, and from the diffusion of surfing to rap EXTENDED ABSTRACT lyrics. Each of these four papers utilizes a longitudinal dataset to Brand managers strategically construct brand backstories from study evolving market stories, as told by firms, news media, rap art- a curated set of facts, signs, symbols, with accompanying evidentiary ists or surfers. The researchers in this session share the conviction support that, when knitted together, provide the firm’s official narra- that discourses in the market represent and construct reality. To study tive of the brand’s origins, including novel insights and revelations them, we engage in diverse forms of text analysis, widely consid- regarding the early days of the brand. Our research question is: How ered a next frontier in marketing research (Berger et al. 2020), that can analysis of multi-sensory archival data reveal how practitioners include theme-coding, content analysis, and automated text analysis. design the backstory, curating brand materials to offer the ‘real’, ‘au- Beyond a shared focus on textual data and analysis, these projects thentic’ story and the ultimate ‘truth’ about the brand? use a variety of data sources and theoretical lenses to contribute to This multi-year, multi-country project focuses on the growing our understanding of how narratives represent and drive the market. practice of the brand backstory, whereby firms grant consumers ac- The presentations will dive deep into the everyday conversations of cess to facts about the brand origins, ‘private’ backstage information. each topic to uncover the subtext behind everyday language and how Often brand backstories unfold in the home of the brand (its head- it reflects deeper realities in the market. quarters or factory) or an authoritative institution, such as a museum. The first presentation examines how brand managers strategi- These sites (headquarters, factory, museum) frame the backstory, im- cally construct brand backstories from a curated set of facts, signs, buing it with an aura of authenticity that bolsters the firm’s authorial symbols, that provide the firm’s official narrative of the brand’s ori- voice. gins. The authors reveal how practitioners design the backstory, cu- We draw on graphic narrative theory, a narratology theory rating brand materials to offer the ‘real’, ‘authentic’ story and the grounded in the study of graphic novels and comics (bandes dessi- ultimate ‘truth’ about the brand. The next presentation focuses on the nées) and which allows a fuller understanding of storytelling via important topic of wealth disparities and the non-linear narratives of words and images. In particular, Thierry Groensteen’s work on the poverty and wealth. Rap music has had a fascination with these nar- System of Comics provides a novel lens through which we can inte- ratives, and their lyrics negotiate and link both ends. Though spend- grate visual, verbal, spatial, and chronological expressions in narra- ing rituals are prominent in hip-hop lyrics, the authors explore how tive analysis. We share the insights gained from our ongoing analysis hip-hop lyrics also may encourage engagement in wealth creation of four brand backstories. Our experience unearthed the challenges rituals. The third presentation explores the timely topic of personal of incorporating visual and other sensory data in primarily text-cen- debt, attempting to uncover how the media constructs the percep- tered consumer research and signals the need for narrative theories tion of debt and what it does for/to people in society. Employing a and rhetorical tools to supplement automated text analyses. variety of automated textual analysis on a large news database, the Our data consist of brand materials across four brand backsto- authors study in particular the role of metaphors framing issues in ries (an entertainment brand, two fast moving consumer good brands public discourse. The final presentation explores the dispersion of an and a nonprofit organization) that varied in the type of spatial brand indigenous practice, surfing, revealing that practice diffusion occurs backstory on offer (museum exhibit, factory tour and corporate head- through practice codification, transposition, and adaptation. quarter exhibit), in the duration (one time event and ongoing) and We structure the session as a series of four empirical presenta- location of the backstory (offsite museum and brand headquarters) tions tied to the theme of market narratives and linked by a method- as well as in the number and type of creators involved in the design ological focus on text. Each presentation will be 12 minutes leaving of the backstory (multiple creators all not members of the brand- 20-25 minutes for discussion and debate. While all the presentations owning firm vs. a sole creator employed by the brand-owning firm). have implications for market creation, public policy, and branding, Practitioners approached for this project shared not only their views

Advances in Consumer Research 1069 Volume 48, ©2020 1070 / Say It Like You Mean It: Revealing Market Narratives through Text Analysis and insights on their involvement with the brand backstory but also overcoming such differences. Art serves a functional purpose within access to historical documents and curatorial practices such as back- the African American community (French 2017; Karenga [1972] story creation manifestos or documentation describing the project 1997). Rap, a musical art form embedded in hip-hop culture, was scope of the brand backstories. once a marginal influence and has emerged as a prominent global Our data consist of in-depth qualitative with practitioners re- socialization force (French 2017; Lena 2006; Motley and Henderson sponsible for crafting and promoting brand backstories for the en- 2008). Motley and Henderson (2008; p. 243) note “Hip-hop culture tertainment brand Outrageous Fortune (OF), the sole creator for the influences styles of behavior…” factory tour of Herr’s Snacks, four practitioners for the headquarters Prior research identifies hip-hop influences on brand consump- of Girl Scouts USA, five for the Calissons du Roy Rene, a specialty tion (Burkhalter and Thornton 2014; Ferguson and Burkhalter 2015; almond delicacy. Interviews were conducted at the backstory site to North and Hargreaves 2007). While spending is integral to consump- enable the researchers to gain firsthand observations and an applied tion, so is saving (Bradford 2015). Studies of hip-hop lyrics suggest experience as a simulated consumer of the multisensorial backstory listeners are socialized to rituals in support of spending (Leung and environments discussed in the interviews. First-hand observations Kier 2010). Though spending rituals are prominent in hip-hop lyrics, were recorded as data through handwritten notes and supplemented we explore how hip-hop lyrics also may encourage engagement in the interviews carried out. The researchers visited each site on mul- wealth creation rituals. tiple occasions for lengthy periods of time. Follow up interviews Hip-hop lyrics often include implicit endorsements of specif- were conducted with three OF backstory practitioners in a group ic forms of consumption (Motley and Henderson 2008; Rehn and interview, one GSUSA backstory practitioner, and two members of Sköld 2005). Consider Petey Pablo’s song featuring Seagram’s gin the Tourism Office in charge of industrial tourism and the Calissons (Burkhalter and Thornton 2014) or Run-DMC and the “My Adidas” tours. Interviews centered on the processes involved in the creation campaign (Ferguson and Burkhalter 2015). Hip-hop music emerged and maintenance of the brand curation, the motivations of the pro- from urban and ethnic enclaves providing an outlet of expression for ducers for the genesis of the brand backstory, the type of informa- those in marginalized communities, and is now found globally across tion included, the narrative structure and spatial organization of the other marginalized communities (French 2017; Motley and Hender- exhibit, the goals and aims of the backstory during and after creation, son 2008). Hip-hop lyrics focus on current issues, where one such and general opinions on consumer reactions to the backstory. issue is the need for wealth creation (vs. spending). The primary em- We supplemented interview data with analysis of the curation phasis in hip-hop lyrics are on those spending rituals related to brand materials: all the OF materials from the Arts Department were pro- acquisition. However, there is evidence that hip-hop lyrics also may vided digitally as well as the Official ‘behind the Scenes’ DVD, the encourage performance of wealth creation rituals. tour guide manuals and instructions for the guided visits at Herr’s, The study explored lyrics of one of the wealthiest hip-hop art- GSUSA, and the Calissons (in as many iterations as were avail- ists, Shawn Carter—“Jay-Z”, who amassed wealth through a di- able), as well as archival documents, when available, outlining the versified portfolio music, real estate, sports management (Akhatar processes for selecting and organization curation materials into the 2018). Jay-Z shares his rags-to-riches story outlining his transition backstory space . We also relied on the hundreds of photos and vid- from poverty to drug-dealing “in the crucible of Brooklyn’s Marcy eos of the exhibits provided by the organizations as well as our own. projects.” (Simmons 2005). Data were collected from studio albums Our analysis of interview transcript data and the curated materi- released between 1996 and 2017 (212 songs). Lyrics were analyzed als were analyzed hermeneutically, noting the relationships between individually, within its album, and across albums. The data were codes and then nomothetically across interviews and data sources coded using codes associated with wealth accumulation from the (Glaser & Strauss 1967; Mick & Buhl 1992; Thompson, Locander & literature and the lyrics. Three themes emerged from that analysis Pollio 1989). A refinement of ideas and themes was achieved by way related to wealth creation rituals. of the constant comparative method whereby data were compared to Prior research on savings explores how individuals employ other sections that contained similar incidents and themes, as well as earmarks and social comparison in support of wealth creation and contrasted with other sections of data with differing themes (Gould- identity projects (Bradford 2009, 2015). Missing is an explanation ing 2002). of how wealth creation rituals may be introduced through music. We highlight the challenges of analyzing multi-sensory data. This research finds such rituals may be introduced through three key The brand backstories in our sample were multi-lingual, multi-cul- mechanisms: social comparison, nostalgia, and aspiration. Studies of tural and multi-industry. The data collection and analysis followed social comparison finds people seek salient others as points of ref- the procedure and protocols of multiple traditions: textual media erence (Festinger 1954). Research on wealth creation finds African data, textual interview data, visual brand collateral (brand adver- Americans employ social comparisons with whites (Bradford 2009). tising and promotional material), and experiential data (factories, The lyrics serve as a metaphorical device that provides a wealth cre- headquarters and museums). Complicating matters, the practitioner ation ritual model, and an alternative social referent for listeners who curated materials represent different punctuated moments with tem- are not likely to achieve acclaim. poral notions of brand attributes and market perceptions. Our data Ritual performance provides a link between the past and the fu- represent a process theoretic approach (Giesler and Thompson 2016) ture. Where nostalgia is employed by marketers to sustain brand vi- centering on the brand backstories, consisting of the process of brand tality (Brown, Kozinets, and Sherry 2003), these lyrics employ nos- becoming, or the ways in which brands enter the market and minds talgia to elevate notions of simplicity found in poverty as noble and of market constituents. worthy of retention amidst the complexity of wealth. Where prior research on voluntary simplicity focuses on reductions in consump- Rituals in Rap: Remaking the Road to Riches tion (Craig-Lees and Hill 2002), we find that the notion of simplicity is not segregated from material possession. Rather, an orientation EXTENDED ABSTRACT toward simplicity is emphasized with respect to consumption to sup- Wealth disparities have long garnered attention in lyrics of rap port performances of wealth creation rituals. Thus, nostalgia is em- music, and continues to do so with more recent lyrics focusing on Advances in Consumer Research (Volume 48) / 1071 ployed to facilitate balance between drives for material possessions being. To this end, we build on the established literature on discur- and security. sive frames (Barsalou 1992, Shoemaker and Reese 1996, Gamson et Continuity of rituals also inspires a future focus which may in- al.1992; Humphreys and Latour 2013) and we strengthen the known clude aspirations. Aspiration in consumption may be found in no- association in academic research between metaphors and frames, or tions of a future self (Bonsu and Belk 2003). We find these lyrics metaphorical frames (Boeynaems et al. 2017, Lee & Schwarz 2014, inspire aspiration not only related to identity but also through an Lakoff 1993, Lakoff and Johnson 2008). By spelling out an issue in adoption of wealth creation rituals. Early research purports a rela- relation to one particular schema and not another, frames become tionship between social class and savings patterns (Martineau 1958), shared understandings not only on what to think about the topic but yet our findings expand Bradford (2009) that illustrates that the im- on how to think about it. In contrast to the immense potential to pact of social class may be mitigated by adoption of rituals. harnessing metaphors, the alternative, unquestioningly echoing the This study contributes to the literature on how behaviors are in- metaphors used by others can backfire or result in the implicit en- fluenced through hip-hop lyrics (Motley and Henderson 2008). More dorsement of ideologies embedded in the language. specifically, we identify how wealth creation rituals within hip-hop Acknowledging its established role influencing and reflecting lyrics challenge notions of spending rituals, and also may inspire public discourse (Deephouse 1996, Humphreys 2010, Humphreys wealth creation rituals for all individuals. Wealth creation rituals and Latour 2013, Warren and Sorescu 2017), we turn to mass news proposed in these lyrics include a broader set of ritual participants, media for an investigation of the metaphors and frames in the topic and encourage performance continuity through social comparison, of debt. We conduct a large-scale text analysis on a comprehensive nostalgia and aspiration. database of ten years of news media articles. Employing textual methods at the forefront of market research innovation (Berger et al. Metaphorical Framings in Market Discourse 2020), we approach the research question: how do metaphors frame market-oriented discourse to guide action? Although we build on lit- EXTENDED ABSTRACT erature investigating discourse in market activism (Humphreys 2010; We communicate using borrowed tools. When businesses, aca- Karababa and Ger 2010; Kurzer and Cooper 2007; Sandlin and Cal- demics, or consumers quote, reference, meme, we are deliberately lahan 2009; Woodly 2015), we align more closely with a rarer and recontextualizing past words to make them do our biding. But all more recent body of research (Giesler and Veresiu 2014; Humphreys words are borrowed. As Bakhtin (1981) puts it, “language has been and Thompson 2014) concerned with the already established insti- taken over, shot through with intentions”, each word acquiring the tutional discourses against which activists sometimes compete. Our taste of a collective, a generation, an ideology. In this intertextual data is analyzed through a three-fold approach of interpretive deep game, market actors can unquestioningly replicate the linguistic reading, computer-assisted discourse analysis, and a deep learning conventions of communications on a topic, or they can strategi- Natural Language Processing (NLP) model of debt conversation. cally design their “borrowings” to fit their agenda. These latter, dis- Consistent with research on framing, we find metaphors to be of course-driving market actors shape public opinion by framing reality central importance not only defining a problem, but also suggesting through a specific lens. a solution. Far from being neutral, metaphors form frames regarding Metaphors are one such lens that influences perception of a topic the topic in terms of valence, responsibility assignment, and action. and drive behavior. Beyond being stylistic devices to attract attention These metaphors tend to cluster around topics, ideologies and lan- and spice conversation, metaphors work as frames, affecting how we guage that support their definition of the frame. We argue that meta- see the world or our definition of an issue. In this article, we explore phors associated with social issues form entrenched framings that how metaphors structure discourses in the context of debt. No other shape perceptions of the topic and our lived experience. While these product of the market economy carries the emotional and moral bag- framings reflect and perpetuate ideologies, we foreground an effect gage of debt. According to market narratives (Peñaloza and Barnhart that can be insidious rather than overt: an endorsement that is often- 2011), debt empowers as much as it constrains borrowers. With in- times tacit rather than explicit. Strategic marketing action requires novations in conversational AI such as Bank of America’s virtual carefully choosing the metaphorical frames that advance the firms assistant Erica (Legters 2019), the Consumer Financial Protection goals. Conversely, acts of resistance against the dominant discourses Bureau increasingly shifting the discourse to one of responsibility can be forced to articulate their criticisms through the same frames put on ordinary citizens (Hayashi 2019), and presidential candidates that work against them, or consciously break the mirage of figurative making debt cancellation a cornerstone of their campaigns, few dia- language that traps them. lectical battles are more current than the framing of consumer-firm responsibility regarding debt. Written in the Sand: Tracing Practice Diffusion Across Whether debt is framed as a tool, a burden, or a prison, the Time and Space power of the words describing it is compounded in the age of social media sharing and content aggregators. Largely repackaging facts EXTENDED ABSTRACT from elsewhere, news outlets engage in a constant parroting of each Prior consumer research underscores the importance of prac- other, much like social media users retweeting, reposting and shar- tices in shaping market-mediated experiences and social structures. ing. The language used to speak of an issue, with its frame and ideol- Little is known, however, about how practices diffuse across various ogy, is echoed by numerous sources and cements in public opinion. social and cultural contexts over time and the consequences of diffu- Persuasive influencers like savvy politicians and firms design their sion on consumption. scripts carefully to guide perception, opinion and action. This research involved 10 years of data collection, 2009-2018, We develop a framework to judge the aptness of metaphors to and multiple rounds of analysis, in which we iterated between data frame marketing-relevant issues such as personal debt to lead public and theories to interpret our findings. We include a variety of market- opinion in a given direction. The implications of this work extend related actors’ narratives that capture different practices and perspec- from the design of persuasive marketing communications to the pro- tives; the sample of interviewees was purposive. We interviewed motion of public policy or nudging behavior towards consumer well- a number of surfing historians and historical figures (e.g., contest 1072 / Say It Like You Mean It: Revealing Market Narratives through Text Analysis organizers, surfboard shapers, and entrepreneurs) in the evolution practices with a broader classification of value co-creation practices of surfing to verify our archival data and to better understand and and their relationship to practice diffusion. This provides a frame- capture specific first-person accounts of particular events.- Differ work for studying value co-creation and practice diffusion among ences in occupation and gender provided a variety of viewpoints and micro- and higher (meso- or macro-) levels of social phenomena perspectives of surfing as well as the history of surfing. We used a (Akaka et al., 2013). variety of archival sources, including newspapers, magazines, pho- This exploration of the dispersion of an indigenous practice, tos, websites, movies, and Observations include participation in the surfing, reveals that practice diffusion occurs through practice codi- practice of surfing, engaging in surfing related events or activities, fication, transposition, and adaptation. Our data reveal that practice spending time in surf-related retail environments, writing for a surf- diffusion is not the wholesale adoption of a practice. We show that ing magazine, and volunteering at a surfing competition. In addition, practice diffusion is influenced by the origins of a practice and the two of the authors spent considerable amounts of time as members of social and cultural contexts in which its diffusion occurs. Practice ad- local surfing communities and cultures prior to the initiation of this aptation is a process that integrates varying perspectives of a practice study.music, which served as surrogate data when first-person ac- into the broader nexus of understandings, doings and sayings associ- counts were unavailable Karababa and Ger (2011). With each source ated with that particular practice. We find that the as a practice dif- of data, we examined each artifact as an individual representation of fuses across contexts it has the potential to lead to social change and surfing and then considered its meaning within the wider scope of the emergence new social structures (subcultures of consumption). the surfing subculture. This approach enabled us to oscillate across micro- to meso- to macro-levels of social phenomena (Karababa and REFERENCES Ger 2011) within the surfing ecosystem (Akaka et al., 2013; Chan- Akaka, Melissa Archpru, Stephen L. Vargo, and Robert F. Lusch dler and Vargo 2011). (2013)”The complexity of context: A service ecosystems We approach these data from an ecosystems perspective (e.g., approach for international marketing.” Journal of International Giesler 2008; Vargo and Lusch 2016) that allows us to consider how Marketing 21.4: 1-20. particular events and practices fit into a broader socio-historic narra- Askegaard, S., & Linnet, J. T. (2011). Towards an epistemology of tive. We are particularly interested in understanding how micro-level consumer culture theory: Phenomenology and the context of social phenomena (e.g., practices) construct broader social structures context. Marketing Theory, 11(4), 381-404. (c.f. Giddens 1984) enabling the spread of practices. We utilize a lon- Akhatar, Allana (2018), “Here’s How Much Jay-Z Is Worth—and gitudinal and historical data set (Karababa and Ger 2011), through a How He Makes Money,” http://time.com/money/5108647/jay- process theorization approach (Giesler and Thompson 2016). z-net-worth/. Chronological accessibility of the emergence and re-emergence Bakhtin, M. (1981). Discourse in the Novel. The Dialogic of specific practices requires the ability to track the spread of a par- Imagination. Four essays, 301-331. ticular local artifact across time and space (Giesler and Thompson Barsalou, L., and Hale, C. (1993). Components of conceptual 2016). While the complete pollination of surfing around the globe is representation. From feature lists to recursive frames. beyond the scope of this particular study, we aim to understand how Berger, J., Humphreys, A., Ludwig, S., Moe, W. W., Netzer, O., practices diffuse across regions and cultures. & Schweidel, D. A. (2020). Uniting the tribes: Using text for Our historical research is guided by the work of Karababa and marketing insight. Journal of Marketing, 84(1), 1-25. Ger (2011) and a process theorization approach that focuses ana- Boeynaems, A., Burgers, C., Konijn, E. A., and Steen, G. J. (2017). lytical attention on change occurring in social collectives (Giesler The effects of metaphorical framing on political persuasion: and Thompson 2016). Change is made manifest in actions, ergo we A systematic literature review. Metaphor and Symbol, 32(2), utilize event-based data and analytical bracketing to organize the 118-134. events (Giesler and Thompson 2016: 502) that, in our context (surf- Bonsu, Samuel K. and Russell W. Belk (2003), “Do Not Go ing) and with our interpretive lens (diffusion) reveal the manner in Cheaply into That Good Night: Death-Ritual Consumption in which surfing spread across multiple locales. Specifically, we exam- Asante, Ghana,” Journal of Consumer Research, 30 (1), 41-55. ine what Giesler and Thompson (2016: 500) call topological change, Bradford, Tonya Williams (2009), “Intergenerationally Gifted Asset or the “shifts in patterns of power relationships.” These shifts from Dispositions,” Journal of Consumer Research, 36 (1), 93-111. one reality to the next disrupt the status quo relations causing to- --- (2015), “Beyond Fungible: Transforming Money into Moral and pological displacements (Giesler and Thompson 2016: 503). Based Social Resources,” Journal of Marketing, 79 (2), 79-97. on the precipitating tensions arising from power struggles regarding Burkhalter, Janée N. and Corliss G. Thornton (2014), “Advertising respectability and reputation, we identify power shifts, manifested to the Beat: An Analysis of Brand Placements in Hip-Hop through events, which enable the spread and perpetuation of surfing Music Videos,” Journal of Marketing Communications, 20 (5), (Giesler and Thompson 2016: 499). 366-382. Part of the challenge of understanding the practice diffusion is Chandler, J. D., & Vargo, S. L. (2011). Contextualization and value- that in studying practices and local culture, there is a tendency to fo- in-context: How context frames exchange. Marketing theory, cus on understanding particular micro-social processes and contexts. 11(1), 35-49. Askegaard and Linnet (2011: 381) argue “there is a need for bridging Craig-Lees, Margaret and Constance Hill (2002), “Understanding the analytical terrain between the anthropological search for thick Voluntary Simplifiers,”Psychology & Marketing, 19 (2), 187- description and deep immersion in the field, and the sociological 210. inclination towards broad social theories and movements that are Deephouse, David L. (1996), Does Isomorphism Legitimate? often quite remote from the emic illustrations of everyday life expe- Academy of Management Journal, 39 (4), 1024–1039. riences.” To investigate communities of practice (Wenger 1998) that Ferguson, Nakeisha S. and Janée N. Burkhalter (2015), “Yo, Dj, underpin surfing’s consumer culture, we draw on the work by Schau That’s My Brand: An Examination of Consumer Response to and colleagues (2009) to identify specific micro-level practices that Brand Placements in Hip-Hop Music,” Journal of Advertising, contribute to the emergence of common structure, and frame these 44 (1), 47-57. Advances in Consumer Research (Volume 48) / 1073

Festinger, Leon (1954), “A Theory of Social Comparison Kurzer, P., and Cooper, A. (2007). Consumer activism, EU Processes,” Human Relations, 7, 117-140. institutions and global markets: The struggle over biotech French, Kenneth (2017), “Geography of American Rap: Rap foods. Journal of Public Policy, 27(2), 103-128. Diffusion and Rap Centers,”GeoJournal , 82 (2), 259-272. Lakoff, G. (1993). “The contemporary theory of metaphor”. In Gamson, W. A., Croteau, D., Hoynes, W., and Sasson, T. (1992). A. Ortony (Ed.), Metaphor and Thought (pp. 202–251). Media images and the social construction of reality. Annual Cambridge: Cambridge University Press. [Second edition]. review of sociology, 18(1), 373-393. Lakoff, G., and Johnson, M. (2008). Metaphors We Live by. Giesler, Markus (2008) “Conflict and compromise: drama in Chicago. University of Chicago Press. marketplace evolution.” Journal of Consumer Research 34.6: Lee, S. W., & Schwarz, N. (2014). Metaphor in judgment and 739-753. decision making. Giesler, M., & Thompson, C. J. (2016). A tutorial in consumer Legters, B. (2019, July 15). Chatbots in Banking Got Smart. This is research: Process theorization in cultural consumer research. How They’ll Make Your Bank Better. Forbes, Retrieved from Journal of Consumer Research, 43(4), 497-508. https://www.forbes.com/sites/boblegters/2019/07/15/ Giesler, M., and Veresiu, E. (2014). Creating the responsible chatbots-in-banking-got-smart-this-is-how-theyll-make-your- consumer: Moralistic governance regimes and consumer bank-better/#19ffe69d70a7 on July 12, 2019. subjectivity. Journal of Consumer Research, 41(3), 840-857. Lena, Jennifer C. (2006), “Social Context and Musical Content of Groensteen, Thierry (2017), The Expanding Art of Comics Rap Music, 1979–1995,” Social Forces, 85 (1), 479-495. (translation by Ann Miller), University Press of Mississippi. Leung, Ambrose and Cheryl Kier (2010), “Music Preferences and Groensteen, Thierry (1988) « Narration as Supplement : An Young People’s Attitudes Towards Spending and Saving,” Archaelogy of the Infra-Narrative Foundations of Comics » Journal of Youth Studies, 13 (6), 681-698. (traduction, par Ann Miller, de mon texte figurant dans Bande Martineau, Pierre (1958), “Social Classes and Spending Behavior,” dessinée, récit et modernité, 1988), in Ann Miller & Bart Journal of Marketing, 23 (2), 121-130. Beaty (ed.), The French Comics Theory Reader, Leuven : Motley, Carol M. and Geraldine Rosa Henderson (2008), “The Presses universitaires de Louvain, 2014, p. 163-181. Global Hip-Hop Diaspora: Understanding the Culture,” Groensteen, Thierry (2013), Comics and Narration (translation by Journal of Business Research, 61 (3), 243-253. Ann Miller), The University of Mississippi Press. North, Adrian C. and David J. Hargreaves (2007), “Lifestyle Hayashi, Y. (2019, July 12). Financial Watchdog Shifts from Correlates of Musical Preference: 3. Travel, Money, Enforcer to Educator. The Wall Street Journal, Retrieved from Education, Employment and Health,” 35 (3), 473-497. https://www.wsj.com/articles/financial-watchdog-shifts-from- Peñaloza, L., and Barnhart, M. (2011). Living US capitalism: The enforcer-to-educator-11562922005 on July 12, 2019. normalization of credit/debt. Journal of Consumer Research, Hirschman, E. C. (1993). Ideology in consumer research, 1980 and 38(4), 743-762. 1990: A Marxist and feminist critique. Journal of consumer Rehn, Alf and David Sköld (2005), “‘I Love the Dough’: research, 19(4), 537-555. Rap Lyrics as a Minor Economic Literature,” Culture & Humphreys, A. (2010). Megamarketing: The creation of markets as Organization, 11 (1), 17-31. a social process. Journal of Marketing, 74(2), 1-19. Sandlin, J. A., and Callahan, J. L. (2009). Deviance, dissonance, Humphreys, A., and Latour, K. A. (2013). Framing the game: and détournement: culture jammers use of emotion in Assessing the impact of cultural representations on consumer consumer resistance. Journal of consumer culture, 9(1), 79- perceptions of legitimacy. Journal of Consumer Research, 115. 40(4), 773-795. Shoemaker, P. J., and Reese, S. D. (1996). Mediating the message Humphreys, A., and Thompson, C. J. (2014). Branding disaster: (pp. 781-795). White Plains, NY: Longman. Reestablishing trust through the ideological containment of Simmons, Russell (2005), “Builders & Titans: Jay-Z,” Time systemic risk anxieties. Journal of Consumer Research, 41(4), International (South Pacific Edition),(15), 73-73. 877-910. Vargo, Stephen L., and Robert F. Lusch (2016) “Institutions and Karababa, E., and Ger, G. (2010). Early modern Ottoman axioms: an extension and update of service-dominant logic.” coffeehouse culture and the formation of the consumer subject. Journal of the Academy of marketing Science 44.1: 5-23. Journal of Consumer Research, 37(5), 737-760. Warren, N. L., & Sorescu, A. (2017). When 1+ 1> 2: how investors Karenga, Maulana ([1972] 1997), “Black Art: Mute Matter Given react to new product releases announced concurrently with Force and Function,” in The Norton Anthology of African other corporate news. Journal of Marketing, 81(2), 64-82. American Literature, Vol. 1973-1977, ed. Henry Louis Gates Woodly, D. R. (2015). The politics of common sense: How social and Nellie Y. McKay, New York: Norton. movements use public discourse to change politics and win acceptance. Oxford University Press. Moving on! The Effect of Movement On, From and Embodiedin Products on Consumer Judgments Chairs: Ana Valenzuela, Baruch College, CUNY & ESADE – Universitat Ramon Llul, USA Eda Sayin, IE Business School – IE University, Spain

Paper #1: Is Fast Feminine? The Effect of Speed of Observed up/down). They suggest that static experiences lead to assimilative Hand-Motor Actions on Consumer Judgment and Behaviors behavior and dynamic experiences lead to complementary behavior Sumit Malik, IE Business School, Spain when they involve aversive situations. Eda Sayin, IE Business School – IE University, Spain Next, Hampton and Hildebrand demonstrate that certain movement delivered by devices, such as mobile vibrations, is consistently Paper #2: Assimilative Versus Complementary Effects of perceived as rewarding and can influence shopping decisions. They Embodied Verticality: Neural and Behavioral Evidence also provide evidence that conditioning drives this vibration-reward Massimiliano Ostinelli, Winthrop University, USA association. David Luna, University of Central Florida, USA Finally, Hadi, Valenzuela, Sridhar and Du develop a novel Torsten Ringberg, Copenhagen Business School, Denmark framework to predict consumer responses to synchronic integration Seidi Suurmets, Copenhagen Business School, Denmark of haptic effects into audiovisual mobile content, particularly move- Paper #3: Good Buzz, Bad Buzz: Using Vibrotactile Feedback to ment. They demonstrate that, while haptic augmentation may indeed Shape Consumer Choice improve consumer responses to advertising, the positive effect is William Hampton, University of St. Gallen, Switzerland moderated by dimensions of brand personality. Christian Hildebrand, University of St. Gallen, Switzerland This session should be of interest to a broad audience of re- Paper #4: Consumer Responses to Haptic Sensing of Movement searchers with substantive interests in the consequences of con- in Mobile Advertising sumer associations with object-directed actions and the experience Rhonda Hadi, University of Oxford, UK of movement within human-technology communication, as well as a Ana Valenzuela, Baruch College, CUNY & ESADE – theoretical interest in embodied cognition, associative processes, and Universitat Ramon Llul, USA information processing. The consequential nature of the dependent Karthik Sridhar, Baruch College, CUNY, USA variables provides insights for practitioners. O. H. Groth, Baruch College, CUNY, USA Is Fast Feminine? The Effect of Speed of Observed Hand SESSION OVERVIEW Movement on Consumer Judgment and Behaviors Personal experiences using hand movements and gestures have been shown to influence attitudes by activating heuristics, visual flu- EXTENDED ABSTRACT ency, and emotional attachment (Labroo & Nielsen, 2010; Hadi & Hand movements, gestures, and actions have been shown to in- Valenzuela, 2014; Streicher & Estes, 2016). This is particularly true fluence consumer attitudes and behavior (Labroo & Nielsen, 2010; since, unlike visual or auditory cues, tactile exchanges require direct Hadi & Valenzuela, 2014; Streicher & Estes, 2016). Extending this contact with a stimulus (Peck, 2010). Touch is considered the most literature, we investigate how speed of observed hand interaction “proximal” sense (Montagu & Matson, 1979) and has an idiosyn- with an advertised product affects consumer judgments. cratic capacity to evoke a sense of psychological closeness (Trope We argue that consumers cognitively relate speedy or fast (than & Liberman, 2010). A sense of movement can also be delivered gentle or slow) movements with masculinity (than femininity). They through a device itself. In fact, computer science research suggests may develop the associations about what it is feminine (vs. mascu- that “mediated touch” is possible, where haptic feedback from de- line) through social learning and repeated exposure to stimuli in their vices symbolizes the experience of the movement delivered by other environment (Becker, Kenrick, Neuberg, Blackwell, & Smith, 2007). (Gallace & Spence, 2010; Haans et al., 2006). For example, in advertisements, consumers oftentimes observe a fe- This session combines four papers that shed light on dimen- male protagonist performing a gentle hand interaction with a product sions, which drive how movement actions on objects, received from (e.g., applying L’Oréal face cream, spreading Nutella on a pancake objects or embodied on the trajectory of objects may have down- etc.) and male character doing a rather speedier hand movement stream consequences for advertising effectiveness, product prefer- (e.g., twisting the handle of Harley Davidson bike, spraying the AXE ences, and choice. body deodorizer etc.) We propose that recurrent exposure to such The first paper investigates how speed of observed hand-motor sensorimotor experiences in advertisements, dance-forms, films, etc. actions affect consumer judgments – driven by schematic associa- can lead the consumers to develop distinct associations of slow (vs. tions between speed of hand-motor actions and femininity-mascu- fast) hand movements with a more feminine (vs. masculine) behav- linity. Specifically, Malik and Sayin propose that observing slow (vs. ior (H#1). These associations may even be taken as an input to form fast) hand-motor actions evokes feminine (vs. masculine) schematic product judgments. We hypothesize that slow (vs. fast) hand move- associations, which leads to perceiving that the product itself is more ment with a product will alter the attribute perception of the product feminine and transfers onto higher product evaluations among fe- – such that the touched product will be perceived more feminine (vs. male (vs. male) consumers. masculine; H#2). Finally, drawing upon the social identity theory, The second paper by Ostinelli, Luna, Ringberg and Suurmets we expect that observing a slow (vs. fast) hand movement with a reconciles disparate findings in the semantic association connected product will evince a higher product evaluation among female (vs. with vertical movements by introducing the conceptual distinction male) consumers (H#3). This is because individuals are motivated to between static embodied experiences (i.e., being up/down in a sta- process information and make product choices that match their social tionary fashion) and dynamic embodied experiences (i.e., moving identity (Berger & Heath, 2007).

In Study 1A, MTurk workers (N = 158) completed the com- Study 3 had a 2(speed: slow vs. fast) x 2(participants’ gender: puter-based Implicit Association Test (IAT; Greenwald, McGhee, & male vs. female) between-subjects design. Postgraduate students Schwartz, 1998). The findings revealed a positive and significant as- (N=89) watched the stimuli video depicting a lateral hand-move- sociation for fast (vs. slow) movement with masculinity relative to ment, which varied on speed (slow: 2.5 cm./sec. vs. fast: 5 cm./sec) femininity [D-score=.67, t(157)=20.40, p<.01, 95% CI =(.60, .73), over a duvet cover (i.e., target product; fig.3). Participants, then, Cohen’s d=1.62]. The association of speed held across the partici- shared their purchase intention. An ANCOVA on purchase intention pant genders. revealed no significant main effect of speed (F(1,82)=.41, p =.52) In Study 1B, MTurk participants (N= 259) were assigned to and participant’s gender (F(1,82)=.08, p=.77) but a significant inter- either slow or fast speed between-subject conditions. Each partici- action effect (F(1,82)=3.75, p=.05; H#3). The effect-size improved pant watched three randomly-presented videos depicting common- upon excluding the covariates (F(1,85)= 5.09, p=.03). ly-encountered hand movements (i.e., ironing, vacuuming, and To summarize, across four studies, we demonstrate that observ- spray-cleaning) that varied on speed of action. They, then, shared ing hand movements, varying on speed, can affect the consumer re- their attribute associations for the action-doer i.e., feminine (gentle, sponses. We demonstrate that consumers evince higher evaluations nurturing, and caring) and masculine (forceful, aggressive, and dom- for not only products but also hand movement speed that matches inant; Hoffman & Borders, 2001), perceived speed, video character- their distinct social identity – such that female (vs. male) consumers istics (manipulation checks), age, and gender (control variables). A evoke higher evaluations for a product depicted with slow (vs. fast) repeated-measures ANCOVA with speed as between-subjects factor hand movement. These findings have direct implications for product and the three contexts as within-subjects factor on the feminine-at- management, advertising, and online marketing. tributes index (average of feminine attributes) showed a significant effect of speed (Mslow=4.91 vs. Mfast=4.35, p<.01; ηp2=.06) suggest- Assimilative Versus Complementary Effects of Embodied ing that slow speed evinced a feminine schematic association. Like- Verticality: Neural and Behavioral Evidence wise, the analysis on masculine-attributes index found that fast speed triggered a masculine schema (Mslow=3.30 vs. Mfast=4.06, p<.01; EXTENDED ABSTRACT ηp2=.07). These effects retained significance without the covariates The embodied cognition literature presents, at times, contradic- (p<.01) as well. tory effects of verticality. Most of the literature draws on the semantic Study 2 demonstrates the use of distinct associations from slow association between power, general valence, and verticality to show (vs. fast) speed as an input to form attribute perceptions of a prod- assimilative effects, where “up” is generally associated with greater uct (i.e., shaving foam; H#2). In a pre-test of 5 shaving foam cans power and with positivity, and “down” is associated with less power (fig.1), MTurk participants (N=103) identified “Ultra-Sensitive” la- and with negativity (Esteky, Wineman, & Wooten, 2018; Guido, Pi- chierri, Nataraajan, & Pino, 2016; Sundar & Noseworthy, 2014). belled foam as significantly lower on perceived masculinity (MUltra- At the same time, other research has shown complementary ef- Sensitive=4.66) compared to “Bold” labelled foam (MBold=5.87, p<.01). In main Study 2, MTurk participants (N=254) were assigned fects of verticality. Complementary responses occur when sensory- to either slow or fast between-subject conditions. The stimuli video motor experiences lead to effects that are inconsistent with, and even showed a male applying shaving foam (fig.2). Participants were opposite of, the underlying conceptual metaphor of “up is good.” asked to choose the foam product they believed that the protagonist These effects result from goal-oriented behaviors aimed at reducing is applying (“Ultra-Sensitive” or “Bold”). They, then, rated the indi- the gap between a current and a desired state (Zhang and Risen 2014). vidual on the schematic attribute associations, video characteristics, For example, Ostinelli, Luna and Ringberg (2014), show that mov- manipulation checks, and control variables as study 1. ing down leads individuals to experience lower self-worth, which We conducted a logistic regression to test whether slow (vs. motivates individuals to obtain better results in cognitive tasks. fast) hand movement led the featured product to be perceived as low/ We reconcile those seemingly disparate findings in the domain “Ultra-Sensitive” (vs. high/ “Bold”) on masculinity. The main effect of verticality by introducing the conceptual distinction between stat- of speed was significant (χ2(6)=3.71, p=.05; H#2) with no significant ic embodied experiences (i.e., being up/down in a stationary fash- covariates (p>.35). Specifically, participants chose less of the high ion) versus dynamic embodied experiences (i.e., moving up/down). We suggest that static experiences lead to assimilative behavior and masculine product (“Bold”) upon observing a slow (Pslow=48.81%) dynamic experiences lead to complementary behavior when they in- vs. fast (Pfast=60.62%) hand movement (χ2(1)=3.56, p=.06). We then ran a MANCOVA on schematic associations [DVs: femininity and volve aversive situations (e.g., moving down). masculinity index]. The analysis showed that slow (vs. fast) speed Study 1 activated a significant feminine schema (Mslow=4.53, Mfast=3.79; Study 1 is an EEG study that measures the brain states that result p<.01) but did not alter the masculine associations (Mslow=3.59, from the manipulation of static versus dynamic verticality. Previous Mfast=3.54; p=.79) for the protagonist. Gender and video characteristics had a significant effect (p<.01) but did not interact with speed to research suggests that the simulation of movement can activate the influence the associations. Subsequently, we ran a mediation analy- brain areas responsible for processing goals to prepare individuals to sis that showed that slow (vs. fast) observed hand movement elicits a cope with changes in the environment (Barsalou 2009; Borghi 2012; higher feminine (vs. masculine) schematic association, which medi- Gallese 2009; Hamilton 2013; Hickok and Hauser 2010; Pezzulo and ates the effect on choice of shaving foam product. Castelfranchi 2009; Semin and Cacioppo 2009; Sartori, Bucchioni, In Study 3, we investigate whether female (vs. male) consum- and Castiello 2013). From that basis, we predict that the simula- ers use speed of observed hand movement to form their purchase tion of downward movement activates the brain areas responsible intention. We expect female (vs. male) consumers to evince higher for compensatory goals aimed at reacting to unpleasant changes in evaluations upon observing a slow (vs. fast) hand interaction with a self-worth. This leads to complementary effects such as a preference product. This is because consumers are influenced by gender-iden- for products with positive associations (Sivanathan and Pettit 2010). tity maintenance motives and choose products that match their own We found that, compared to static scenarios, theta synchroni- social identity. zation in dynamic scenarios was higher in the parietal lobe. Syn- 1076 / Moving on! The Effect of Movement On, From and Embodiedin Products on Consumer Judgments cronization of theta power in the parietal lobe has previously been through the use of EEG and behavioral evidence. The implications of described as reflective of redirecting attention to goal-relevant se- our findings extend beyond the domain of verticality, as our frame- mantic information (Atienza, Crespo-Garcia, & Cantero, 2011). This work can help explain potential discrepancies in other embodied finding suggests that dynamic scenarios activate goal-directed be- cognition areas such as the experiences of weight and/or approach havior, which may drive the desire to compensate for an aversive behavior. state in the moving-down scenario. In addition, alpha synchronization was higher for dynamic imagery in the frontal, central and Good Buzz, Bad Buzz: Using Vibrotactile Feedback to parietal regions. This implies that dynamic scenarios do trigger an Shape Consumer Choice increased internally focused attention. Synchronization of alpha oscillations in parietal regions may also reflect increased process- EXTENDED ABSTRACT ing of motor, semantic and perceptual aspects of an imagined scene Although smartphones have been vibration-capable for over a (Bartsch et al., 2015). decade, mobile vibration features have been relatively unutilized by firms. Within the last two years, however, large e-commerce plat- Study 2 forms such as Amazon have quietly begun to pair mobile vibration Study 2 shows that down conditions activate more negative as- with certain consumer behaviors such as shopping cart adds. An- sociations than the up conditions, both in dynamic and static simula- ecdotally, people tend to report that a vibration from their mobile tions. The results show a main effect of direction such that respon- evokes an approach response, hinting at its potential as rewarding dents in the up condition had a higher self-worth than respondents stimulus, yet scholarly research examining the subjective percep- in the down condition (Mup = 3.23, SDup = 1.79; Mdown = 2.66, SDdown tions of mobile vibration is lacking. The current research integrates = 1.45, F(1,234) = 8.60, p < .01), controlling for task enjoyment. and builds upon existing haptic and reward perception research to Similarly, we analyzed the top three words that came to respondents’ examine three fundamental research questions relating to mobile vi- minds while imagining the vertical scenario. The results again show brations: (1) how they are subjectively perceived, (2) whether they a main effect of direction such that the up condition led to typing influence consumer choice, and (3) by what mechanism they exert more positive words than the down condition (Mup = .14, SDup = influence, focusing on the role of classical conditioning.

1.88; Mdown = -.36, SDdown = 1.70, F(1,234) = 4.54, p = .03). Almost every modern consumer now carries a vibrating mobile device that they interact with for up to 10 hours per day (Rob- Study 3 erts, Yaya, & Manolis, 2014). This massive uptick in mobile usage Study 3 investigates whether dynamic verticality leads to com- has been accompanied by increasing reports of mobile dependency plementary effects and static verticality leads to assimilative effects. (Salehan & Negahban, 2013), heightened impulsivity (Wilmer, The study was a 2 (direction: up vs. down) x 2 (movement: dynamic Hampton, Olino, Olson, & Chein, 2019), as well as experiences of vs. static) between-subjects design. A significant interaction emerged “phantom vibrations,” i.e. imaginary vibrations (Drouin, Kaiser, & (F(1, 212) = 13.94 , p < .001). In the dynamic condition, participants Miller, 2012). Academic research of mobile vibration has focused who imagined moving downward were willing to pay a greater price primarily on its use as an alerting or communicative stimuli (Sahami, premium for the high-status product relative to the low-status prod- Holleis, Schmidt, & Häkkilä, 2008; Saket, Prasojo, Huang, & Zhao, uct than those who imagine moving upward (Mmoving down = 18.68%, 2013), though one recent marketing study does suggest that mobile

SDmoving down = 34.42%; Mmoving up = 4.23%, SDmoving up = 28.79%, t(212 vibrations can boost the persuasiveness of certain text messages ) = -2.19, p = .03). This is consistent with a complementary effect of (Hadi & Valenzuela, 2019). Nonetheless, it remains unclear how vertical position. In the static condition, however participants in the consumers perceive, and are affected by, mobile vibrations in con- “being down condition” reported a lower price premium for the high sumptive contexts such as online shopping. We therefore conducted status iPad relative to the low status Klu tablet (Mbeing down = 4.82%, an initial pretest to establish a base understanding of consumer per-

SDbeing down = 31.04%; Mbeing up = 24.73%, SDbeing up = 39.71%; t(212) = ception of different mobile vibrations and to guide our selection of 3.10, p < .01). This pattern of results is consistent with assimilative vibrational stimuli. effects of vertical position (Zhang and Risen 2014). In the pretest, online participants (n=150) reported their subjective perceptions mobile vibrations (ranging from 25ms to 3200ms; Study 4 random presentation) along a variety of dimensions (reward, pleas- Study 4 provides support for a motivational goal priming ex- antness, arousal, etc.). The findings of this pretest demonstrate planation of complementary effects, including the mediating role of that reward perception varies by vibration duration F(3,147)=6.52, goal activation, through a moderation-of-process approach (Spencer, p<.001, with reward perception peaking at 400ms. Zanna, and Fong 2005). We introduced a goal satiation condition, Our pretest finding that 400ms mobile vibrations are reward- in which respondents had to write about their own success, and a ing, together with robust evidence that reward influences choice (Pa- control condition in which responents had to write about a trip to pies, Barsalou, & Press, 2015), led to our main Study 1 hypothesis: the grocery store. We found that in the dynamic down condition, consumers will prefer an action that produces a rewarding vibration, writing about one’s success led to lower willingness to pay for status compared to one that does not. Using a within-subjects design, we products than writing about a trip to the grocery store (Mgoal satiation = presented participants (n=150) with two visually-identical boxes

21.11%, SD goal satiation = 20.54% vs. Mno goal satiation = 36.9%, SDno goal satia- (analogous to online shopping carts). Participants were instructed to tion = 27.6% vs; t(202) = -2.09, p = .04). add items to either box as many times as they liked during a fixed Our work provides two main theoretical contributions: First, timeframe. As predicted, we find that participants added more items we introduce a goal systems framework to the study of verticality, re- to the vibration-producing box (MVibrationClicks=14.69; MNon- solving discrepancies in the existing research. Second, we introduce vibrationClicks=10.71; F(2,148)=4.08, p<.05). Interestingly, the and explain the difference between static and dynamic simulations extent to which participants found vibrations rewarding was sig- in embodied cognition. Methodologically, we add to the incipient nificantly correlated with the positive-valence of their daily mobile body of work that integrates neural and behavioral empirical studies notifications (r=.29, 95%-CI=[.11,.41], p<.01, suggesting that con- Advances in Consumer Research (Volume 48) / 1077 ditioning may underlie mobile vibration reward perception (further to utilize haptic feedback technology, which takes advantage of the examined in Study 3). sense of touch by applying vibrations to the user in synch with dis- In Study 2, we extend our findings with a lab-controlled study played movement and content (Brewster et al., 2007). In the market- (n=138, Swiss sample) examining the effect of rewarding vibro- place, some brands have begun to distribute such “haptified” experi- tactile stimuli on consumer choice in a highly-ecological online encd (e.g., mobile ads for Stoli vodka, users feel their phone vibrate shopping setting. Participants in this study used identical tablets when a woman shakes a cocktail; Johnson, 2015). to complete a custom-built online grocery shopping task. In the Importantly, unlike visual or auditory cues, tactile exchanges control condition, as in a typical online shopping interface, add- require direct contact with a stimulus (Peck, 2010). Touch is consid- ing an item updated the shopping cart, but produced no other feed- ered the most “proximal” sense (Montagu & Matson, 1979) and has back. In the vibration condition however, clicking the add-to-cart an idiosyncratic capacity to evoke a sense of psychological close- button triggered a rewarding vibration. We find that consumers in ness (Trope & Liberman, 2010). Computer science research suggests the vibration condition had significantly higher basket totals com- that “mediated social touch” is possible, where haptic feedback from pared to the control group (MVibrationBasketTotal=72.78CHF ; devices symbolizes the touch of another person (Gallace & Spence, MControlBasketTotal=57.84CHF; F(1,138)=10.01, p<.001, and that 2010; Haans et al., 2006). this effect is moderated by impulsivity. This result establishes that Drawing from this interdisciplinary literature, we propose that mobile vibrations can be rewarding and influence consumer choice, haptic sensations experienced in a mobile phone advertisement can but does not address the underlying mechanism. be perceived as a form of touch from the sender (i.e., the brand). Fur- In Study 3 we test the hypothesis that perception of mobile vi- ther, because haptic sensations are so uniquely associated with im- brations as rewards stems from classical conditioning (Gormezano mediate proximity and contact, haptic feedback in advertising may & Moore, 1966). To test this hypothesis, we adapted a classic as- have the ability to make these consumer-brand exchanges feel more sociative learning, Multi-armed Bandit (MAB) task (Gittins, Glaze- engaging and personal, ultimately augmenting mobile advertising brook, & Weber, 2011). In our incentive-compatible, two condition effectiveness. However, any positive effect of haptics will likely de- MAB task, we presented online participants (n=215, U.S. sample) pend on the brand’s characteristics since the source of haptic sensa- with four visually-identical slot machines, and tasked them with tions plays an important role in their interpretation (Martin, 2012). finding the most rewarding slot machine (each with unknown win IPG Media Labs recruited 1,136 Android mobile phone users in probability) to maximize their points across 4 blocks of 30 trials (120 the U.S. from an online panel (MedianAge = 25-34 years; 51.67% total; probability assignment randomized between blocks). Partici- female). Participants took a mobile survey, ostensibly about video pants were randomly assigned to either a (1) vibration-win condition content of their choosing. Before viewing the selected video con- in which wins were visually-salient and carried an approximately tent, all participants were served one of four pre-roll advertisements, neutral mobile vibration (100ms), while losses were less visually- which were either haptically augmented or not. Several other fac- salient and carried no vibration or (2) a vibration-loss condition, in tors were manipulated, including: density of the haptics (high ver- which the opposite was true: losses were salient and carried the vi- sus low), the presence/absence of a “bumper” notifying participants bration. In this way we conditioned participants to associate a pre- they were experiencing haptics, whether or not the “bumper” was viously-neutral vibration with either a positive stimulus (a win) or branded, and the ability to skip the ad or not. We accessed this data- an aversive stimulus (a loss). Participants then completed a MAB set and analyzed it to test whether the effect of haptic augmentation task questionnaire, followed by the shopping task (Study 2, but with was universal or contingent on the brand being advertised. To do so, 100ms vibration for item-adds) and final questionnaire battery. we employed a logit specification to investigate customer’s purchase Study 3’s findings support that consumers in the vibration-win intentions (Top Box and Second Box of purchase intentions were as- condition had significantly higher basket totals compared to the con- signed a value of 1, as we did not have access to the continuous vari- trol group (MVibrationWinBasketTotal=$158.13; MVibrationLoss- able) as a function of haptic augmentation at the category level while BasketTotal=$126.37; F(1,213)=9.62, p<.01) and, in line with our controlling for gender (due to the nature of the categories) and ad- theorizing, the effect of vibration condition on shopping basket total vertisement evaluation (see details in Exhibit 1). We were also able was mediated by how rewarding participants found the MAB task. to control for other properties of advertisements: (i) the density of The current study demonstrates that mobile vibration is parsed haptic effects (ii) whether the ad was skippable, and if so, whether it as a reward that can systematically influence consumer choice. Our was watched to its completion and, (iii) whether there was a branded findings show that even when the touch modality is kept constant, bumper/notification at the beginning of the advertisement. Overall, mobile vibrations evoke shifts in consumer preference toward stim- we found that while the overall impact of haptics on purchase inten- uli that are paired with a vibration. We further provide evidence a tion was positive and significant, a closer look revealed the effect classic conditioning mechanism likely underlies vibration reward only manifested for certain brands. The model also confirms (as sug- perception and its influence on consumer decision making. Together, gested by IPG itself) that high haptic density and the possibility to our findings have important implications for the effective design of skip the advertisement has a negative impact on purchase intentions, haptic human-machine interfaces in marketing and the role of vibro- but that adding a notification of haptics generates a positive effect tactile stimuli as a novel form of reward. instead. Notably, brands can differ not only in how familiar they are to Consumer Responses to Haptic Sensing of Movement in consumers (Kent & Allen, 1994) but also in personality dimensions Mobile Advertising (Aaker 1997; differences in perceived brand warmth are particularly robust; Kervyn et al., 2012). We expect that haptic feedback deliv- EXTENDED ABSTRACT ered from warm brands will exert a positive effect on downstream While consumers’ prolific smartphone usage provides a fruitful consumer responses, whereas the same feedback delivered from advertising avenue, small screen and limited ad visibility present a unfamiliar or cold brands will not. We conducted a series of experi- challenge to marketers (Bart et al., 2014). One way to circumvent ments to explore these predictions. All haptic effects were profes- this shortcomings and exploit the handheld nature of smartphones is sionally integrated by a developer of haptic feedback technology. 1078 / Moving on! The Effect of Movement On, From and Embodiedin Products on Consumer Judgments

Study 2 held brand familiarity constant, varied product category Bart, Yakov, Andrew T. Stephen and Miklos Sarvary (2014). as a within-subjects variable, and manipulated brand warmth and “Which products are best suited to mobile advertising? A haptics between-subjects. Study 3 (N=174) was a 2(Haptics: absent field study of mobile display advertising effects on consumer versus present) x 2(Brand Warmth: low versus high) x 2(Product attitudes and intentions,” Journal of Marketing Research, Category: Insurance versus Retail) mixed-model design. Within a 51(3), 270-285. video reel, participants viewed two advertisements: one for a retailer Berger, J., & Heath, C. (2007). “Where consumers diverge from and one for an insurance provider. In the high brand warmth condi- others: Identity signaling and product domains”. Journal of tion, participants viewed advertisements for Target and Geico, and Consumer Research, 34(2), 121-134. in the low brand warmth condition, participants viewed advertise- Borghi, Anna M. (2012), “Language Comprehension: Action, ments for Walmart and Allstate (all brands were selected based on Affordances and Goals,” inLanguage and Action in Cognitive pretest results). Afterwards, participants indicated their brand atti- Neuroscience, ed. Y. Corello and A. Bartolo, 125-43. tudes and those in the haptics-present conditions described how the Brewster, Stephen, Faraz Chohan, and Lorna Brown (2007). haptic effects felt to them. A repeated-measures ANOVA revealed “Tactile Feedback for Mobile Interactions,” in Proceedings no significant main effect of haptics on brand attitude, butasig- of the SIGCHI Conference on Human Factors in Computing nificant interactive effect of haptics and brand warmth onbrand Systems, ed. Mary B. Rosson and David Gilmore, New York, attitude (F(1,170)=5.91, p<.02). Planned contrasts revealed that as NY: ACM, 159-62. expected, haptics had a positive effect on attitudes towards high- Brough, A. R., Wilkie, J. E., Ma, J., Isaac, M. S., & Gal, D. (2016). warmth brands (F(1,170)=3.83, p<.06) but not towards low-warmth “Is eco-friendly unmanly? The green-feminine stereotype and brands. Notably, this pattern did not differ across product categories its effect on sustainable consumption.”Journal of Consumer (F(1,170)=.06, p>.80). In the haptics-present condition, independent Research, 43(4), 567-582. coders coded participants’ description of how the haptics felt. We Dohi, I., Yamada, F., & Asada, H. (2001). “The relationship also found a significant effect of brand warmth on affective reactions between masculinity and the Type A behavior pattern: The to the haptic effects (F(1,72)=8.80, p<.01). moderating effects of femininity.” Japanese Psychological Study 3 examined the effect of haptic advertising on a conse- Research, 43(2), 83-90. quential downstream variable: product choice. Participants (N=85) Drouin, M., Kaiser, D. H., & Miller, D. A. (2012). “Phantom viewed one of two 30-second advertisements on mobile phones ac- vibrations among undergraduates: Prevalence and associated cording to a 2 level (Haptics: absent versus present) between-subjects psychological characteristics.” Computers in Human Behavior, experimental design. The advertisement was for the chocolate brand 28(4), 1490–1496. Cadbury, which pre-test results revealed to be high in both familiar- Esteky, S., Wineman, J. D., & Wooten, D. B. (2018). The Influence ity and warmth. Afterwards, participants were offered a snack as a of Physical Elevation in Buildings on Risk Preferences: gift (either a Cadbury chocolate bar or an equivalently-sized cereal Evidence from a Pilot and Four Field Studies. Journal of bar). Binary logistic regression results revealed a main effect of hap- Consumer Psychology, 28(3), 487-494. tics on choice (χ2=3.80, p=.05). Participants in the haptics-present Gallese, Vittorio (2009). “Motor Abstraction: A Neuroscientific condition were more likely to choose the chocolate than those in the Account of How Action Goals and Intentions Are Mapped and haptics-absent condition (80% versus 59%, respectively). Understood,” Psychological Research, 73 (4), 486-98. This is the first academic research to explore the effects of hap- Gallace, Alberto and Charles Spence (2010). “The Science of tic advertising on consumer responses. We investigate an intersec- Interpersonal Touch: An Overview,” Neuroscience and Bio- tion between haptic feedback and source characteristics allowing us behavioral Reviews, 246–59. to differentiate between the nature of the haptic sensation andits Gittins, J., Glazebrook, K., & Weber, R. (2011). Multi-armed bandit evaluative implications. allocation indices. John Wiley & Sons. Gormezano, I., & Moore, J.W. (1966). Classical conditioning. REFERENCES Experimental Methods and Instrumentation in Psychology, 1, Aaker, J. L. (1997). “Dimensions of brand personality.” Journal of 385–420. Marketing Research, 34 (August), 347-57. Guido, G., Pichierri, M., Nataraajan, R., & Pino, G. (2016). Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and “Animated logos in mobile marketing communications: The interpreting interactions, Newbury Park, CA: Sage. roles of logo movement directions and trajectories.” Journal of Atienza, M., Crespo-Garcia, M., & Cantero, J. L. (2011). “Semantic Business Research, 69(12), 6048-6057. congruence enhances memory of episodic associations: Role Haans, Antal and Wijnand IJsselsteijn (2006). “Mediated Social of theta oscillations.” Journal of Cognitive Neuroscience, Touch: A Review of Current Research and Future Directions,” 23(1), 75–90. Virtual Reality, 149–59. Barsalou, Lawrence W. (2009). “Simulation, Situated Hadi, R., & Valenzuela, A. (2014). “A meaningful embrace: Conceptualization, and Prediction,” Philosophical Contingent effects of embodied cues of affection.”Journal of Transactions of the Royal Society B: Biological Sciences, 364 Consumer Psychology, 24(4), 520-532. (1521), 1281-89. Hadi, R., & Valenzuela, A. (2019). “Good Vibrations: Consumer Bartsch, F., Hamuni, G., Miskovic, V., Lang, P. J., & Keil, A. Responses to Technology-Mediated Haptic Feedback.” (2015). “Oscillatory brain activity in the alpha range is Journal of Consumer Research, forthcoming. modulated by the content of word-prompted mental imagery.” Hamilton, Antonia F. de C. (2013). “The Mirror Neuron System Psychophysiology, 52(6), 727–735. Contributes to Social Responding,” Cortex, 49 (10), 2957-59. Hanslmayr, S., & Staudigl, T. (2014). “How brain oscillations form memories - A processing based perspective on oscillatory subsequent memory effects.”NeuroImage , 85, 648–655. Advances in Consumer Research (Volume 48) / 1079

Hickok, Gregory and Marc Hauser (2010). “(Mis) Understanding Sahami, A., Holleis, P., Schmidt, A., & Häkkilä, J. (2008). “Rich Mirror Neurons.” Current Biology, 20 (14), R593-R94. tactile output on mobile devices.” European Conference on Hoffman, R. M., & Borders, L. D. (2001). “Twenty-five years Ambient Intelligence, 210–221. Springer. after the Bem Sex-Role Inventory: A reassessment and new Saket, B., Prasojo, C., Huang, Y., & Zhao, S. (2013). “Designing issues regarding classification variability.” Measurement and an effective vibration-based notification interface for mobile Evaluation in Counseling and Development, 34(1), 39-55. phones.” Proceedings of the 2013 Conference on Computer Johnson, L. (October 29, 2015). “Stoli’s Mobile Ads Let You Supported Cooperative Work, 149–1504. ACM. Actually Feel a Cocktail Being Made in Your Hand.” Ad Week. Salehan, M., & Negahban, A. (2013). “Social networking on Retrieved from www.adweek.com smartphones: When mobile phones become addictive.” Kent, Robert J. and Chris T. Allen (1994). “Competitive Computers in Human Behavior, 29(6), 2632–2639. Interference Effects in Consumer Memory for Advertising: Sartori, Luisa, Giulia Bucchioni, and Umberto Castiello (2013). The Role of Brand Familiarity,” Journal of Marketing, 58(3), “When emulation becomes reciprocity,” Social Cognitive and 97-105. Affective Neuroscience, (6),8 662–669. Kervyn, Nicolas, Susan T. Fiske, and Chris Malone (2012). “Brands Semin, Gün R. and John T. Cacioppo (2009), “From Embodied as Intentional Agents Framework: How Perceived Intentions Representation to Co-Regulation,” in Mirror Neuron Systems, and Ability Can Map Brand Perception,” Journal of Consumer ed. J.A. Pineda: Humana Press, 107-20. Psychology, 22(2), 166–76. Sivanathan, Niro and Nathan Pettit (2010),”Protecting the Self Krishna, A., & Morrin, M. (2008). “Does touch affect taste? The through Consumption: Status Goods as Affirmational perceptual transfer of product container haptic cues.” Journal Commodities,”Journal of Experimental Social of Consumer Research, 34(6), 807-818. Psychology,46(3),564-70. Kwon, M., & Adaval, R. (2018). “Going against the flow: The Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). “Establishing effects of dynamic sensorimotor experiences on consumer a Causal Chain: Why Experiments Are Often More Effective choice.” Journal of Consumer Research, 44(6), 1358-1378. than Mediational Analyses in Examining Psychological Labroo, A. A., & Nielsen, J. H. (2010). “Half the thrill is in the Processes.” Journal of Personality and Social Psychology, chase: Twisted inferences from embodied cognitions and brand 89(6), 845-851. evaluation.” Journal of Consumer Research, 37(1), 143-158. Streicher, M. C., & Estes, Z. (2016).”Multisensory interaction in Montagu, Ashley, and Floyd W. Matson (1979). The Human product choice: Grasping a product affects choice of other seen Connection. St Louis, MO: McGraw-Hill. products.” Journal of Consumer Psychology, 26(4), 558-565. Ostinelli, Massimiliano, David Luna, and Torsten Ringberg (2014). Sundar, A., & Noseworthy, T. J. (2014). “Place the logo high or “When up brings you down: The effects of imagined vertical low? Using conceptual metaphors of power in packaging movements on motivation, performance, and consumer design.” Journal of Marketing, 78(5), 138-151. behavior.” Journal of Consumer Psychology, 24(2), 271–283. Trope, Yaacov and Nira Liberman (2010). “Construal-Level Theory Peck, Joann (2010). “Does Touch Matter? Insights from Haptic of Psychological Distance,” Psychological Review, 117(2), Research in Marketing,” in Sensory Marketing: Research on 440–63. the Sensuality of Products, ed. Aradhna Krishna, New York, Wilmer, H. H., Hampton, W. H., Olino, T. M., Olson, I. R., & NY: Routledge, 17-31. Chein, J. M. (2019). “Wired to be connected? Links between Papies, E. K., Barsalou, L. W., & Press, G. (2015). “Grounding mobile technology engagement, intertemporal preference and desire and motivated behavior: A theoretical framework and frontostriatal white matter connectivity.” Social Cognitive and review of empirical evidence.” The Psychology of Desire, Affective Neuroscience, 14(4), 367–379. 36–60. Zhang, Yan, and Jane L. Risen (2014). “Embodied Motivation: Pezzulo, Giovanni and Cristiano Castelfranchi (2009). “Thinking Using a Goal Systems Framework to Understand the as the Control of Imagination: A Conceptual Framework for Preference for Social and Physical Warmth,” Journal of Goal-Directed Systems,” Psychological Research, 73 (4 ), Personality and Social Psychology, 107(6), 965–977. 559-77. Roberts, J., Yaya, L., & Manolis, C. (2014). “The invisible addiction: Cell-phone activities and addiction among male and female college students.” Journal of Behavioral Addictions, 3(4), 254–265. Self-Brand Connection in the Age of Technology Chair: Claire Linares, HEC Paris, France

Paper #1: We Can Look Like Our Brands: Facial Appearance Finally, contributing to expanding research on online reviews Anne-Laure Sellier, HEC Paris, France and electronic word-of-mouth, Yoon and Khamitov share an unex- Claire Linares, HEC Paris, France pected finding regarding highly self-brand connected consumers. Al- though one might expect such consumers to generate positive word- Paper #2: From Sweetheart to Scapegoat: Brand Selfie-taking of-mouth, they show that highly self-brand connected consumers Shapes Consumer Behavior feel psychologically threatened when the brand is used by others. In Reto Hofstetter, University of Lucerne, Switzerland turn, psychological threat leads participants who do not want to share Gabriela Funk, University of Lucerne, Switzerland the brand to engage in negative word-of-mouth. Leslie John, Harvard Business School, USA Overall, these papers demonstrate how self-brand connection Paper #3: When and Why Metaphors Affect Consumer–Brand can unfold in rapid and fundamental ways, and particularly resonates Relationships in this age of technology. We believe this session can create a rich Alberto Lopez, Tecnológico de Monterrey, Mexico discussion, hopefully renewing interest for research on this topic. Martin Reimann, University of Arizona, USA This session is likely to attract researchers working on a broad array Raquel Castaño, Tecnológico de Monterrey, Mexico of topics, brands, identity, the self, but also technology, face percep- Paper #4: Hands off My Brand: Strong Self-Brand Connection tion, embodied cognition, metaphors or word-of-mouth. and Psychological Threat Nari Yoon, Indiana University, USA We Can Look Like Our Brands: The Manifestation of Mansur Khamitov, Nanyang Technological University, Adopted Brands in Facial Appearance Singapore EXTENDED ABSTRACT SESSION OVERVIEW Considerable branding research suggests that brands can evoke Consumers live in a digitalized environment in which they ex- general brand user human age and/or body type stereotypes (e.g., a perience brands both on- and offline. Extensive research has stud- brand user is a grandfather, Belk, Bahn, and Mayer 1982; Absolut ied self-brand connection offline, the extent to which brands overlap Vodka is a hip 25-year old, Aaker 1997) and that possessions ac- with the self, arguing that brands can be part of the self (Belk 1988) tively contribute to our identities (e.g., Belk 1988; Berger and Heath and showing that brands help consumers enact their social identities 2007; Escalas and Bettman 2003). We ask the reverse question: can (Escalas and Bettman 2003). Further research is needed, however, consumers’ faces evoke the brands they are loyal to? If the bulk of to investigate how self-brand connection can form, how it can affect research on facial appearance has focused on how it influences so- consumers in powerful ways, and what prospects it can open in an cial perceptions (e.g., Ballew and Todorov 2007), recent research increasingly technological world. shows that social perceptions can shape facial appearance. In par- This session explores different facets of self-brand connection. ticular, Zwebner, Sellier, Rosenfeld, Goldenberg, and Mayo (2017) How can self-brand connection be reflected in our self and in our demonstrated that given names produce a face-name matching effect, actions? How can perceptions of self-brand connection arise? How the finding that both a social perceiver and a computer - canaccu is self-brand connection further shaped in our technological environ- rately match a person’s given name to their face, above chance level. ment? The four papers provide insight into these questions and more. Zwebner et al. (2017) pinpoint the existence of shared face-name In particular, they should foster a discussion about the interplay be- stereotypes (e.g., Gordon and Tanaka 2011; Tanaka 2001) to explain tween self-brand connection and different forms of technology, such the effect. They suggest that we live up to these stereotypes to meet as facial recognition, selfies, apps and electronic word-of-mouth. the social expectations of how someone with our given name should First, at a time when facial recognition technologies are ever look. more precise and pervasive, Sellier and Linares investigate if brands The present research examines whether brands that consumers that consumers are loyal to can manifest in their faces. They find are loyal to can similarly manifest physically in our faces. In three support for a face-brand matching effect, whereby a social perceiver studies, we find support for a face-brand matching effect, whereby can accurately match a person’s perfume brand to their face, above a social perceiver can accurately match a person’s perfume brand to chance level. This means that we can look like our brands. their face, above chance level. We propose that brands that consum- Next, Hofstetter, Funk, and John show how a benign act such ers actively link to their identities may similarly manifest in their as taking a selfie with a brand can trigger self-brand connection and faces. Unlike given names, brands are social tags that individuals in turn impact brand preference. Eight studies provide evidence that choose to embrace. Considerable research suggests that consumer consumers’ brand preference is increased after taking a selfie with it, identity (both actual and ideal) markedly influences the connection even if the brand is of lower utility than the alternative. Consumers’ to brands (e.g., Escalas and Bettman 2003; Levy 1959; Malär et al. self-inferences account for this effect. 2011; McCracken 1986) and that brands actively help consumers In a field study, and using a new app transmitting relationship build their self-identities (e.g., McCracken 1989; Escalas and Bet- metaphors and tracking real-time behavior, Lopez, Reimann, and tman 2019). It is the resulting interdependence between consumers Castaño explore the role of self-brand connection in the influence of and their brands that leads to the development of a proper relation- relationship metaphors (e.g., complete a puzzle and breaking some- ship (e.g., Fournier 1998). It may be, then, that after enough time in thing) on brand relationships. They show that in the presence (ab- a brand relationship, consumers begin to look like a stereotype of the sence) of a brand transgression, a negative (positive) metaphor will brand user, if only to signal to others around them that they “carry” decrease (increase) the strength of a consumer’s brand relationship. the social identities enacted by the brand.

Advances in Consumer Research 1080 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1081 The repeated finding of a face-brand matching effect means that From Sweetheart to Scapegoat: Brand Selfie-taking some evocation of the perfume brand we have been loyal to for a Shapes Consumer Behavior long time may get tattooed in our faces. This means that a social tag may show up in our face more rapidly than was known to date. EXTENDED ABSTRACT A second possibility is that people choose brands because they look Increasingly, consumers are taking self-photos and marketers, like the brand user stereotype ex ante and self-select into brands that eager to capitalize on this trend, routinely encourage consumers they can more easily live up to. to take “brand-selfies”—self-photos with a brand logo or product This research focuses on the case of perfume brands, as per- from that brand in tow. Despite the pervasiveness of selfie-taking, fume users symbolically identify with their fragrances (Aaker 1997; and marketers’ attempts to promote brand selfie-taking in particular, Belk 1988), particularly in France, where our data was collected little is known about the effect of this behavior on the selfie-takers (Statista 2017). In each study, female participants saw the headshots themselves. of 10 to 30 targets that they did not know prior to the study. For each We suggest that consumers’ compliance with such requests headshot, participants chose, out of two perfume brands, the one they sparks a self-inferential process that leads the consumer to feel con- thought was the target’s actual perfume: one brand was the target’s nected to the brand (e.g., “If I took the brand selfie, I must feel con- true perfume, the other was a filler. We consistently found evidence nected to this brand”). We base this on self-perception theory which of a face-brand matching effect: Across all headshots in each study, holds that people have a tendency to infer who they are by looking to participants picked the targets’ true perfumes above chance level their behavior (Bem 1967; Chaiken and Baldwin 1981; Fazio, Zan- (here, 50%). We extensively calibrated the stimuli to minimize con- na, and Cooper 1977). The theory has already been useful in explain- founds such as brand familiarity, the perfume brands’ release dates, ing many consumer behavioral and social phenomena (e.g., Grewal ethnicity, socioeconomic factors, target age and any difference be- et al. 2019; Summers, Smith, and Reczek 2016; Xu, Shen, and Wyer tween the target and filler brands that we used. 2012). Building on prior findings, we propose that taking a self-pho- In study 1, we photographed the faces of 10 Caucasian women to together with a brand can trigger self-inferences that increase self- who had been loyally wearing a well-known perfume for at least brand connection. Prior work in consumer behavior indicates that, four and up to 15 years. In this study, the photographs were framed in turn, feeling connected or attached to a brand can foster positive to include the face, the shoulders and the beginning of the chest. behaviors toward the given brand, for example increased purchase We paired their true perfumes with fillers that were perceived as intention (Whan Park et al. 2010). However, based on attribution equivalently well-known. All the perfumes were from the high-end theory (Festinger 1957; Heider 1958) we further suggest that when perfume market and were overall in an equivalent price range. Re- a consumer encounters a less-than-stellar image of herself in a brand sults from a t test and a logistic mixed-effect regression that accounts selfie, instead of revising her impression of her own attractiveness for the random selection of faces showed that participants found the (downward)—a threatening undertaking because it could be dam- targets’ true perfumes significantly more than random chance (50%). aging for the ego—she may instead use the brand as a scapegoat, In study 2, we further controlled for differences related to the projecting her personal dissatisfaction onto the brand. two perfume options and for age. Headshots were also more tightly Eight studies support this account. In a dataset of 283,140 user framed on the oval of the face. We developed two sets of 10 head- reviews from Yelp, study 1 documented a positive association be- shots, such that each headshot in set A had a counterpart in set B for tween a reviewer’s propensity to take a brand selfie and the star rat- whom the true perfume was the filler option of set A. With this de- ing he gives the restaurant. sign, an intrinsic preference or any other difference between the true Studies 2a to 2c establish causality. Study 2a tests brand selfie perfume and the filler that might have caused an accurate face-brand taking vs taking no photo whatsoever on purchase intention and matching in set A is controlled for by having set B. In particular, studies 2b and 2c test whether it is specifically brand selfie-taking, pooling the accuracy rates of the two corresponding faces in sets and not simply taking a selfie (without the brand), that affects brand A and B effectively controls for any bias linked to the perfume op- choice. In all three studies we find a positive and significant effect of tions themselves. Headshots were also paired to minimize the age brand selfie-taking over the control on brand preference. difference. Across the two sets of headshots, participants guessed Study 3 tests the specificity of the effect and rules out the pos- the true perfumes above chance, controlling for socioeconomic, age sibility that the effect is driven by the higher exposure to the brand and preference factors, as well as familiarity and release dates of the in the brand selfie-condition. Therefore, in study 3, we added an ad- perfumes. ditional control condition, in which participants were instructed to Study 3 provided an identical and conceptual replication of simply take a photo of the brand (without the self). Participants were study 2 on a total of 30 headshots. Participants first saw the 20 assigned to one of three photo-taking tasks: selfie with the brand headshots of study 2 in two different orders. They then saw 10 ex- (brand selfie), a selfie without the brand (selfie) or a photo of the tra headshots. Across the two orders, the face-brand matching effect brand in front of a neutral background (brand photo). We still find, held once again. that purchase intention is higher for the brand selfie-takers relative to In summary, our studies repeatedly support the existence of a the mere selfie-takers and to the mere brand photo-takers. And that face-brand matching effect, whereby social perceivers can accurately the effect is fully mediated by self-brand connection (indirect effect match a person’s actual perfume brand to a headshot of their face, = .35, SE = .13, 95% CI [.10, .62]). above chance level. Critically for marketing, a contribution of this Study 4 was preregistered and designed in order to provide pro- research is to show that our faces alone can evoke a brand user ste- cess evidence via serial mediation. Participants were instructed to reotype. On a practical level, facial recognition systems may be able take either a brand selfie or a selfie, and to then report their intention to push certain brands as a result of an estimated good match with to purchase that brand. Critically however, in addition to assessing targeted consumers’ faces in the not so distant future. self-brand connection as we did in studies 2a, 2c and 3, we also assessed participants’ propensity to look to their selfie-taking behavior to make self-inferences. In line with our conceptual framework we find that self-inferences and self-brand connection serially mediated 1082 / Self-Brand Connection in the Age of Technology the relationship between brand selfie-taking and purchase intention of “putting something together”) and negative metaphors associ- (indirect effect = .402, BootSE = .111, 95% CI = [.197, .634]). ated with ending a romantic relationship (i.e., the behavioral act of In study 5, instead of measuring self-inferences as in study 4, “breaking something”). We hypothesize that relationship metaphors’ we sought to manipulate them, by varying participants’ facial ex- effect on consumer–brand relationships is contingent on the (non-) pressions in their selfies. Specifically, participants were instructed to occurrence of a brand transgression, in that positive (negative) rela- either have a negative expression on their face (e.g., frown), a posi- tionship metaphors affect the relationship only in the absence (pres- tive expression (e.g., smile), or, in a control condition, were given ence) of a transgression. Furthermore, we hypothesize that the inclu- no instruction on how to appear. This manipulation was crossed with sion of a brand in the self mediates the relationship. our standard manipulation of selfie type (brand selfie vs. selfie). We predicted and found an interaction such that the facial expression Study 1 manipulation affects purchase intention only among those in the We relied on archival data by scraping 1,000 reviews from the brand selfie condition, and not among those assigned to simply take best-selling products on Amazon.com. Of these reviews, 418 were a selfie (without the brand). Particularly, self-brand connection me- classified as containing metaphorical speech, in which consumers diates the effect of brand selfies on purchase intention for those in frequently utilized positive metaphors such as “[this brand] and I the positive expression condition (conditioned indirect effect = .550, make a better couple than me and my boyfriend” and “I could marry BootSE = .131, 95%CI = [.296, .801]) and for those in the control [this brand],” as well as negative metaphors such “I am breaking up condition (conditioned indirect effect = .191, BootSE = .075, 95% CI with [this brand].” Our data show that consumers actively and ex- = [.045, .336]) but not for those in the negative expression condition tensively use metaphors when referring to their brand-relationships. (conditioned indirect effect = -.168, BootSE = .111, 95% CI = [-.385, .053]; index of moderated mediation = .359 BootSE = .096, 95% CI Study 2 = [.170, .549]). Study 2 employed a 3 (metaphor: positive, negative, absent) × Finally, study 6 tests whether satisfaction with the selfie mod- 2 (brand transgression: present, absent) experimental design, with erates its effect on self-brand connection. Participants again took a metaphor and brand transgression as between-subjects independent brand selfie or a selfie, and in addition to assessing self-brand con- variables and brand relationship strength as dependent variable. A nection and purchase intention we also assessed participants’ satis- new brand called snapholio, a virtual portfolio app, was developed. faction with the selfie. We find that the effect of brand selfies on After giving informed consent, 196 participants were randomly as- self-brand connection is moderated by the level of selfie satisfaction, signed to one of the six conditions. Participants in the brand trans- i.e., the positive effect of brand selfie-taking on brand preference is gression present conditions had to endure brand-transgressive events attenuated by dissatisfaction with one’s selfie (index of moderated involving severe failures of the app. Participants in the negative mediation = .16, BootSE =.075, 95% CI = [.011, .305]). metaphor condition were shown images of broken things and asked By testing how a new form of user-generated content affects to write about times when they had broken something. Participants self-brand connection formation we contribute to the literature on in the positive condition were asked to complete a series of simple brand relationship formation and to the literature on consumer pho- puzzle tasks. Participants in the control condition were shown a neu- to-taking experiences, and branded user-generated content. tral image and asked to write about their day (Larson and Billetera With respect to practice, our investigation offers guidance to 2013; Marin, Reimann, and Castaño 2014). To measure brand re- marketers on when and why encouraging consumers to take brand lationship strength, we employed an established scale (Aaker et al. selfies may augment or detract from attempts to foster a consumer- 2004), which was measured five times (T1-T5). Inclusion of brand brand connection. Our findings imply that marketers may be overin self was measured employing an established brand inclusion Venn concerned with the appearance of the brand in the photo, and under- diagram (Reimann et al. 2012). Data were submitted to a two-way concerned with helping the consumer to feel satisfied with her own repeated-measures ANOVA. Results revealed a significant interac- appearance. tion effect between metaphor and brand transgression on brand relationship strength, F(2, 185) = 7.27, p < .01, ηp2 = .073. Direct When and Why Metaphors Affect Consumer–Brand effects of metaphor, F(2, 185) = 20.70, p < .001, ηp2 = .18, and brand Relationships transgression, F(1, 185) = 572.98, p < .001, ηp2 = .76 were also significant. At T1, all conditions were the same (baseline). At T2-T4, EXTENDED ABSTRACT it was systematically found that the effect of relationship metaphors In the interaction between brands and consumers, metaphors on consumer–brand relationships is contingent on the (non-)occur- with relationship-related meanings often appear (e.g., in slogans, rence of a brand transgression, in that positive (negative) relation- brand names, logos). Moreover, metaphors seem to be increasing- ship metaphors affect the relationship only in the absence (presence) ly used by consumers themselves (Reimann, Nuñez, and Castaño of a transgression (all p’s of contrasts < .05). To test whether our 2017). hypothesized effects were short-lived or long-lasting, we collected However, despite their common usage, it is unclear how exactly additional measures (T5) four days after having the app was no relationship-related metaphors affect consumer–brand relationships. longer required, revealing that metaphors have a lasting effect on Indeed, the extant literatures in psychology, linguistics, and con- consumers’ brand relationships. Furthermore, in the presence of a sumer research are inconclusive regarding whether metaphors affect brand transgression, only 13% of participants in the negative meta- consumer–brand relationships. Classic views insist that metaphors phor condition chose to share the brand (vs. 41%) p < .05. In the represent the one universal way in which both language (Hawkes absence of a brand transgression, 42% of participants in the positive 1972; MacCormac 1985) and marketing (Bremer and Moonkyu metaphor condition shared the brand (vs. 31%), p < .05. We also as- 1997; Fournier 1998) work. However, recent investigations imply sessed whether or not participants had kept the app on their phones. that the effectiveness of metaphors is uncertain (Landau, Meier, In the presence of a brand transgression, only 9% of participants in and Keefer 2010). We focus on positive metaphors associated with the negative metaphor condition kept it (vs. 31%) p < .05. In the starting/maintaining a romantic relationship (i.e., the behavioral act absence of a brand transgression, 73% of participants in the positive Advances in Consumer Research (Volume 48) / 1083 metaphor condition kept it (vs. 47%) p < .05. Following mediation- was framed very special to them personally because they have been testing guidelines by Hayes and Preacher (2014), we constructed using it for a long time since they received it from their grandparents, three regression models that support our hypothesis that the inclu- and because it fits their personal values very well. In control condi- sion of brand in self mediates the relationship. tion, the scenario did not reference SBC but mentioned they like the brand very much. We measured participants’ psychological threat Study 3 when Rondo is used by others. As predicted, participants who have The goal of Study 3 was to test the effect of metaphors when a high SBC (vs. control) felt their bond with the brand is threatened consumers actively write (vs. are merely exposed to) metaphorical (Msbc=2.71 vs. Mcontrol=2.15, p=.042) and their sense of identity was reviews. We employed the same measures as in Study 2. We present- endangered (Msbc=2.79 vs. Mcontrol=2.06, p=.008). Importantly, there ed participants with a positive or negative hypothetical brand inter- was no effect of uniqueness seeking (p=.18), ruling out the alterna- action (Park and John 2018). Participants in the negative (positive) tive explanation that people feel threatened because usage by others metaphor condition were instructed to write a metaphorical review may take away their uniqueness. of the brand using words like break, breaking, and broken (match, Experiment 2 extended experiment 1 to real brands and exam- matching, and together). Data were submitted to an analysis of vari- ined whether consumers’ perceived psychological threat translates ance. Results revealed an interaction effect between metaphor and into negative behavioral intentions. Students (N=358) wrote either brand transgression on brand relationship strength, F(1, 178) = 2.36, about a brand which they like so much that they do not want to share 2 p < .09, ηp = .026. The direct effects of both metaphor, F(1, 178) = with others (not share) or a brand they like so much (control). In 2 6.59, p < .01, ηp = .071, and brand transgression, F(1, 178) = 34.68, addition to perceived psychological threat items (4 items, α=.903), 2 p < .001, ηp = .17, were significant. The results suggest that the ef- we measured consumers’ intention to say something negative about fect of metaphors on consumer–brand relationships holds regardless the brand just to make sure others do not try it. Participants in the of the mode of exposure. not share condition exhibited greater psychological threat (M=2.25) compared to the control condition respondents (M=1.56; p<.001). General discussion Additionally, participants showed greater intention to spread nega- In summary, across different methods (i.e., lab, longitudinal tive WOM when they recalled the brand they did not want to share field experiment, archival study) and different dependent variables (Mnot share=1.82 vs. Mcontrol=1.48, p=.008). Mediation results revealed (i.e., psychometric measures, actual behaviors, written accounts participants experienced greater psychological threat from sharing from online reviews), the present research not only demonstrates the brand which they like so much that they do not want to share ver- that relationship metaphors depend on the (non-)occurrence of brand sus the brand they simply like, which in turn led to greater negative transgression in order to influence consumer–brand relationships but WOM intentions (B=.36; 95% CI: .22, .52). also provides insight into the mechanism underlying this effect. Our Experiment 3 extended our findings beyond physical posses- empirical investigations illustrate that the inclusion of the brand in sions and examined feelings of happiness associated with sharing. the consumers’ self represents a key explanation for the role of meta- Online participants (N=204) read a scenario about a café named phors in consumer–brand relationships. Janko’s that they either perceive high connection with or they enjoy visiting. Strong SBC consumers were marginally less likely to feel Hands off My Brand: Strong Self-Brand Connection and happy if others visit their café (M=5.88) compared to participants in Psychological Threat the “enjoy visiting the café” condition (M=6.22; p=.055). Interest-

ingly, participants felt less happy if their acquaintances (Msbc=5.66 EXTENDED ABSTRACT vs. M =5.99, p=.074) or random strangers (M =4.95 vs. M The dominant paradigm in the consumer-brand relationships re- control sbc control=5.49, p=.020) started going to Janko’s, but it did not bother them search stream suggests that when consumers love the brand, they will if their family (p=.83) or friends (p=.74) started going there. This do everything in their power to influence others to consider or buy suggests interpersonal proximity plays a potential role. the brand and happily spread positive word-of-mouth (WOM) about We contribute to the literature on consumer-brand relation- it (Batra et al. 2012; Fournier 1998; Park et al. 2010). The current ships (Fournier 1998; Escalas 2004; Khamitov et al. 2019). While work proposes that this may not always be the case. Specifically, we the predominant paradigm within this research stream suggests that pose that when some consumers exhibit elevated self-brand connec- consumers who have a strong brand relationship would go to great tion (SBC), they would instead be more likely to willingly engage lengths to support their favored brands and thus should be encour- in negative WOM and feel unhappy about others trying the brand. aged to become public brand advocates, our findings imply that high We theorize that this “brand for me and not for you” effect should be SBC individuals are likely reluctant to spread positive WOM to the driven by perceived psychological threat generated by the prospect extent that the prospect of other consumers laying hands on their of other consumers using the ‘coveted’ brand. That is, building on brand comes across as threatening. While the extant literature tends the self-expansion theoretical account (Aron et al. 1992; Reimann to view high SBC levels as an invaluable asset (Escalas 2004; Esca- and Aron 2009), we anticipate that when highly self-brand connected las and Bettman 2003, 2005), the present research underscores an in- consumers engage in self-expansion of their self-concept, once they stance where typical benefits of high SBC are unlikely to accrue and deeply incorporate the brand into the sense of self, the prospect of instead may reverse. Last but not the least, our work adds a nuance other consumers trying the brand should result in a sense of psycho- to the brand community research stream (McAlexander et al. 2002; logical threat. This should in turn make such consumers more willing Muniz and O’Guinn 2001) by providing a conceptual and empirical to engage in negative WOM, make them feel less happy, and prefer account of what can be construed as the opposite notion of brand to selfishly keep the brand for themselves. community where instead of a shared brand-enabled bond consum- Experiment 1 assessed consumers’ reactions to others using the ers appear to harbor a rather selfish and egoistic SBC they want to brand which they got primed to perceive high connection with. On- keep for themselves. line participants (N=206) were asked to imagine that they have been Prior WOM research established that when consumers feel fa- using Rondo, a diary notebook brand. In high SBC condition, Rondo vorably toward a product or a service, a number of them exhibit a 1084 / Self-Brand Connection in the Age of Technology tendency to leave positive reviews, encourage their friends and ac- McCracken, Grant (1986), “Culture and Consumption: A quaintances to give it a try, and put in a good word (Berger 2014; Theoretical Account of the Structure and Movement of the Lovett et al. 2013; Moore and Lafreniere 2020). Mapping our results Cultural Meaning of Consumer Goods,” Journal of Consumer on this literature, the findings nuance some of the WOM work by Research, 13(June), 71–84. providing a type of boundary condition wherein instead of engag- ——— (1989), “Who Is the Celebrity Endorser? Cultural ing in increased WOM, consumers with strongly favorable product Foundations of the Endorsement Process,” Journal of and service opinions are actually engaging in more negative WOM. Consumer Research, 16(3), 310–21. Lastly, we contribute to the related consumer research on secrecy Statista (2017), “Vous Personnellement, Portez-Vous du Parfum? (Rodas and Roedder John 2020; Thomas and Jewell 2019) that has [Do You Personally Wear Perfume?],” Retrieved from https:// largely focused on the desire to maintain secrecy before, during, or fr.statista.com/statistiques/708172/utilisation-parfum-france/ after consumption. To this end, our findings reinforce the value and Tanaka, James W. (2001), “The Entry Point of Face Recognition: importance of studying phenomena wherein consumers would rather Evidence for Face Expertise,” Journal of Experimental keep their consumption objects private and for themselves. Psychology: General, 130(3), 534–543. Zwebner, Yonat, Anne-Laure Sellier, Nir Rosenfeld, Jacob REFERENCES Goldenberg and Ruth Mayo (2017), “We Look Like our Names: The Manifestation of Name Stereotypes in Facial Session Overview Appearance,” Journal of Personality and Social Psychology, Belk, Russell W. (1988), “Possessions and the Extended Self,” 112(4), 527–554. Journal of Consumer Research, 15(2), 139–168. Escalas, Jennifer E. and James Bettman (2003), “You Are what From Sweetheart to Scapegoat: Brand Selfie-taking They Eat: The Influence of Reference Groups on Consumers’ Shapes Consumer Behavior Connections to Brands,” Journal of Consumer Psychology, Bem, Daryl J. (1967), “Self-Perception: An Alternative 13(3), 339–348. Interpretation of Cognitive Dissonance Phenomena,” Psychological Review, 74(3), 183-200. We Can Look Like Our Brands: The Manifestation of Chaiken, Shelly and Mark W. Baldwin (1981), “Affective- Adopted Brands in Facial Appearance Cognitive Consistency and the Effect of Salient Behavioral Aaker, Jennifer L. (1997), “Dimensions of Brand Personality,” Information on the Self-Perception of Attitudes,” Journal of Journal of Marketing Research, 34(3), 347–356. Personality and Social Psychology, 41(1), 1-12. Ballew, Charles C. and Alexander Todorov (2007), “Predicting Fazio, Russell H., Mark P. Zanna, and Joel Cooper (1977), Political Elections from Rapid and Unreflective Face “Dissonance and Self-Perception: An Integrative View of Judgments,” Proceedings of the National Academy of Sciences Each Theory’s Proper Domain of Application,” Journal of of the United States of America, 104(46), 17948–17953. Experimental Social Psychology, 13(5), 464-79. Belk, Russell W. (1988), “Possessions and the Extended Self,” Festinger, Leon (1962), A Theory of Cognitive Dissonance, Journal of Consumer Research, 15(2), 139–168. Stanford, CA: Stanford University Press. Belk, Russell W., Kenneth D. Bahn, and Robert N. Mayer (1982), Grewal, Lauren, Jillian Hmurovic, Cait Lamberton, and Rebecca “Developmental Recognition of Consumption Symbolism,” Walker Reczek (2019), “The Self-Perception Connection: Journal of Consumer Research, 9(1), 4–17. Why Consumers Devalue Unattractive Produce,” Journal of Berger, Jonas and Chip Heath (2007), “Where Consumers Diverge Marketing, 83(1), 89-107. from Others: Identity Signaling and Product Domains,” Heider, Fritz (1958), “The Naive Analysis of Action,” in The Journal of Consumer Research, 34(2), 121–134. Psychology of Interpersonal Relations, Hoboken, NJ: Wiley, Escalas, Jennifer E. and James Bettman (2003), “You Are what 79–124. They Eat: The Influence of Reference Groups on Consumers’ Park, C. Whan, Deborah J. MacInnis, Joseph Priester, Andreas B. Connections to Brands,” Journal of Consumer Psychology, Eisingerich, and Dawn Iacobucci (2010), “Brand Attachment 13(3), 339–348. and Brand Attitude Strength: Conceptual and Empirical ______and ______(2019), “Connecting With Celebrities: How Differentiation of Two Critical Brand Equity Drivers,” Journal Consumers Appropriate Celebrity Meanings for a Sense of of Marketing, 74(6), 1-17. Belonging,” Journal of Advertising, 46(2), 297-308. Summers, Christopher A., Robert W. Smith, and Rebecca Walker Fournier, Susan (1998), “Consumers and Their Brands: Developing Reczek (2016), “An Audience of One: Behaviorally Targeted Relationship Theory in Consumer Research,” Journal of Ads as Implied Social Labels,” Journal of Consumer Research, Consumer Research, 24(March), 343-373. 43(1), 156-78. Gordon, Iris and James W. Tanaka (2011), “Putting a Name to a Xu, Jing, Hao Shen, and Robert S. Wyer Jr. (2012), “Does the Face: The Role of Name Labels in the Formation of Face Distance Between us Matter? Influences of Physical Proximity Memories,” Journal of Cognitive Neuroscience, 23(11), to Others on Consumer Choice,” Journal of Consumer 3280–3293. Psychology, 22(3), 418-23. Levy, Sidney J. (1959), “Symbols for Sale,” Harvard Business Review, 37 (4), 117–24. When and Why Metaphors Affect Consumer–Brand Malär, Lucia, Harley Krohmer, Wayne D. Hoyer, and Bettina Relationships Nyffenegger (2011), “Emotional Brand Attachment and Brand Aaker, Jennifer, Susan Fournier, and Adam Brasel (2004), “When Personality: The Relative Importance of the Actual and the Good Brands Do Bad,” Journal of Consumer Research, 31(1), Ideal Self,” Journal of Marketing, 75(July), 35-52. 1–16. Advances in Consumer Research (Volume 48) / 1085

Bremer, Kristine and Lee Moonkyu (1997), “Metaphors in Escalas, Jennifer E. (2004), “Narrative Processing: Building Marketing: Review and Implications For Marketers,” Consumer Connections to Brands,” Journal of Consumer Advances in Consumer Research Volume, 24(1), 419–424. Psychology, 14(1-2), 168-180. Fournier, Susan (1998), “Consumers and Their Brands: Developing Escalas, Jennifer E. and James R. Bettman (2003), “You Are what Relationship Theory in Consumer Research,” Journal of They Eat: The Influence of Reference Groups on Consumers’ Consumer Research, 24(4), 343–53. Connections to Brands,” Journal of Consumer Psychology, Hawkes, Terence (1972), “Metaphor,” London: Methuen and Co. 13(3), 339-348. Hayes, Andrew F. and Kristopher J. Preacher (2014), “Statistical Escalas, Jennifer E. and James R. Bettman (2005), “Self-Construal, Mediation Analysis with a Multicategorical Independent Reference Groups, and Brand Meaning,” Journal of Consumer Variable,” British Journal of Mathematical and Statistical Research, 32(3), 378-389. Psychology, 67(3), 451–470. Fournier, Susan (1998), “Consumers and Their Brands: Developing Landau, Mark J., Brian P. Meier, and Lucas A. Keefer (2010), Relationship Theory in Consumer Research,” Journal of “A Metaphor-Enriched Social Cognition.,” Psychological Consumer Research, 24(4), 343-373. Bulletin, 136(6), 1045–1067. Khamitov, Mansur, Xin (Shane) Wang, and Matthew Thomson Larson, Jeffrey S. and Darron M. Billetera (2013), “Consumer (2019), “How Well Do Consumer-Brand Relationships Drive Behavior in ‘Equilibrium’: How Experiencing Physical Customer Brand Loyalty? Generalizations from a Meta- Balance Increases Compromise Choice,” Journal of Marketing Analysis of Brand Relationship Elasticities,” Journal of Research, 50(4), 535–547, Consumer Research, 46(3), 435-459. MacCormac, Earl (1985), “A Cognitive Theory of Metaphor,” Lovett, Mitchell J., Renana Peres, and Ron Shachar (2013), “On Cambridge, MA: MIT Press. Brands and Word of Mouth,” Journal of Marketing Research, Marin, Alex, Martin Reimann, and Raquel Castaño (2014), 50(4), 427-444. “Metaphors and Creativity: Direct, Moderating, and Mediating McAlexander, James H., John W. Schouten, and Harold F. Koenig Effects,” Journal of Consumer Psychology, 24(2), 290–297. (2002), “Building Brand Community,” Journal of Marketing, Park, Ji Kyung and Deborah Roedder John (2018), “Developing 66(1), 38-54. Brand Relationships after a Brand Transgression: The Role of Moore, Sarah G. and Katherine C. Lafreniere (2020), “How Online Implicit Theories of Relationships,” Journal of the Association Word‐of‐Mouth Impacts Receivers,” Consumer Psychology for Consumer Research, 3(2), 175–187. Review, 3(1), 34-59. Reimann, Martin, Raquel Castaño, Judith Zaichkowsky, and Muniz, Albert M. and Thomas C. O’guinn (2001), “Brand Antoine Bechara (2012), “How We Relate to Brands: Community,” Journal of Consumer Research, 27(4), 412-432. Psychological and Neurophysiological Insights into Park, C. Whan, Deborah J. MacInnis, Joseph Priester, Andreas B. Consumer-Brand Relationships,” Journal of Consumer Eisingerich, and Dawn Iacobucci (2010), “Brand Attachment Psychology, 22(1), 128–142. and Brand Attitude Strength: Conceptual and Empirical Reimann, Martin, Sandra Nuñez, and Raquel Castaño (2017), Differentiation of Two Critical Brand Equity Drivers,” Journal “Brand-Aid,” Journal of Consumer Research, 44(3), 673–691. of Marketing, 74(6), 1-17. Reimann, Martin and Arthur Aron (2009), “Self-Expansion Hands off My Brand: Strong Self-Brand Connection and Motivation and Inclusion of Brands in Self: Toward a Theory Psychological Threat of Brand Relationships,” in Handbook of Brand Relationships, Aron, Arthur, Elaine N. Aron, and Danny Smollan (1992), ed. Deborah J. MacInnis, C. Whan Park, and Joseph R. “Inclusion of Other in the Self Scale and the Structure of Priester (pp. 87-103), Armonk, NY: M. E. Sharpe. Interpersonal Closeness,” Journal of Personality and Social Rodas, Maria A. and Deborah Roedder John (2020), “The Secrecy Psychology, 63(4), 596-612. Effect: Secret Consumption Increases Women’s Product Batra, Rajeev, Aaron Ahuvia, and Richard P. Bagozzi (2012), Evaluations and Choice,” Journal of Consumer Research, “Brand Love,” Journal of Marketing, 76(2), 1-16. forthcoming. Berger, Jonah (2014), “Word of Mouth and Interpersonal Thomas, Veronica L. and Robert D. Jewell (2019), “I Can’t Get Communication: A review and Directions for Future You Out of My Head: The Influence of Secrecy on Consumers’ Research,” Journal of Consumer Psychology, 24(4), 586-607. Self‐Brand Connections,” Journal of Consumer Psychology, 29(3), 463-471. Consumers and Politics: When Companies Take a Stance Chair: Rhia Catapano, University of Toronto, Canada

Paper #1: Political Neutrality Aversion: When and Why ‘Staying threat to culture/values), product status, and actor-product fit mediate Out of It’ Backfires this effect. Ike Silver, University of Pennsylvania, USA Finally, Paper 4 (Catapano and Tormala) explores how compa- Alex Shaw, University of Chicago, USA nies should frame their stances in order to maximize impact. They find that consumers believe that expressing attitudes in terms of sup- Paper #2: Chasing Political Review Storms port, rather than opposition, is more expressive of their values, and Johannes Boegershausen, University of Amsterdam, The will lead to more liking from others. Thus, when brands talk about Netherlands the policies they support, rather than policies they oppose, consum- Jared Watson, New York University, USA ers are more likely to propagate these messages through sharing. Paper #3: The Changing Face of America: When Majority Together, these papers provide multiple perspectives and en- Group Consumers Prefer vs . Are Threatened By Ethnic Diversity hance our understanding of when, why, and how, companies should in Advertising weigh in on political issues in order to maximize benefits to their Steven Shepherd, Oklahoma State University, USA brand. Whereas limited research to date has explored the emerging Tanya L. Chartrand, Duke University, USA trend of companies weighing in on political issues, the work pre- Gavan J. Fitzsimons, Duke University, USA sented in this session begins to address this gap, enhancing our un- Aaron C. Kay, Duke University, USA derstanding of consumers in a changing world. Paper #4: Support or Oppose? The Effects of Political Attitude Framing on Sharing Behavior Political Neutrality Aversion: When and Why ‘Staying Rhia Catapano, University of Toronto, Canada Out of It’ Backfires Zakary Tormala, Stanford University, USA EXTENDED ABSTRACT SESSION OVERVIEW Business leaders, celebrities, and politicians wield significant More and more, firms are taking public stances on political and influence over many facets of life, from the laws that pass tothe social issues. In the past few years, Nike created an ad campaign products and ideas that spread. However, public figures derive much centered around Colin Kaepernick, Dick’s Sporting Goods and Delta of their power from public opinion and consumer support. How do publicly severed ties with the NRA, and dozens of CEOs from major consumers decide whom to trust, buy from, or vote for? companies made public statements opposing the 2017 immigration One important consideration is a public figure’s commitment ban. Although firms have always been involved in politics, ithas to social causes. Indeed, consumers increasingly care where public traditionally been behind the scenes (Lawton McGuire, and Rajwani figures and the organizations they represent stand on hot-button is- 2013), rather than in public. Moreover, there is little precedent for sues like abortion, climate change, and immigration, and they will firms taking stances on political and social issues that are notdi- boycott, condemn, and punish actors who take positions they deem rectly relevant to the firm’s value proposition. For this reason, little unacceptable (Haidt 2012; Skikta 2010). Perhaps as a result, public is known about how taking a stance impacts consumer behavior. Do figures frequently choose to remain neutral when asked to weigh in firms benefit from taking public political stances? What happens on such issues, citing the merits of each side or expressing a prefer- when firms do choose to take a stance? And, when they do choose ence not to get involved at all (Shaw et al. 2017). to take a stance, how can they maximize the impact on consumers? This paper explores how consumers react to political neutrality. Our session sheds light on these questions. Paper 1 (Silver and We propose that despite its intuitive appeal, remaining neutral often Alex Shaw) explores why it may be beneficial for firms to take politi- resembles a ‘false signal’ (Jordan and Bloom 2017), which can un- cal stances, even at the risk of alienating some consumers. They find dermine trust and erode support. that despite its intuitive appeal, remaining neutral can harm trust, When public figures express political neutrality, consumers fre- even relative to outright opposition. This effect occurs because con- quently have information about their intended audience (e.g., view- sumers interpret neutrality as concealed opposition and therefore as ers of a particular media outlet, constituents from a particular politi- a dishonest impression management tactic. cal party). We hypothesize that this audience information can shift The next two papers explore what happens when companies do how neutrality is interpreted in predictable ways. Specifically, we take a stance. Paper 2 (Boegershausen and Watson) focuses on what predict that remaining neutral in front of a more liberal audience will happens to reviews after a brand takes a political stance. They find signal underlying conservative convictions and vice versa. that the resulting political review storms include reviews from sup- Moreover, to the extent that expressing neutrality over a di- porters and opposers, leading to a set of reviews that provide very lit- visive issue seems like a form of strategic non-disclosure (John, tle information diagnostic of product quality. The authors investigate Barasz, and Norton 2016), neutral public figures may seem insincere, the downstream consequences for the brand involved, exploring the dishonest, and untrustworthy. Indeed, remaining neutral (and signal- differential effects on liberals and conservatives. Paper 3 (Shepherd, ing concealed opposition) may harm trust even relative to opposing Chartrand, Fitzsimons, and Kay) focuses on the impact of political a consumers’ viewpoint directly, a phenomenon we term political ideology on how consumers respond to advertising that implicitly neutrality aversion. represents a political stance. Specifically, they find that conserva- Study 1 (n=301) sought to demonstrate that audience informative (liberal) ideology is associated with more negative (positive) tion predictably shapes how expressions of neutrality are interpreted. attitudes and behavioral intentions when ads feature ethnic diver- Participants read one of three stimulus-sampled vignettes which each sity/non-White actors. Symbolic threat (seeing the ethnic group as a described a public figure expressing neutrality in a public forum. Be- tween-subjects, we manipulated whether the actor was speaking to

Advances in Consumer Research 1086 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1087 a predominantly liberal or a predominantly conservative audience. with the majority of their peers on a divisive political issue. When Participants were then asked to make attributions about the neutral asked which approach would effectively win trust and support from actor’s personal beliefs on scales from -3 (holds strongly liberal be- peers, the majority chose ‘remain neutral’ (66.8%), more than ‘side- liefs) to +3 (holds strongly conservative beliefs). As predicted, we against’ (24.2%) and ‘side-with’ combined (9.0%; p<.001). found a main effect of audience (F(1,295)=111.14,p <.001), such that Together, these studies illuminate the surprising costs of remain- participants believed an identical expression of neutrality to be con- ing neutral, and they reveal an attribution process which incentivizes cealing conservative beliefs when addressing a liberal audience, and side-taking and may ultimately instigate polarization. Furthermore, liberal beliefs when addressing a conservative audience. This effect actors who do choose to remain neutral on important social issues held separately in each vignette-condition (ps<.001). may be inadvertently mismanaging their reputations. Studies 2a-b (n=187, n=300) replicated the belief inference effects from Study 1 with real press video clips of public figures Chasing Political Review Storms (A member of the Backstreet Boys, the owner of the Kansas City Chiefs) expressing neutrality in live answers to reporters’ questions EXTENDED ABSTRACT about divisive political issues (support for President Trump, NFL Today, brands who take a political stance may receive reviews players kneeling during the national anthem). We again manipulated from supporters and opposers, resulting in a political review storm the actor’s audience by describing the reporter in each clip as being that provides the consumer very little information diagnostic of from either a conservative or a liberal news outlet. Once again, par- product quality. For example, a bakery in Colorado refused to bake a ticipants saw neutrality as signaling conservative beliefs when ad- wedding cake for a same-sex couple resulting in a flood of 1-star re- dressing a liberal audience and vice versa (2a: t(185)=1.91, p=.058; views from liberals and 5-star reviews from conservatives (Gajanan 2b: t(298)=6.15, p<.001). Moreover, participants who agreed with 2018). Utilizing the accessibility-diagnosticity framework (Feldman the speaker’s audience (and thus inferred that the neutral actor tacitly and Lynch 1988) and moral foundations theory (Graham, Haidt, opposed their viewpoint) reported that they would find him more and Nosek 2009), we propose that liberals and conservatives will trustworthy if he had explicitly opposed their viewpoint instead of differentially respond to political review storms. While reviews are remaining neutral (3-item trust scale, αs>.75: -3: Much less trust- generally helpful in forming product quality inferences, like other worthy to +3: Much more trustworthy; 2a: t(96)=2.67, p=.009; 2b: attributes (Rao and Monroe 1989), we propose that product qual- t(146)=1.92, p=.057; vs. scale midpoint). ity is not the only attribute that matters amidst a political review Study 3 (pre-registered, n=541) sought to demonstrate that neu- storm. Our theorizing suggests that liberals will attempt to utilize trality harms trust, but only when it resembles strategic concealment, both product quality information and a brand’s political ideology to and not when it signals sincere middle-ground beliefs. Participants form their decisions, whereas conservatives will rely on the brand’s read a scenario in which a prominent businesswoman is asked to political ideology. To reconcile their need for ingroup loyalty, the weigh in on immigration policy at a press conference. In all condi- review voice (i.e., is it written by a liberal or conservative) will lead tions, we described the businesswoman’s audience as holding the conservatives to discount the credibility of reviews written by liber- same views on the issue as the participant. This time, we manipu- als. On the other hand, liberals will recognize the bias in both review lated whether the businesswoman remains neutral (public neutral- voices as they sift through the reviews for diagnostic quality infor- ity condition) or sides against her audience (and the participant) mation. Across three studies, we begin by demonstrating the effect outright (opposition condition). We also included a third condition of political review storms on perceived review credibility, then ulti- in which the businesswoman remains neutral in public at the press mately the consequence of this on brand intentions. conference, but later expresses a neutral viewpoint in private to her Study 1 (N = 327) employed a 2(review voice: liberal, con- spouse (private neutrality condition). As predicted, public neutrality servative) x 2(political ideology: liberal, conservative), between- signaled concealed opposition, but private neutrality did not (F(2, subjects design. Participants were randomly assigned to the reviews 522)=114.44, p<.001). In turn, public neutrality was seen as less factor and we measured their ideology with the item “with which trustworthy than both outright opposition (t(347)=12.09, p<.001) political party do you most closely identify?”. Participants were and private neutrality (t(347)=10.61, p<.001). asked to view the information for a bakery on Yelp. The business Study 4 (pre-registered, n=600) aimed to examine neutrality’s was framed as being conservative and this was revealed through the effects on trust in an incentive-compatible economic game. Partici- voice manipulation. Review voice was manipulated with reviews pants first indicated their views about gun control in a forcedbi- that mentioned claims like “can’t believe they support a conserva- nary choice and shared these beliefs with a partner. We manipulated tive agenda [1-star]” (liberal voice) or “proud to have a local busi- whether the partner (a confederate) then signaled back agreement, ness fight back against corruptive liberal ideologies [5-star]” (con- opposition, or neutrality (a preference not to take sides). Finally, servative voice). When it came to the perceived review credibility, participants rated attitudinal trust (3-item scale, α>.9) and played a 2x2 ANOVA yielded a significant interaction (F(1,323) = 28.709; an incentivized prisoner’s dilemma game. Participants judged neu- p < .001), qualified by main effects of the review voice (F(1,323) trality to be substantially less trustworthy than outright opposition = 11.518; p = .001) and political ideology (F(1,323) = 5.968; p = (t(396)=6.98, p<.001). Additionally, although participants were sig- .015). For a conservative voice (e.g., 5-star reviews), conservatives nificantly more likely to cooperate when their partner signaled ideo- found the reviews to be much more credible than liberals (Mconserva- 2 logical agreement (90% cooperation), they were no more likely to tives = 5.22, Mliberals = 4.11; F(1,323) = 32.306; p < .001; partial η cooperate with a neutral partner (70%) than with one who opposed = .162). Whereas for a liberal voice (e.g., 1-star reviews), liberals them outright (73%; Neutrality vs. Opposition: p>.5). found the reviews to be more credible than conservatives (Mconservatives 2 Finally, Study 5 (n=190) investigated whether participants = 3.97, Mliberals = 4.39; F(1,323) = 4.016; p = .046; partial η = .025). would intuit neutrality’s costs when asked to weigh in on a politi- Importantly, we see a much smaller effect for liberal voices relative cal issue themselves. They largely did not. Participants in a business to conservative voices, suggesting that ideology impacts perceptions school’s behavioral lab were told to imagine that they were running of credibility differentially. for a leadership position in an organization and that they disagreed 1088 / Consumers and Politics: When Companies Take a Stance

In study 2 (N = 621), we sought to generalize our findings same time, companies feel pressure to take a stance on social issues and determine downstream consequences of these political review (Global Strategy Group 2012). storms. Study 1 positioned the brand as “promoting Conservative Drawing from work on ideology (Jost, Glaser, Kruglanski, ideologies” or “opposing Liberal ideologies”. However, we real- and Sulloway 2003; Jost, Nosek, and Gosling 2008) and intergroup ize that these two positions are not equal (Catapano and Tormala, attitudes (Stephan and Stephan 2000), we posit that conservative working paper). To avoid this potential confound, study 2’s voice ideology will be associated with more negative attitudes toward di- manipulation disclosed that the business refused to seat a member versity in advertising, whereas liberals will actually prefer ads with of the Republican [Democratic] congress. Thus, the political review (vs. without) ethnic diversity/non-White actors. We also test relevant voice was always negative (i.e., 1-star). We also included a control moderators of this effect and offer process evidence, including the condition where none of the reviews had a political voice. Political role of symbolic threat (i.e., seeing immigrants as a threat to one’s ideology was assessed as in the prior study. We replicated the previ- culture), product status, and the product’s cultural origins. We argue ous patterns on review credibility as we again found a significant that our results are not explained by felt targetedness (Aaker, Brum- interaction between voice and political ideology (F(2,616) = 6.60; p baugh, and Grier 2000; Grier and Brumbaugh 1999), or similarity to < .01), which qualified the main effect of voice (F(2,616) = 60.45; the target (Aaker et al. 2000). p < .001). The main effect of political ideology was not significant In Study 1 (n = 170, White, US undergraduates), participants (F(1,616) = 2.47; p > .11). For conservative reviews, conservatives were asked to rate their attitudes toward a series of ads that either perceived the review significantly more credible (Mconservatives = 4.65) featured White or Non-White (Sikh or Muslim) actors. Participants than liberals (Mliberals = 3.97; t(207) = 3.75; p < .001). In contrast, completed a measure of symbolic threat (e.g., extent to which immi- there were no significant difference between liberals and conserva- grants undermine or enhance core American values, American iden- tives for liberal reviews (Mconservatives = 4.23, Mliberals = 4.33; t(206) = tity, etc.). After a delay, participants also completed a measure of

.50; p > .62) or non-politicized control reviews (Mconservatives = 5.44, support for the current sociopolitical status quo (system confidence;

Mliberals = 5.54; t(203) = .68; p > .49). Kay and Jost 2003) as a measure of conservative ideology (Jost et Lastly, study 3 employed a 2x2x2 design where we first mea- al., 2003). The two-way interaction was significant; those higher in sured political ideology (conservative or liberal), then manipulated system confidence rated the non-White ads more negatively than the overall quality (2- or 4-star average rating) and (review voice: con- White ads, and the opposite was found for those low in system conservative, liberal). We matched participants with a business that fidence. Higher system confidence scores were also associated with aligned with their ideology, then randomly assigned them to a review lower non-White ad evaluations. System confidence also predicted voice condition that supported their position (5-star) or opposed their symbolic threat, and symbolic threat mediated the effect of system position (1-star). A 2x2x2 ANOVA on brand intentions yielded a sig- confidence on ad attitudes for the ads with non-White actors, but nificant 3-way interaction (F(1,495) = 4.436;p = .036). Due to space not when the ads had White actors (moderated mediation effect was constraints, we will not fully detail the patterns here, but we will significant). summarize them by saying that when the rating was consistent with Study 2 replicated the interactive effect of ad content and ideol- the reviews (e.g., 2-star business with 1-star reviews), liberals indi- ogy on ad attitudes, but with ads for businesses and services ostensi- cated directionally higher intentions than conservatives. However, bly available around campus. Ads either featured all White actors or when there was inconsistency (e.g., 2-star business with 5-star re- a diverse range of ethnicities (including White). Ideology was mea- views), conservatives indicated greater intentions than liberals, sug- sured using the social dominance orientation (SDO) scale (Ho et al. gesting that they will support a business with a Conservative agenda, 2015), which is a conservative ideology (Ho et al., 2015; Jost et al. regardless of brand quality. 2003) where hierarchical differences and inequality between groups Taken together, these studies demonstrate that liberals and con- is seen as desirable. Those high (low) in SDO rated the diversity servatives are both influenced by their political identity; however, ads more negatively (positively) than the no-diversity ads. They also liberals will also use quality as diagnostic information when deter- reported less likelihood of visiting the business. mining whether to engage with a brand. Conservatives, on the other To further test the role of symbolic threat in the relationship hand, appear more concerned with ingroup loyalty than quality con- between ideology and attitudes toward diversity in ads, Study 3 (n = cerns, whereas Liberals place less concern on ingroup loyalty. More 395, White US residents) manipulated the description of an ostensi- research is needed to determine if this stems from some relationship bly real ethnic group (Camarians), such that their culture and values to perceived review credibility or reviewer intentions or if it arises were described as either similar or dissimilar to American culture from a tribe mentality given the perceptions of the political climate and values (adapted from Maio et al. 1994, Stephan et al. 2005). Par- currently. This work is currently limited with its focus on the U.S. ticipants viewed and rated three ads said to feature Camarians. The population, but we hope to expand the investigation to include other interaction between the symbolic threat manipulation and ideology countries to disentangle these potential pathways. (system confidence: low vs. high) was significant. Those high (vs. low) in system confidence reported more negative attitudes toward The Changing Face of America: When Majority Group the ads when symbolic threat was high, but this effect was eliminat- Consumers Prefer vs . Are Threatened By Ethnic ed and ad attitude ratings were significantly higher when symbolic Diversity in Advertising threat was low. Conservative ideology is also associated with valuing vertical EXTENDED ABSTRACT differentiation and status (Kim, Park, and Dubois 2018; Ordabayeva By reflecting the diverse composition of American society, ad- and Fernandes 2018). Given this, and given the relationship between vertising has the potential to say something about society and cul- conservative ideology and attitudes toward immigration and vari- ture (Pollay 1986). However, diversity in ads may take on political ous ethnic groups (Esses, Jackson, and Armstrong 1998; Ho et al. meanings given the recently stoked culture wars regarding ethnic- 2015; Jost et al. 2003), we posited that conservatives would see the ity, immigration, and the call to “make America great again”. At the product in the non-White ad condition as lower status, which would then predict negative ad attitudes. Study 4 supported this. Partici- Advances in Consumer Research (Volume 48) / 1089 pants (n = 243, White American residents) saw an ad for sweet tea statement (Strongly Oppose (0) and Strongly Support (100)), and that featured a White family or Sikh family. Participants rated ad their intentions to share their attitudes on a 3-item index (e.g., “How attitudes, indicated their WTP ($0-$5), and rated the status of the likely would you be to share your views on this topic with friends or product. Finally, they completed the SDO scale. The ad content X family who disagree with you on this issue?”). This index was used ideology interaction was significant. Those higher in SDO responded across studies. Across topics, participants were more likely to share more negatively to the non-White ad (vs. White ad), whereas those their views when they considered statements they supported rather low in SDO showed the reverse effect. Similar effects were observed than statements they opposed, t(2172.02) = 17.00, p < .001. This on WTP and perceived status. Moderated mediation analysis found effect was consistent across statements, and we found no interaction that those higher in SDO saw the product as lower in status in the between attitude framing and attitude valence (p = .96). non-White (vs. White) ad condition, which predicted more negative Study 2 aimed to provide causal evidence for the attitude- ad attitudes and decreased WTP. framing effect in a more controlled setting. Participants in the lab Finally, we reasoned that if conservative ideology is associated were assigned to either a support-framing or oppose-framing condi- with seeing various ethnic groups as a symbolic threat and main- tion. We began by asking participants to select which statement best taining boundaries between groups, then it follows that those higher characterized their attitude towards gun control: “More gun control in conservative ideology will particularly dislike ethnic diversity would be (helpful/harmful).” In the support-framing condition, par- in ads for products that are associated with American culture com- ticipants were asked about the statement they supported (the one pared to those that are not. Study 6 (n = 362 White American resi- they selected), whereas in the oppose-framing condition, participants dents) tested this. Participants saw the same ad with a Sikh family were asked about the statement they opposed (the one they did not from Study 4, except the ad was either for Southern Breeze sweet select). Thus, in both conditions, participants were asked about an tea (American brand and product), or Southern Breeze bael fruit tea attitude position that was consistent with their own view. For ex- (American brand, Indian product). Actor-product fit (Pounders and ample, a participant who indicated that more gun control would be Mabry-Flynn 2016) was also measured (e.g., “choosing this fam- beneficial would be randomly assigned to answer questions about ily for this product is appropriate”). High SDO was associated with either their support for the beneficial statement or their opposition to seeing less actor-product fit particularly for the domestic product the harmful statement. Then, participants were then asked how likely (vs. foreign product), and also liking the ad for the domestic product they were to share their attitudes on the statement with another par- less (moderated mediation effect was significant). Thus, despite the ticipant currently completing the study. Participants reported greater foreign-origin product being unfamiliar to participants and not be- sharing intentions in the support-framing than in the oppose-framing ing targeted by the ad, the high SDO participants liked the ad for condition, t(206) = 2.13, p = .033. the foreign-origin product more. Although counterintuitive, this is Study 3 aimed to provide evidence for the proposed value ex- consistent with conservative ideology as a hierarchy- and status quo- pression and impression management mechanisms. We manipulated maintaining ideology. attitude-framing as described in Study 2. After indicating their likeli- Even though the US has long been a diverse society, ethnic di- hood of sharing, participants were asked to imagine that they had versity in ads is not ideologically neutral. We provide evidence for decided to share their view, and reported their prediction for the mes- the role of ideology in shaping attitudes toward culturally diverse sage recipient’s impression of them (2 items; e.g., “How positively ads. Moreover, we find evidence for theoretically-relevant modera- do you think the message recipient would view you as a person”) tors and mediators that also have marketing implications. and how value expressive sharing this position would be (2 items, e.g., “How representative of your values is your attitude [in favor Support or Oppose? The Effects of Attitude Framing on of/against] this statement?”). Participants reported greater sharing Sharing Behavior intentions when the issue was framed in terms that they supported relative to terms that they opposed, t(823) = 4.68, p < .001). In addi- EXTENDED ABSTRACT tion, support framing led to greater value expressiveness, Z = 9.40, p What factors affect whether consumers share their views with < .001, which in turn predicted greater sharing intentions, Z = 4.69, p others? We identify a novel determinant of whether consumers < .001; indirect effect= .12, Z = 4.28, p < .001. Support framing also share—attitude framing, defined as whether consumers think of their led to higher positive impression ratings, Z = 2.53, p = .01, which own attitude in terms of what they support or what they oppose. At- predicted greater sharing intentions, Z = 7.64, p < .001; indirect ef- titude framing is distinct from attitude valence, as the same attitude fect = .05, Z = 2.35, p = .02. can be framed in terms of support (e.g., I support that this policy is In Study 4, we replicated this effect, finding that the effect holds bad) or opposition (e.g., I oppose that this policy is good). regardless of whether the target of sharing agrees or disagrees on a Six preregistered experiments provide evidence for an attitude- given topic. In Study 5, we replicated this effect with another topic, framing effect, whereby consumers are more likely to share, or and showed that the effect is attenuated when individuals do not want express, attitudes framed in terms of positions they support rather to be liked. than positions they oppose. This effect occurs due to two interaction How can companies leverage this knowledge? More and more, goals: value expression and impression management. In our final companies are weighing in on social issues through their advertising study, we show how brands can leverage this psychological tendency campaigns and social media, and consumers are deciding whether to to encourage sharing on social media. share and propagate these messages. The current research suggests Study 1 examined the basic relationship between attitude fram- that if companies want consumers to share their messages on social ing and sharing across a variety of attitudes. We asked participants media, it may be more beneficial to talk about what a company sup- their attitudes toward seven statements regarding companies and ports, rather than what they oppose – even when a tweet relays the company policies (e.g., “Facebook has had a negative effect on same underlying message. To test this proposition, in Study 6 we America), varying the valence of the statements in order to ensure asked participants their attitudes towards immigration (“Immigration that “support” was not consistently associated positive attitudes. policies should be implemented that decrease/increase the number For each statement, participants reported their attitudes towards the of people coming into the US”), and used their answers to assign 1090 / Consumers and Politics: When Companies Take a Stance participants to support- or oppose- framed messages, as in previous Jordan, Jillian J., Sommers, Roseanna, Bloom, Paul, and Dave G. studies. Then participants saw a Tweet from Uber consistent with Rand (2017), “Why do we hate hypocrites? Evidence for a their own attitude, framed in terms of either what they supported theory of false signaling,” Psychological Science, 28, 356-368. (e.g., “We support policies that increase immigration”) or what they Jost, John T., Jack Glaser, Arie W. Kruglanski, and Frank J. opposed (e.g., “We oppose policies that decrease immigration.”), Sulloway (2003), “Political Conservatism as Motivated Social modelled from corporate tweets after the 2017 immigration ban. Cognition,” Psychological Bulletin, 129(3), 339-375. We then asked participants how likely they would be to retweet, Jost, John T., Brian A. Nosek, and Samuel D. Gosling (2008), and asked them to give us their email address to contact them about “Ideology: Its Resurgence in Social, Personality, and Political retweeting tweets like this one. Support framing led to greater shar- Psychology,” Perspectives on Psychological Science, 3 (2), ing intentions than oppose framing, t(975) = 2.89, p = .004. Partici- 126–136. pants who viewed support-framed attitudes were also more likely to Kay, Aaron C. and John T. Jost (2003), “Complementary Justice: agree to be contacted (27.5%) than participants who viewed oppose- Effects of ‘Poor but Happy’ and ‘Poor but Honest’ Stereotypes framed attitudes (17.5%), x2(2) = 13.92, p < .001. Thus, the cur- on System Justification and Implicit Activation of the Justice rent research suggests that support-framing allows companies to best Motive,” Journal of Personality and Social Psychology, 85(5), maximize their return on their investment for these sorts of decisions 823–37. and tweets. Kim, Jeehye Christine, Brian Park, and David Dubois (2018), “How Consumers’ Political Ideology and Status-Maintenance REFERENCES Goals Interact to Shape Their Desire for Luxury Goods,” Aaker, Jennifer L., Anne M. Brumbaugh, and Sonya A. Grier Journal of Marketing, 82(6). 132-149. (2000), “Nontarget Markets and Viewer Distinctiveness: The Lawton, Thomas, Steven McGuire, and Tazeeb Rajwani. Impact Of Target Marketing on Advertising Attitudes,” Journal “Corporate political activity: A literature review and research of Consumer Psychology, 9, 127—140. agenda.” International Journal of Management Reviews 15, Catapano, R. and Tormala, Z. (working paper). Do I Support Gun no. 1 (2013): 86-105. Control or Oppose Guns? Supporters are More Likely to Share Maio, Gregory R, Victoria M. Esses, and David W. Bell (1994), than Opponents. “The Formation of Attitudes Toward New Immigrant Groups,” Esses, Victoria M., John F. Dovidio, Lynne M. Jackson, and Tamara Journal of Applied Social Psychology, 24 (19), 1762—1776. L. Armstrong (2001), “The Immigration Dilemma: The Role Ordabayeva, Nailya, and Daniel Fernandes (2018), “Better or of Perceived Group Competition, Ethnic Prejudice, and Different? How Political Ideology Shapes Preferences for National Identity,” Journal of Social issues, 57(3), 389-412. Differentiation in the Social Hierarchy,” Journal of Consumer Feldman, J. M., & Lynch, J. G. (1988). Self-generated validity and Research, 45(2), 227-250. other effects of measurement on belief, attitude, intention, and Pollay, Richard W. (1986), “The Distorted Mirror: Reflections on behavior. Journal of Applied Psychology, 73(3), 421. the Unintended Consequences of Advertising,” Journal of Gajanan, M. (2018). Yelp Reviews Have Tanked for the Colorado Marketing, 50 (2), 18—36. Bakery at Center of Same-Sex Wedding Cake Case. Fortune Pounders, Kathrynn, and Amanda Mabry-Flynn (2016), “Consumer Magazine. (accessed on February 25, 2020 at: https://fortune. Response to Gay and Lesbian Imagery: How Product Type com/2018/06/05/masterpiece-cakeshop-anti-lgbt-supreme- and Stereotypes Affect Consumers’ Perceptions,” Journal of court-yelp/). Advertising Research, 56 (4), 426-440. Global Strategy Group (2012), “Business and Politics: Do Rao, A. R., & Monroe, K. B. (1989). The effect of price, brand They Mix,” http://globalstrategygroup.com/wp-content/ name, and store name on buyers’ perceptions of product uploads/2014/10/2014-Business-and-Politics-Report.pdf. quality: An integrative review. Journal of Marketing Research, Graham, J., Haidt, J., & Nosek, B. A. (2009). Liberals and 26(3), 351-357. conservatives rely on different sets of moral foundations. Shaw, Alex, DeScioli, Peter, Barakzai, Anam, and Robert Kurzban Journal of personality and social psychology, 96(5), 1029. (2017), “Whoever is not with me is against me: The costs of Grier, Sonya A. and Anne M. Brumbaugh (1999), “Noticing neutrality among friends,” Journal of Experimental Social Cultural Differences: Ad Meanings Created by Target and Psychology, 71, 96-104. Non-Target Markets,” Journal of Advertising, 28 (1), 79—93. Skitka, Linda J. (2010), “The psychology of moral conviction,” Haidt, Jonathan (2012), The righteous mind: Why good people are Social and Personality Psychology Compass, 4, 267-281. divided by politics and religion, Vintage Press. Stephan, Walter G. and Cookie White Stephan (2000), “An Ho, Arnold K., Jim Sidanius, Nour Kteily, Jennifer Sheehy- Integrated Theory of Prejudice,” in Reducing Prejudice and Skeffington, Felicia Pratto, Kristin E. Henkel, Rob Foels, and Discrimination: The Claremont Symposium on Applied Social Andrew L. Stewart (2015), “The Nature of Social Dominance Psychology, ed.S. Oskamp, Mahwah, NJ: Lawrence Erlbaum Orientation: Theorizing and Measuring Preferences for Associates. Intergroup Inequality Using the New SDO7 Scale,” Journal of Stephan, Walter G., C. Lausanne Renfro, Victoria M. Esses, Personality and Social Psychology, 109(6), 1003-1028. Cookie White Stephan, and Tim Martin (2005), “The Effects John, Leslie K., Barasz, Kate, and Michael I. Norton (2016), of Feeling Threatened on Attitudes Toward Immigrants,” “Hiding personal information reveals the worst,” Proceedings International Journal of Intercultural Relations, 29 (1), 1—19. of the national academy of sciences, 113, 954-959. When You Need a Crystal Ball: Factors that Affect the Riskiness and Accuracy of Consumer Predictions Chairs: Uri Barnea, Bocconi University, Italy Jackie Silverman, University of Delaware, USA

Paper #1: How Order Affects People’s Choices in Sequences of The last two papers examine what factors affect consumers’ ac- Independent Predictions curacy when making predictions of uncertain events. The third paper Jackie Silverman, University of Delaware, USA investigates what conditions lead people to follow advice for predict- Uri Barnea, Bocconi University, Italy ing future events. Seven incentive-compatible studies show that people are (significantly or directionally) more likely to follow advice Paper #2: Consumers are (Relatively) Risk Seeking in the that includes a confidence interval around the specific best guess. Domain of Error Thus, assuming that advisors give good advice, including confidence Berkeley J. Dietvorst, University of Chicago, USA intervals when giving advice can improve the accuracy of consum- Lin Fei, University of Chicago, USA ers’ predictions. Finally, the fourth paper focuses on the accuracy Paper #3: Should Advisors Provide Confidence Intervals of more personal, financial predictions, such as consumers’ expec- Around Their Estimates? tations of their future expenses and income. Nine studies featuring Celia Gaertig, University of Chicago, USA unique field data (e.g., budget forecasts made with a personal finance Joseph P. Simmons, University of Pennsylvania, USA app, longitudinal spending reports from credit union members), find Paper #4: A Prototype Theory of Consumer Financial that people rely on their “most common” past experience when mak- Misprediction ing predictions, leading them to under-predict the magnitude of their Ray Charles “Chuck” Howard, University of British Columbia, upcoming expenses, but over-predict the magnitude of their earnings. Canada Taken together, these papers shed light on how consumers make David Hardisty, University of British Columbia, Canada predictions, and when such predictions may be suboptimal. Given Dale Griffin, University of British Columbia, Canada the pervasiveness of uncertainty about the future, understanding con- Marcel Lukas, Heriot-Watt University, Scotland sumer predictions and their accuracy is critical for consumers and Abigail Sussman, University of Chicago, USA marketers alike.

SESSION OVERVIEW How Order Affects People’s Choices in Sequences of Consumers often make choices with a high degree of uncer- Independent Predictions tainty regarding the outcomes. For example, the decision to purchase an unfamiliar product could ultimately increase or decrease utility, EXTENDED ABSTRACT and consumers typically do not know beforehand what that outcome How does the order in which consumers make predictions might be. Similarly, consumers spend trillions of dollars each year on about independent events (e.g., gambles) affect those predictions? stocks, bonds, gambles, and other bets, which suggests that predict- We find in five studies (four pre-registered) that consumers more of- ing what might occur in the future is a common – and potentially ten choose that an unlikely outcome will occur later (vs. earlier) in lucrative – aspect of consumers’ lives. a sequence of predictions; that is, consumers more often choose the When it comes to understanding consumers’ predictions in “underdog” over the “favorite” for their second, third, and/or fourth these various contexts, two key questions stand out: what predictions prediction (versus first). do they actually make, and how accurate is that choice? The four We propose that when consumers predict the outcomes of mul- papers in this session explore how different aspects of the prediction tiple events, they make their first prediction without necessarily environment play an important role in consumers’ actual choices re- considering the other events for which they will make predictions, garding future events and the accuracy of such predictions. and thus rationally tend to predict that the more likely outcome will The first two papers examine what factors affect consum- occur. However, because people often falsely perceive independent ers’ predictions, and specifically, their willingness to choose more events as related to each other (i.e., the belief in the law of small “risky” options. The first paper explores when people are more likely numbers: Tversky and Kahneman 1971), they believe that the un- to choose the “risky” (i.e., less likely) outcome relative to the “safe” likely outcome is bound to happen at some point. This belief dispro- (i.e., more likely) outcome across a sequence of predictions; for portionately affects predictions later in the sequence, making con- example, when predicting outcomes in a sports tournament, when sumers more likely to predict that the unlikely outcome will occur might consumers pick the “underdog” to win over the “favorite?” later (versus earlier). Five studies show that people more often predict that unlikely out- By demonstrating that the (arbitrary) order in which consum- comes will occur later (versus earlier) in the series of predictions of ers make predictions affects their choices, this research contributes independent events because they believe that the unlikely outcome to work on the representativeness heuristic (Kahneman and Tversky must occur at some point even within a small sample. The second 1972), and specifically its effects on sequential bets (gambler’s falla- paper investigates how the potential to make forecasting errors, like cy: Croson and Sundali 2005) and predictions of multiple outcomes choosing the wrong team to win by X points, affects willingness to made simultaneously (probability matching: Gal and Baron 1996). select riskier gambles. Four studies show that because people have We are the first to show that the representativeness heuristic applies a diminishing sensitivity to this type of error, they are more likely to multiple sequential predictions for which the outcomes are un- to choose the riskier gamble or outcome when the gamble involves known (as opposed to a prediction following a known sequence of prediction error in addition to monetary consequences. outcomes, such as predicting “heads” after seeing several coin flip outcomes). Additionally, our findings contribute to the broad literature on variety seeking. The fact that consumers in our studies vary Advances in Consumer Research 1091 Volume 48, ©2020 1092 / When You Need a Crystal Ball: Factors that Affect the Riskiness and Accuracy of Consumer Predictions their predictions over time suggests that people seek variety even in From placing bets to predicting the weather for a multi-day contexts in which satiation, preference uncertainty, and social pres- vacation, consumers often have to make predictions about multiple sure (e.g., Kahn 1995) are unlikely to play a role. future events. Thus, our finding – that the order in which consumers In Study 1 (N = 402), we show the predicted effect “within make predictions has a systematic effect on their choices – is impor- subject:” consumers are more likely to choose the underdog later in tant for marketers and consumers alike. a sequence of predictions (compared to their first prediction). Par- ticipants predicted the outcomes of three randomly ordered hypo- Consumers are (Relatively) Risk Seeking in the Domain thetical basketball games in which one team slightly outranked the of Error other (e.g., Team A ranked 12th vs. Team B ranked 15th). A repeated measures logit model revealed that more participants predicted that EXTENDED ABSTRACT the underdog would win in the second (38%) and third games (40%) In most theories of choice under risk, preferences are elicited they made predictions about (vs. the first: 31%;p s < .003). over prospects (outcomes and associated probabilities; see Starmer, Study 2 (N = 601) replicated this effect in a different paradigm 2000). When testing these theories, most researchers have used pure- with real incentive-compatible choices (making predictions about ly monetary gambles (e.g. 50% chance of $1,000 vs 100% chance upcoming Rugby World Cup games). Participants predicted the out- of $450; Kahneman & Tversky, 1979). People’s choices over these comes of one close game (Wales, 5th vs. Australia, 6th) and two other prospects are supposed to inform us about their choices in the real games with clear favorites (e.g., Ireland, 1st vs. Russia, 20th). We ran- world. However, real world prospects usually have outcomes that are domized the order such that participants predicted the outcome of not purely monetary. Namely, they also produce an error – the dif- the close game either first or last. We also manipulated whether par- ference between a prediction and the realized outcome. For example, ticipants were incentivized to make correct predictions or not: half betting that a die will land on “6” when it actually lands on “4” proof the participants knew they would receive a bonus if they made a duces an error of two. Sports betters’ realize an error when betting correct prediction, and the other half did not receive any bonuses. A on a spread, manufacturers realize an error when deciding how much logit model revealed a main effect of order condition, such that more of a good to produce, and investors realize an error when selecting participants chose Australia (i.e., the underdog) when they predicted assets based on projections. If this context affects people’s appetite that game’s outcome last (59%) versus first (39%; p < .001). Sepa- for risk, then people’s preferences in the abstract scenarios used in rate chi-square analyses showed that the effect held both with and past research may not accurately represent consumer’s preferences without incentives (ps < .01; ANOVA interaction p > .40). An ad- in the real world. ditional study using four real NCAA “March Madness” tournament In this project, we find that consumers make fundamentally dif- basketball games replicated this order effect (Study 3, N = 199: first ferent choices when gambles produce an error in addition to a mon- = 31% vs. last = 49%, p = .010). etary outcome. Recent work has found evidence that people have In Study 4 (N = 178), we examined the effect of choice order diminishing sensitivity to forecasting error, which makes them risk in a more generalizable, non-sports context: predictions about what seeking over forecasting error (Dietvorst, & Bharti, 2019). Follow- color ball (red or black) would be drawn from an urn. As in Studies ing this work, we predict that people will be significantly more risk 2 and 3, participants predicted the outcome of a draw from a focal seeking when their choice in a gamble will produce an error in addi- “close” urn (with 27 red balls and 23 black balls) either before or tion to a monetary gain or loss. after predicting the outcomes of two draws from urns with a large In Study 1, 755 participants chose between safe and risky op- majority of red balls (e.g., 43 out of 50). Again, more participants tions and we manipulated whether their choice produced an error. chose the unlikely option (black) when they made the focal predic- Participants in the “Gain” chose between a 50% chance of $0.30 tion last (43%) versus first (6%;p < .001). We also found evidence of and $0.12 for sure. They learned that the computer would flip a coin the psychological process driving this effect: we asked participants to determine the outcome, but they did not learn how this process how much they were thinking about their other predictions when would work. Thus, their choice did not produce an error because it making their prediction for the focal urn. Participants in the last con- was not clear what outcome they were betting on when they chose dition reported considering their other predictions more (M = 5.21) the risky option. Participants in the “Heads” chose between a gain of than in the first condition (M = 3.10; p < .001), which mediated the $0.30 if the coin landed on heads (50% chance), and $0.12 for sure. effect of order on prediction (Indirect effect = -0.05, SE = 0.03, 95% Participants in the “Choice” condition faced the same gamble, but CI = [-0.13, -0.01]). could choose whether heads or tails resulted in the $0.30 gain. Thus, Study 5 (N = 2000) tested our theory through moderation. We the risky option in the Choice and Heads conditions produced an hypothesized that if participants’ belief that the unlikely outcome error because participants knew what outcome they were betting on. must occur at some point in the sequence drives the effect, then it We found that participants in the Gain condition were significantly should be attenuated when participants learn that the events are in more likely to pick the safe option (44%) than participants in the fact independent. As in Study 4, participants predicted what color Heads (31%, X2(1, n=504)=9.34, p=.002) and Choice (25%, X2(1, ball would be drawn from three different urns, with the same focal n=504)=20.43, p<.001) conditions. We replicate this effect in Study urn appearing either first or last. We also manipulated if, in the in- 2, with a different coin flip gamble. structions phase, participants completed a short tutorial on how sepa- In Study 3, 401 participants bet on a gamble that would be re- rate sequential events (e.g., coin tosses) are statistically independent solved with a die roll. Participants in the Gain condition chose be- or not. The order effect replicated for participants who did not do the tween three options without learning how the outcomes would be tutorial (last = 8% vs. first = 36%; p < .001) and was still present but determined. The two riskier options each offered a 40% chance of smaller among participants who did the tutorial (last = 8% vs. first = $0.30, a 20% chance of $0.10, and a 40% chance of $0; the safer 25%; p < .001, interaction p = .048). Importantly, among participants option offered a 20% chance of $0.30 and an 80% chance of $0.10. who made the focal prediction last, participants who completed the Participants in the Error condition learned how these gambles cor- tutorial were less likely to choose the black ball (p = .001). responded to the outcome of the die roll: 1) the die was 5-sided and displayed the numbers “1, 1, 2, 3, 3”, 2) they could bet on “1”, “2”, Advances in Consumer Research (Volume 48) / 1093 or “3”, and 3) they would earn $0.30 for a perfect prediction, $0.10 how advisees react to uncertain advice in the form of confidence for a prediction off by 1, and $0 for a prediction off by 2. Thus, the intervals. two riskier options corresponded to betting on “1” and “3”, while the In the current project, we set out to test whether people are safer option corresponded betting on “2”. We displayed the odds of more or less likely to follow advice that is accompanied by a con- each monetary outcome the same way to both groups to ensure that fidence interval. For example, the website Kayak.com may recom- they had the same information about the monetary gamble. We found mend to buy a flight ticket now because “prices will rise in the next that participants were significantly more likely to pick the safer op- two weeks” (certain advice) or because “there is a 90% chance that tion in Gain condition (56%) than the Error condition (39%, X2(1, prices will rise between $20 and $30 in the next two weeks” (uncer- n=401)=11.89, p=.001). tain confidence-interval advice). We tested participants’ reactions to In Study 4, we explore this phenomenon in a completely dif- these different types of advice in the domain of sports predictions, ferent context. Participants saw two different forecasting options as this allowed us to incentivize participants for accurate predictions (Option E and Option S) make 20 trial forecasts in a prediction and to give them high-quality advice. task, and then chose between those options for a final incentivized Across seven studies (N = 9,471), we found that participants are forecast. Participants earned $0.50 for a prefect forecast, and this directionally more likely to follow advice that is accompanied by a bonus decreased by $0.10 for each unit of error in the forecast. Un- confidence interval. The effect was significant in four of the seven like the previous studies, participants only learned about the odds studies, and non-significant in the other three studies. Our results and outcomes associated with their options through experience. We suggest that advisors may be more persuasive if they provide confi- configured the predictions so that Option E was always off by 1, and dence intervals around their estimates. Option S was equally likely to be perfect or off by 2. Thus, Option In Study 1 (N = 399), we asked participants to predict the out- E was a safe option that always produced $0.40, and Option S was comes of upcoming NBA basketball games. We presented partici- a risky option that produced $0.50 and $0.30 with equal probability. pants with 9 different games, and, for each game, we asked them to We manipulated the instructions and feedback that participants got predict how many points one of the teams would score in the game between conditions. Participants in the Gain condition never learned (e.g., “How many points will the Philadelphia 76ers score in this about the prediction component of the task – they only saw the gains game?”). Participants first made a prediction for each of the games. that the two options generated in each of the 20 trials. Participants in Then, we asked them to choose whether they would want to sub- the Error condition only learned about the prediction task; they saw mit their own prediction or a prediction made by an algorithm as the predictions, the gains, and the correct answers in each trial. We their official, incentivized prediction. Before making their choice, found that participants were significantly more likely to pick the safe participants saw how the algorithm had performed in the past. For option (Option E) in the Gain condition (53%) than the Error condi- each game, participants were then shown the algorithm’s prediction tion (40%, X2(1, n=401)=6.59, p=.010). alongside their own prediction, and were asked to choose which they These results suggest that consumers are more likely to make would like to submit as their official, incentivized prediction. risky choices when their choices produce an error in addition to a We manipulated the certainty of the algorithm’s advice by ma- monetary gain or loss. These results can help to explain how the nipulating whether or not the advice was accompanied by a con- same consumer can simultaneously buy insurance (a choice without fidence interval. In the certain advice condition, participants only a clear prediction component) and gamble on sports (choices with a saw the algorithm’s best guess (e.g., “The statistical model’s best clear prediction component) – two actions that suggest very different guess is that the Philadelphia 76ers will score 111 points.”). In the risk preferences. Further, they suggest that marketers can manipulate uncertain advice condition, the best guess was accompanied by a people’s tolerance for risk by highlighting or hiding the prediction confidence interval (e.g., “According to the statistical model, there component of a risky choice. For example, consumers may value is a 90% chance that the Philadelphia 76ers will score between 84 uncertain promotions (see Mazar, Shampanier, & Ariely, 2017) more and 118 points. Its best guess is that the Philadelphia 76ers will score when those promotions generate an error in addition to a monetary 111 points.”). We found that participants were more likely to follow outcome. the algorithm’s advice when it was accompanied by a confidence interval than when it was not accompanied by a confidence interval. Should Advisors Provide Confidence Intervals Around Specifically, in Study 1, 25% of participants followed the algorithm’s Their Estimates? advice when it was accompanied by a confidence interval, but only 18.7% of participants followed the advice in the certain advice con- EXTENDED ABSTRACT dition, b = .063, SE = .026, p = .015. Many of a consumer’s everyday decisions require them to make In six additional studies, we replicated this procedure using dif- accurate predictions about inherently uncertain events. For example, ferent incentives and different sports prediction tasks. In all of our consumers may need to forecast how likely it is that the price of a studies, participants were directionally more likely to follow the al- product will fall, that a new model for a product will be released, or gorithm’s advice when it was accompanied by a confidence interval that their investment will be profitable. To navigate such uncertain- than when it was not accompanied by a confidence interval. This ties, consumers often ask for and rely on advice. Despite the potential effect was significant in three of the additional six studies (p = .032, perils of following overly precise advice, advice is often given with p = .004, and p = .007), and non-significant in the remaining three too much certainty (Moore and Healy 2008; Moore, Tenney, and Ha- studies (p = .112, p = .314, and p = .439). ran 2016; Radzevick and Moore 2011; Soll and Klayman 2004). Taken together, our results suggest that presenting a confidence Recent research suggests that consumers actually do not inher- interval alongside advice may be appealing to customers. Our results ently dislike uncertain advice, as they do not judge advisors more have important practical implications and are relevant to any compa- negatively when these advisors include uncertainty into their advice ny or individual who seeks to effectively communicate their advice. (Gaertig and Simmons 2018). However, we do not yet know (1) whether people are more likely to follow uncertain advice, and (2) 1094 / When You Need a Crystal Ball: Factors that Affect the Riskiness and Accuracy of Consumer Predictions A Prototype Theory of Consumer Financial that the effect of distribution condition on predictions is mediated by Misprediction what participants perceive to be their “average” spending (indirect effect = 3.41, SE = 1.32, 95% CI = [1.05, 6.37]). EXTENDED ABSTRACT In Studies 4–5 we test H2 in the context of expenses. In Study Consumers frequently make financial predictions (Peetz et al. 4 we partnered with Canada’s largest credit union to run a five week 2016, Zhang and Sussman 2018), and the accuracy of these predic- longitudinal field study with 187 of its members. In weeks one tions can be consequential for their financial well-being. For exam- through four participants were asked to predict their spending at the ple, under-predicting future expenses is associated with high-interest beginning of the week then report their actual spending at the end of payday loan use (Pew 2012) and revolving credit card debt (Yang the week. In the fifth week of the study participants were randomly et al. 2007). Similarly, over-predicting future income may lead con- assigned to predict their expenses in a control condition or an inter- sumers to over-spend because expected income growth typically vention condition that prompted them to consider three reasons why leads to increased spending (e.g., Berman et al. 2016; Carroll et al. their spending for the next week would be different from a typical 1994). However, relatively little is known about the prevalence and week before they made their prediction. During the first four weeks magnitude of expense and income prediction biases, why these bi- of the study participants under-predicted their weekly spending be- ases occur, and how these biases can be neutralized. tween 11% and 19% (p’s < .01). During the fifth week of the study In the present research we develop a prototype theory of con- participants in the control condition once again under-predicted their sumer financial misprediction that helps explain: 1) why consum- weekly spending (mean bias = 18%, p < .01). However, supporters display an expense prediction bias in which they under-predict ing H2, under-prediction was virtually eliminated in the interven- their future spending, 2) why consumers who face variable income tion condition (mean bias = -1.3%, p = .85). Study 5 (n = 1,048, display an income prediction bias in which they over-predict their nationally representative sample of US consumers) extends Study future income, and 3) how consumers’ expense and income predic- 4 by demonstrating that the intervention makes it easier to think of tion accuracy can be improved. The logic of the prototype theory atypical expenses when predicting, and that the number of atypical is that financial predictions are based on prototype attributes that expenses that come to mind mediates the effect of the intervention come to mind easily when predictions are being constructed. These on predictions (indirect effect = .05, SE = .02 95% CI = [.02, .09]). attributes represent a consumer’s average spending or income, where In Studies 6–9 we test H1 and H2 in the context of income pre- “average” refers to the mode of the relevant distribution rather than dictions. Our primary expectations were that consumers who work the mean. This leads consumers to under-predict expenses because in the gig economy tend to over-predict their future earnings because expenses are positively skewed with mode < mean, and over-predict gig income is often negatively skewed with mode > mean (Hall and income when it is negatively skewed with mode > mean. One impor- Krueger 2018), and that our “atypical” intervention would therefore tant implication of this framework is that predictions will be more make income predictions more accurate by making them lower. To accurate when the relevant distribution is approximately normal briefly summarize the results of these studies: 1) Uber drivers (n = (with mode = mean) than when it is skewed (H1). A second key im- 57) and Mturkers (n = 200) over-predict their weekly gig income by plication is that when outcomes are naturally skewed prediction ac- as much as 61% (p’s < .001), 2) Income predictions are more accu- curacy can be improved by prompting consumers to consider reasons rate when income is normally distributed than when it is negatively why their expenses (income) might be different than usual because skewed, and 3) Prompting people to consider reasons why their gig most values in a positively (negatively) skewed distribution that are income for the next week might be different than usual reduces over- different from the mode are also higher (lower) than the mode (H2). prediction. In Studies 1–3 we test H1 in the context of expenses. In Study 1 we use longitudinal data from a popular personal finance app REFERENCES (+400 million transactions across +65,000 users from 2014 to 2018) André, Quentin, Nicholas Reinholtz, and Bart de Langhe. “Variance to examine monthly prediction accuracy across expense categories Spillover in Intuitive Statistical Judgments.” ACR North that naturally vary in terms of skew. On average, consumers under- American Advances (2017). predict their monthly spending by 21.1% (p < .001). However, sup- Berman, Jonathan Z., An TK Tran, John G. Lynch Jr, and Gal porting H1, prediction accuracy is higher in expense categories char- Zauberman. “Expense neglect in forecasting personal acterized by more normally distributed spending (e.g., Fuel: Skew finances.” Journal of Marketing Research 53, no. 4 (2016): = .10, Mean under-prediction = .03%, n.s.) and lower in categories 535-550. characterized by more positively skewed spending (e.g., Dining and Carroll, Christopher D., Jeffrey C. Fuhrer, and David W. Wilcox Drinking: Skew = .38, Mean under-prediction = 34.22%, p < .001). (1994), “Does Consumer Sentiment Forecast Household Study 2 extends the descriptive findings of Study 1 by examin- Spending? If So, Why?” American Economic Review, 84 ing the causal effect of skew on expense predictions. Participants (December), 1397–1408. viewed 52 weekly expense values in quick succession (Andre et al. Croson, Rachel, and James Sundali (2005) “The Gambler’s Fallacy 2017), then predicted their spending for the next week. The experi- and the Hot Hand: Empirical Data from Casinos.” Journal of mental manipulation was that the values were drawn in random or- risk and uncertainty, 30 (3),195-209. der without replacement from either a positively skewed distribution Dietvorst, B., & Bharti, S. (2019). People Reject Algorithms in with mode < mean (Mode = $180, Mean = $200, SD = 38.14, Skew Uncertain Decision Domains Because They Have Diminishing = 3.13) or a normal distribution with mode = mean (Mode = $200, Sensitivity to Forecasting Error. Available at SSRN 3424158. Mean = $200, SD = 28.14, Skew = 0.00). Supporting H1, and consis- Gaertig, Celia and Joseph Simmons (2018). “Do People Inherently tent with the descriptive results of Study 1, predictions in the normal Dislike Uncertain Advice?” Psychological Science, 29(4), distribution condition were remarkably close to the $200 mean of 504-520. the underlying distribution (M = $200.32, p = .83), but predictions in Gal, Ido, and Jonathan Baron (1996), “Understanding Repeated the positive skew condition were significantly lower than $200 (M = Simple Choices,” Thinking and Reasoning, 2, 81–98. $195.14, p < .02). Study 3 directly replicates Study 2 and also shows Advances in Consumer Research (Volume 48) / 1095

Hall, Jonathan V., and Alan B. Krueger. “An analysis of the labor Radzevick, Joseph R. and Don A. Moore (2011). “Competing to market for Uber’s driver-partners in the United States.” ILR Be Certain (But Wrong): Market Dynamics and Excessive Review 71, no. 3 (2018): 705-732. Confidence in Judgment.”Management Science, 57(1), 93- Kahn, Barbara E. (1995), “Consumer Variety-Seeking Among 106. Goods and Services: An Integrative Review,” Journal of Soll, Jack B. and Joshua Klayman (2004). “Overconfidence in Retailing and Consumer Services, 2 (3), 139–148. Interval Estimates.” Journal of Experimental Psychology: Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis Learning, Memory, and Cognition, 30(2), 299-314. of decision under risk. Econometrica, 47, 263–291. Starmer, C. (2000). Developments in non-expected utility theory: Kahneman, Daniel, and Amos Tversky (1972) “Subjective The hunt for a descriptive theory of choice under risk. Journal Probability: A Judgment of Representativeness,” Cognitive of economic literature, 38(2), 332-382. Psychology, 3 (3), 430-454. Trusts, Pew Charitable. “Payday lending in America: Who Mazar, N., Shampanier, K., & Ariely, D. (2017). When retailing borrows, where they borrow, and why.” Washington, DC: Pew and Las Vegas meet: Probabilistic free price promotions. Charitable Trusts (2012). Management Science, 63(1), 250-266. Tversky, Amos, and Daniel Kahneman (1971), “Belief in the Law Moore, Don A., Tenney, Elizabeth R., and Uriel Haran (2016). of Small Numbers,” Psychological Bulletin, 76 (2), 105. “Overprecision in Judgment.” In G. Wu & G. Keren (Eds.), Yang, Sha, Livia Markoczy, and Min Qi. “Unrealistic optimism Handbook of Judgment and Decision Making (pp. 182–212). in consumer credit card adoption.” Journal of Economic New York: Wiley. Psychology 28, no. 2 (2007): 170-185. Moore, Don and Paul J. Healy (2008). “The Trouble With Zhang, C. Yiwei, and Abigail B. Sussman. “Perspectives on mental Overconfidence.”Psychological Review, 115(2), 502-517. accounting: An exploration of budgeting and investing.” Peetz, Johanna, Melanie Simmons, Jingwen Chen, and Roger Financial Planning Review 1, no. 1-2 (2018): e1011. Buehler. “Predictions on the go: Prevalence of spontaneous spending predictions.” Judgment and Decision Making 11, no. 1 (2016): 48. When Beneficence Backfires: Negative Consequences of Doing Good Chairs: Yonat Zwebner, The InterDisciplinary Center (IDC) Herzliya, Israel Shalena Srna, University of Michigan, USA

Paper #1: Unable to Rebound: The Downside of Being a do-gooders themselves may act in a way that harms their side of Nonprofit That Errs a political cause (e.g., gun control). Counterintuitively, when faced Shalena Srna, University of Michigan, USA with the choice between taking money from their side of a cause or Yonat Zwebner, Interdisciplinary Center (IDC), Israel allocating money to the opposing side, people take money from their own side. In fact, the more one cares about their cause, the more Paper #2: Doing Good for (Maybe) Nothing: How Reward likely they are to harm their side rather than help the opposition, Uncertainty Shapes Consumer Responses to Prosocial Behavior undermining their ability to do good for their cause. Ike Silver, University of Pennsylvania, USA Taken together these four projects highlight the importance of Jackie Silverman, University of Delaware, USA understanding the negative responses to actors who aim to do good. Paper #3: Damned Either Way: Hypocrisy Judgments When Counterintuitively, this session identifies that beneficence can back- Goals and Commitments Conflict fire and reveals the potential negative consequences of doing good. Jonathan Z. Berman, London Business School, UK In line with the conference theme, the impact and widespread appli- Graham Overton, Bocconi University, Italy cability of these papers’ findings could attract researchers, practitio- Daniel A. Effron, London Business School, UK ners, and policy-makers. Paper #4: The Cost of Opposition: Preferring to Punish our Own Rather than Help our Opponent Unable to Rebound: The Downside of Being a Nonprofit Rachel Gershon, University of California, San Diego, USA That Errs Ariel Fridman, University of California, San Diego, USA EXTENDED ABSTRACT SESSION OVERVIEW Organizations often err. Understanding peoples’ responses to Doing good is generally beneficial in terms of reputation and organizational mistakes and apologies is of utmost managerial im- status (Flynn 2003; Flynn et al. 2006; Hardy and Van Vugt 2006). portance (Koehn 2013; Pace et al. 2010). Previous work examining Further, people not only have greater goodwill towards prosocial these responses has focused on for-profits and not nonprofits (e.g., agents, but also evaluate these agents more positively in unrelated Gregoire and Fisher 2008; Joireman et al. 2013; Reimann et al. domains (i.e., a moral halo effect; Chernev and Blair 2015). How- 2018). Yet, how do consumers respond to a nonprofit that errs and ever, when might consumers respond negatively to do-gooders? In apologizes? this session, we focus on instances when beneficence backfires. Spe- Nonprofits are generally perceived more positively than for- cifically, four papers identify cases in which people respond nega- profits (Aaker et al. 2010). Indeed, Bhattacharjee, Dana, and Baron tively towards those who do good (in terms of willingness-to-spend, (2017) document that people hold anti-profit beliefs such that non- consumer decisions, judgments, and monetary allocations). This profits are judged as doing less harm and more societal good than for- novel experimental work highlights the importance of understanding profits. Further, prosocial actors receive reputational benefits such responses to do-gooders and contributes to the literature on altruism, as high status (Flynn 2003; Hardy and Van Vugt 2006), and a be- prosociality, CSR, and social and political causes. nevolent halo effect (Chernev and Blair 2015; Sen and Bhattacharya The first two papers show how beneficence backfires when pro- 2001). Given these findings, one would expect that people would be social actors behave in a manner that is seemingly inconsistent with more forgiving towards a nonprofit than a for-profit that transgresses. their mission. Despite generally positive perceptions of nonprofits Counter to this, we find that consumers are more resentful of an or- relative to for-profits, Paper 1 shows that people are actually more ganization that transgresses when it is a nonprofit than when it is a resentful and less forgiving of a nonprofit than a for-profit that com- for-profit. Further, consumers feel more exploited by a nonprofit that mits the same wrongdoing. These findings are particularly important errs than when a for-profit errs. Therefore, even after apologizing, as evident by the dramatic decrease in Red Cross donations follow- nonprofits still have a harder time rebounding than for-profits. ing news that it mismanaged Haitian relief money. Consistent with Study 1 (N=501, MTurk) was a 2(between: nonprofit vs. for- past work, Paper 2 finds that actors who do good in exchange for cer- profit) x 3(within: wrongdoing) design. Participants evaluated 3 dif- tain rewards suffer negative consequences because they are judged ferent organizations that had done wrong (released pollutants, pub- as less purely motivated. However, interestingly, this paper finds that lished inaccurate information on its website, had misleading ads). this effect is attenuated when rewards for doing good are uncertain; The organizations were described identically except they were either for example, consumers prefer to buy products from and write more nonprofits or for-profits. Across all 3 wrongdoings, participants who positive reviews of companies whose prosocial initiatives face un- evaluated nonprofits were less likely to continue donating/spending certain profit outlooks (i.e., they may earn no profit at all) versus and perceived the organization more negatively than those who eval- certain profits of equivalent expected value. uated for-profits (ps<.01). Thus, nonprofits that transgress are penal- The last two papers examine how identity plays a role in nega- ized more than for-profits for the same transgression. tive responses to those who do good in the context of social and Study 2 (N=200, Mturk) builds on Study 1 by examining re- political causes, an area particularly relevant in today’s ideologically sponses towards an organization that has not only erred (misleading charged world. Paper 3 examines situations in which do-gooders are ads) but also apologized. We find that people are less forgiving of a faced with a tradeoff between acting inconsistently with their past nonprofit than a for-profit. Specifically, participants were still more behavior versus acting inconsistently with their underlying values, likely to avoid donating/spending at a nonprofit (M=5.33) than at a and finds that in these situations actors are condemned to be judged for-profit (M=4.48) after the organization apologized for their mis- as hypocrites regardless of what is chosen. Paper 4 demonstrates that take (p<.001).

Study 3 (N=167, lab) ensures that differences in baseline will- vilified after it mismanaged and misrepresented the use of donation ingness to donate/spend cannot explain the effect. After reading a de- money intended to support the victims of the 2010 Haiti earthquake. scription about the organization (nonprofit or for-profit), participants Following news of the scandal, the Red Cross had the lowest amount were equally likely to give to/spend at the organizations (M=5.50 of donations in over 15 years (The Washington Post 2015). Thus, vs. 5.23, p = .18). However, after they learned about the wrongdoing paradoxically, those who do good have trouble rebounding after they and apology, participants were less willing to give to/spend at the transgress. nonprofit than the for-profit (M=3.25 vs. 4.42, p<.001; interaction: p<.001). Doing Good for (Maybe) Nothing: How Reward Next, we examined why consumers have more negative re- Uncertainty Shapes Consumer Responses to Prosocial sponses towards nonprofits who erred and apologized compared to Behavior for-profits who did the same. Consistent with Expectation-Discon- firmation Theory (Oliver 1977), we reason that consumers feel ex- EXTENDED ABSTRACT ploited by a nonprofit that errs because they expect the organization Scholars have long studied decision-making under reward un- to do good in exchange for their money, while this expectation does certainty (i.e., when there is an ex ante chance of no reward at all; not exist for for-profits. In Study 4 (N=404, lab), after reading about Kahneman and Tversky 1979). Here, we examine how people evalu- an organization that donates (vs. sells) warm clothing and had apolo- ate the decisions of others in response to reward uncertainty. Specifi- gized for doing wrong (misleading ads), participants again indicated cally, we hypothesize that actions taken when rewards are uncertain they were less likely to continue donating/spending at the nonprofit seem more purely (i.e., intrinsically) motivated, whereas actions tak- (M=3.57) compared to the for-profit (M=4.90; p<.001). Further- en when rewards are certain seem more tainted by extrinsic benefits, more, participants also felt more exploited by the erring organiza- holding reward value constant. tion when it was a nonprofit (M=4.60) than when it was a for-profit We investigate this effect in the domain of prosocial behavior, (M=3.85; p<.001). Importantly, this reduced forgiveness towards the where inferences about motive purity powerfully guide judgment nonprofit (vs. the for-profit) was mediated by how exploited partici- and evaluation (Critcher and Dunning 2011). Although prosocial be- pants’ felt (B=-.62; 95% CI [-.85, -.39]). Lastly, participants were havior can lead to substantial reputational benefits for the actor, good asked whether they would prefer to transact with a different organi- deeds can backfire if they appear tainted by extrinsic benefit (Small zation, as opposed to the erring organization, that has lower quality and Cryder 2016). In extreme cases, doing good and turning a profit of clothing but has never erred. Consistent with our main finding, can provoke stronger backlash than doing no good at all (Newman 56.9% of participants in the nonprofit condition preferred to switch and Cain 2014). to the organization that had never erred, while only 7.4% indicated We theorize that when selfish rewards for doing good are un- they would switch in the for-profit condition p( <.001). certain ex ante, observers infer that prosocial actors would have Finally, we rule out an alternative account. Specifically, if the been willing to do good for nothing. For example, an individual who unfavorable responses towards nonprofits that err are because people volunteers in exchange for a raffle ticket with a large (but unlikely) feel personally exploited by the organization, then we should not payout seems like she would have volunteered for free (even if she observe these unfavorable responses towards nonprofits when advis- wins); a volunteer receiving a smaller sure reward enjoys no such ining someone else on whether or not to continue transacting with the ference. Analogously, a brand that launches a sustainability initiative organization. Accordingly, in Study 5 (N=603, Mturk) we employed despite an uncertain profit outlook seems more praiseworthy than a 2(organization type: nonprofit vs. for-profit) x 2(self vs. other) be- one that launches an identical initiative with a certain profit outlook. tween-subjects design. Similar to the previous studies, the organiza- Studies 1a-b (n=128, n=199) test our prediction in the con- tions were described identically except they were either nonprofits or text of product choices and customer reviews. In Study 1a, partici- for-profits. Participants were asked whether they would (vs. advise pants read about two chocolate brands which recently switched to their friend to) continue donating to/shopping at an organization. fair-trade cacao beans. For one brand, the fair-trade program was We found a main effect for organization type (nonprofit: M=4.05 vs. sure to boost profits by $100,000 (certain reward). For the other, for-profit: M=4.74, p<.001), and no significant interaction (p=.304). the program had a 25% chance of boosting profits by $400,000 (un- Replicating the effect observed in the previous studies, participants certain reward; equivalent in expectation). Participants subsequently who were asked about their own money, were less willing to keep evaluated the brand facing uncertain profits as more purely - moti donating at a nonprofit (M=4.20) than they were to keep spending vated (p<.001, d=.87), and 64% preferred to take home a chocolate at a for-profit (M=5.03) after the organization apologized for their from that brand (Z=3.18, p<.01). In Study 1b, participants read a mistake (p<.001). Furthermore, this effect was also observed in press release about a new corporate philanthropy project from a ma- the other condition in which participants advised their friend (non- jor wireless provider. We manipulated, this time between-subjects, profit: M=3.91 vs. for-profit: M=4.46; p=.003). Unexpectedly, we industry reports about the project’s likely impact on the brand’s bot- also observed a main effect for self vs. other (self: M=4.61 vs. other: tom line (certain reward: 100% chance of $100,000 profit, uncertain M=4.18, p=.001) suggesting people in general are less willing to ad- reward: 10% chance of $1,000,000 profit). Participants again rated vise others to transact with an organization after it errs. Together, the brand facing uncertain rewards as more purely motivated (p<.01, these results suggest that feeling exploited towards the nonprofit is d=.43), and this time they also wrote more positive customer reviews not because of a deeper personal connection with an organization. (p=.001, d=.48). In sum, previous research has mostly focused on corporate rep- Studies 2a-b (n=150, n=301) replicate this effect of reward un- utations and consumers’ reactions to for-profits who err; however, certainty on motive inference across different profit magnitudes and less is known about how people respond to nonprofits that err. Our prosocial acts. These studies also find effects of reward uncertainty findings suggest that, instead of giving nonprofits that transgressed on downstream inferences about moral character, likeability, and slack the opposite is true: after erring, consumers are more resentful predictions about future good behavior (ds>.20). and less forgiving of nonprofits than for-profits. This has important Study 3 (n=248) asks whether these effects can be explained by social and financial implications. For instance, the Red Cross was an alternative mechanism: differences in perceived value between 1098 / When Beneficence Backfires: Negative Consequences of Doing Good certain and uncertain rewards. We first asked participants to imag- socially risky. For instance, individuals often denigrate the character ine receiving a raffle ticket with a 10% chance of winning a $250 of do-gooders if those prosocial others are seen as implicitly im- gift card and to indicate how much sure compensation they would pugning the character of the self (Minson and Monin 2012). More require to trade in the ticket. Two weeks later, we invited the same generally, people tend to be suspicious that prosocial actors harbor participants back for our main study. Returning participants read self-interested motives (Berman, et al. 2015; Critcher and Dunning about a donor who gives blood either in exchange for the same raffle 2009; Miller 2001). Such suspicions can grow into perceptions that ticket (uncertain reward) or a gift card in the value the participant the actor is a hypocrite – someone who benefits from appearing vir- had previously indicated (certain reward). Although the rewards re- tuous without paying the price of actually being virtuous (e.g., Mo- ceived were matched on participants’ perceived value, the donor in nin and Merritt 2012). the uncertain reward condition was still seen as more purely moti- Hypocrisy often involves inconsistency but not all inconsistent vated (p<.001, d=.56) and likely to donate blood again in the future behavior is labeled hypocrisy. To perceive hypocrisy, an audience (p<.001, d=.64). must interpret inconsistency as signaling that a person appears or Study 4 (n=605) tests whether the impact of reward uncertainty feels more virtuous than they deserve (Effron et al. 2018). However, on motive inference is robust to ex post reward outcomes. Partici- the conditions under which inconsistency invites this interpretation pants were asked to evaluate a manufacturing company that launched are poorly understood. a new recycled product with either a certain profit forecast (100% We investigate how people perceive two types of inconsistency chance of $100,000) or an uncertain profit forecast (10% chance of that a prosocial consumer could display: commitment-based incon- $1 million). This time, participants also learned about profit out- sistency, whereby the consumer’s present actions are inconsistent comes, yielding a three-condition design (certain reward, realized with past commitments made; and goal-based inconsistency, where- uncertain reward, and unrealized uncertain reward). The brand was by the consumer’s actions fail to maximize the stated ethical goal judged to be more purely motivated in both uncertain reward con- they are trying to achieve. ditions relative to the certain reward condition (ps<.001, ds>.70). Consumers sometimes must choose between acting inconsis- Here, even if a prosocial initiative turned a 10x profit ex post, it was tently with their past commitments or their stated goals. Consider still seen as more purely motivated. someone who adopts vegetarianism to benefit the environment. Now Studies 5a-b (n=369, n=1,419) explore boundary conditions imagine an invasive fish species is destroying local habitats, and en- based on reward probability. Study 5a finds that the effect is specific vironmentalists recommend sell the fish commercially for consump- to situations in which uncertain profits include a chance of earning tion in order to limit the damage. Suppose the consumer is offered nothing, and not present in broader cases of risk over multiple pos- the invasive fish for dinner. She must choose between acting incon- sible outcomes (i.e., a chance of either a smaller or larger profit). sistently with her commitment to avoid meat versus acting inconsis- Study 5b tests the effect over seven profit probabilities, finding that tently with her goals to help the environment. motives begin to appear substantially less tainted when the chance of Across four studies, we find evidence for a “Damned Either no reward reaches 50%. Way” effect such that consumers are labeled hypocrites for display- Study 6 (n=1,007) finds that the effect is attenuated if we make ing either commitment- or goal-based inconsistency (Studies 1-4). salient that the actor needs the large possible reward in the uncer- Study 1 presents an initial demonstration of the effect and Study 2 as- tain condition. Participants read about a student who volunteers in sess its robustness. Study 3 shows that this effect occurs when evalu- exchange for a raffle ticket with a 5% chance at $400 (uncertain re- ating consumers who provide a reason for their commitment, but not ward) or a sure payment of $20 (certain reward). As we found previ- for those who believe that a given behavior represents a categorical ously, participants judged the volunteer as having less tainted mo- imperative (Kant, 1781/1908). Finally, Study 4 finds that motivated tives in the uncertain reward condition (p<.001, d=.40). However, reasoning determines which form of inconsistency that evaluators this result actually reversed when we told participants that the stu- evoke when making hypocrisy judgments. dent was $400 short on his rent (p<.01, d=.27; interaction p<.001). Study 1’s participants (N = 301) read a scenario involving a Presumably, this additional information attenuated the inference that consumer, Jeff, who became a vegetarian for environmental reasons. the student would have been willing to volunteer if no reward were In the Control condition, the scenario simply ended. In the other two offered. conditions, participants read that Jeff learns that environmentalists Finally, Study 7 (n=449) asks whether revealing a prosocial recommend the public eat a specific invasive fish species in order initiative’s uncertain profit outlook can earn more credit than not to protect local habitats. Shortly thereafter, Jeff finds himself at a mentioning profits at all. Here, participants rated a bank’s decision restaurant that has this fish species on the menu. In the Commitment- to invest in urban revitalization as more purely motivated when it based inconsistency condition, Jeff orders the fish, thereby acting in faced an uncertain profit forecast relative to both a certain profit fore- a manner inconsistent with his past commitment to vegetarianism. cast (p<.001, d=.57) and no mention of profits whatsoever (p<.01, In contrast, in the Goal-based inconsistency condition, Jeff orders d=.38). the vegetarian option, thereby failing to maximally help the environ- Our findings challenge conventional wisdom that extrinsic ben- ment. Consistent with a “Damned Either Way” effect, we find that, efits always lead to cynical motive inferences and deepen our under- relative to the control condition (M = 1.96), participants judged Jeff standing of how observers evaluate actions taken when rewards are to be more of a hypocrite regardless if he ordered the fish (M = 2.79, uncertain. p < .001) or the vegetarian option (M = 2.76, p < .001). In Study 2 (pre-registered: https://aspredicted.org/blind. Damned Either Way: Hypocrisy Judgments When Goals php?x=m6ab5q, N = 311), we replicate these findings in a within- and Commitments Conflict subjects design and utilizing three different scenarios to test the effect. In all three scenarios, we find a significant effect that behav- EXTENDED ABSTRACT ioral-based inconsistency leads to judgments hypocrisy relative to While those who behave ethically are often rewarded with repu- control (ps < .001), while in two of the three scenarios we find a tational benefits (e.g., Flynn et. al. 2006), behaving ethically can be Advances in Consumer Research (Volume 48) / 1099 significant effect of goal-based inconsistency leading to attributions group norms – those who chose to harm their own group rather than of hypocrisy relative to control (ps < .001). concede resources to the opposition believed others in their group In Study 3 (pre-registered: http://aspredicted.org/blind. would do the same. Changing individuals’ beliefs about the norm php?x=3uf789, N = 600), we ask whether the Damned Either Way changed their behavior. effect depends on the stated goal. In particular, we compared reason- In our primary study (preregistered: http://aspredicted.org/ based goals (e.g., eating fish to reduce environmental harm) versus blind.php?x=sx6vb7, N = 800, MTurk), participants were asked to categorical imperatives (e.g., refusing to eat fish because it is inher- modify donation amounts for three important and polarizing causes: ently wrong to do so). Consistent with our pre-registered hypoth- abortion access, gun control, and political parties. We first measured esis, we replicate the DEW effect in the reason-based goal condition: the degree to which participants identified with each side of the three relative to the control (M = 1.86), the actor was again judged to be issues. Participants in the counter-attitudinal condition were asked to a more of a hypocrite regardless if he ate the meat (M = 3.24, p < choose between subtracting $1 from a donation to an organization .001) or the vegetarian (M = 2.52, p < .001) option. However, we do on their side of the cause or adding $1 to a donation to an organi- not find a DEW effect when the actor’s behavior is driven by a cat- zation on the opposing side. Given these two unfavorable options, egorical imperative. In this case, relative to the control (M = 2.20), most participants (71%) chose to subtract $1 from the organization he was only considered a hypocrite when he chose to eat meat (M = on their side. This choice to hurt one’s own organization, rather than 4.16, p < .001), but not when he ate the vegetarian option (M = 1.99, help an opposing organization, conflicts with a harm minimizing p = .30). strategy, as individuals rate organizations on their side as more ef- In Study 4 (pre-registered: http://aspredicted.org/blind. fective with funds (p < .001). The implication is that the participants’ php?x=cp87p5, N = 806) we test the Damned Either Way effect in own cause suffers greater overall harm when their organization is a new context, and further show that both commitment- and goal- deprived of funding than when the opposing organization is given based hypocrisy judgments are shaped by motivated reasoning. Par- that same funding. ticipants first reported attitudes towards illegal immigrants. They Participants in the pro-attitudinal condition were given opposite then evaluated a scenario in which a man who commits to oppos- choices: to add $1 to a donation to an organization on their side of ing immigration for economic reasons learns that immigrants in fact the cause or to subtract $1 from a donation to the opposing side. The benefit the economy. A canvasser accepting donations in support of majority of participants (75%) chose to help their own side, indi- immigrant rights then approaches him. The individual either donates cating that choices do not stem from a preference to simply hinder (commitment-based inconsistency) or does not donate (goal-based the opposition above all else. The results from both conditions also inconsistency). A control condition was also included to assess base- hold at various levels of aggregation, including for each of the three line ratings of hypocrisy. We additionally reversed the scenario such causes individually, as well as for both sides of each cause. We find that the individual who supports immigrant rights for economic rea- that the choice to help (vs. harm) is moderated by the strength of parsons learns that immigrants are harmful to the economy. A canvasser ticipants’ attitudes toward the cause – those with stronger attitudes accepting donations in support of deporting illegal immigrants then are more likely to choose to harm their side (vs. help the other) in the approaches him. counter-attitudinal condition, and help their side (vs. harm the other) Results show a Damned Either Way effect across both versions in the pro-attitudinal condition. of the scenario (ps < .03). We further find that judgments of hypoc- An additional study (N = 403, MTurk) elicited the amount of risy were moderated by participants’ own attitudes about immigra- money participants would prefer to be subtracted from a donation to tion. Specifically, participants were more likely to judge the actor their side of a cause, in order to feel indifferent towards a $1 increase as a hypocrite if his final actions contradicted their personal beliefs in a donation to the opposing side. After removing large outliers, we regardless of the nature of the inconsistency they judged (ps < .06). conservatively estimated the average amount to be $2.33 (compared In sum, we find evidence that both commitment-based and goal- to $1: p < .001), suggesting a preference to harm one’s side to a based inconsistency leads ethical actors to be judged as hypocrites. greater magnitude than helping the opposition. To investigate the role of group norms in counter-attitudinal The Cost of Opposition: Preferring to Punish our Own choices, we ran a study (N = 653, MTurk) manipulating the norms Rather than Help our Opponent towards helping the opposition. In the control condition, participants were informed that the experimenter would be making a donation to EXTENDED ABSTRACT a pro-life and a pro-choice organization, and that they would have to Imagine that you are pro-choice and are given two options: add choose how to alter the amount. In the norm-add condition, partici- $5 to a donation going to a pro-life organization or subtract $5 from pants were also told that in a previous study, 70% of MTurkers on a donation going to a pro-choice organization. Would you choose their side of the cause chose to add to the opposing group rather than to help the opposing group or harm your own? In three studies (and subtract from their own. In the norm-subtract condition, participants two supplementary studies), we find that when given two options were instead told that 70% of previous participants on their side of – to help the opposition through greater funding or hurt one’s own the cause had chosen to subtract from their own group rather then side through diminished funding – research participants consis- add to the opposing group. The proportion of participants choosing tently choose to harm their own cause. This replicates across polar- to add $1 was less than 50% in the control group (39.2%, p =.002), ized contexts: abortion access, gun control, and political affiliation. and was not significantly different from the proportion in the norm- Moreover, we find that individuals prefer to harm their own interests subtract group (36.7%, p = .68), suggesting that subtracting is the despite the fact that they generally believe their own cause is more norm. However, there was a significant increase in the proportion effective with funds. of participants choosing to add $1 in the norm-add group, relative We propose that that the preference to harm one’s own group to both the control and norm-subtract conditions (57.7%, ps < .001), rather than help an opposing group stems from the belief that help- suggesting that behavior in this context is guided by perceptions of ing the opposition is a stronger violation of group norms. We find group norms. evidence that individuals act in accordance with their beliefs about 1100 / When Beneficence Backfires: Negative Consequences of Doing Good

When making decisions in polarized contexts, individuals do Hardy, Charlie L., and Mark Van Vugt (2006), “Nice guys finish not always make choices to minimize harm. Our findings add to a first: The competitive altruism hypothesis,” Personality and literature on the power of social influence and how psychological Social Psychology Bulletin, 32(10), 1402-1413. barriers can impede the advancement of a cause. In contexts in which Holly, Peter (2015), “The Red Cross Had $500 Million in Haitian accommodating two groups’ desires is crucial for progress, how do Relief Money, but It Built Just 6 Houses,” The Washington we compromise when both sides would rather harm their own cause Post, https://www.washingtonpost.com/news/morning-mix/ than make concessions in which the opposition benefits? Given our wp/2015/06/04/the-red-cross-had-500-million-in-haitian- highly polarized political system, determining how to shift these relief-money-and-it-built-just-6-houses/. norms is critical. Otherwise, paradoxically, the individuals who care Joireman, Jeff, Yany Gregoire, Berna Devezer, and Thomas the most may be most likely to forestall the other group’s progress at M. Tripp (2013), “When Do Customers Offer Firms a the expense of their own cause. ‘Second Chance’ Following a Double Deviation? The Impact of Inferred Firm Motives on Customer Revenge and REFERENCES Reconciliation,” Journal of Retailing, 89(3), 315– 37. Aaker, Jennifer, Kathleen D. Vohs, and Cassie Mogilner (2010), Kahneman, Daniel and Amos Tversky (1979), “Prospect Theory: “Nonprofits Are Seen as Warm and For-Profits as Competent: An Analysis of Decision under Risk,” Econometrica, 47(2), Firm Stereotypes Matter,” Journal of Consumer Research, 263-292. 37(2), 224-237.‏ Kant, Imannuel. (1908). Critique of pure reason. 1781. Modern Bhattacharjee, Amit, Jason Dana, and Jonathan Baron (2017), Classical Philosophers, Cambridge, MA: Houghton Mifflin, “Anti-Profit Beliefs: How People Neglect the Societal Benefits 370-456. of Profit,” Journal of Personality and Social Psychology, Koehn, Daryl (2013), “Why Saying “I’m Sorry” Isn’t Good 113(5), 671-696. Enough: The Ethics of Corporate Apologies,” Business Ethics Berman, Jonathan Z., et al. (2015), “The Braggart’s Dilemma: On Quarterly, 23(2), 239-268.‏ the Social Rewards and Penalties of Advertising Prosocial Miller, Dale T. (1999), “The Norm of Self-Interest,” American Behavior,” Journal of Marketing Research, 52(February), Psychologist, 54 (12), 1053–60. 90-104. Minson, Julia A. and Benoit Monin (2012), “Do-gooder derogation: Chernev, Alexander and Sean Blair (2015), “Doing well by doing Disparaging morally motivated minorities to defuse good: The benevolent halo of corporate social responsibility,” anticipated reproach,” Social Psychological and Personality Journal of Consumer Research, 41(6), 1412-1425. Science, 3(2), 200-207. Critcher, Clayton R. and David Dunning (2011), “No Good Deed Monin, Benoit and Anna Merritt, A. (2012), “Moral hypocrisy, Goes Unquestioned: Cynical Reconstruals Maintain Belief in moral inconsistency, and the struggle for moral integrity.” the Power of Self-Interest,” Journal of Experimental Social Oliver, Richard L. (1977), “Effect of Expectation and Psychology, 47(6), 1207–1213. Disconfirmation on Postexposure Product Evaluations - an Effron, Daniel A., et al. (2018), “From inconsistency to hypocrisy: Alternative Interpretation,” Journal of Applied Psychology, When does “saying one thing but doing another” invite 62(4), 480–486. condemnation?” Research in Organizational Behavior, 38, Pace, Kristin M., Tomasz A. Fediuk, and Isabel C. Botero (2010), 61-75. “The Acceptance of Responsibility and Expressions of Regret Flynn, Francis J. (2003), “How much should I give and how often? in Organizational Apologies After a Transgression,” Corporate The effects of generosity and frequency of favor exchange Communications: An International Journal, 15(4), 410-427.‏ on social status and productivity,” Academy of Management Reimann, Martin, Deborah J. MacInnis, Valerie S. Folkes, Arianna Journal, 46(5), 539-553. Uhalde, and Gratiana Pol (2018), “Insights into the Experience Flynn, Francis J., Ray E. Reagans, Emily T. Amanatullah, and of Brand Betrayal,” Journal of the Association for Consumer Daniel R. Ames (2006), “Helping One’s Way to the Top: Research, 3(April), 240–54. Self-Monitors Achieve Status by Helping Others and Knowing Sen, Sankar, and Chitra Bhanu Bhattacharya (2001), “Does Doing Who Helps Whom,” Journal of Personality and Social Good Always Lead to Doing Better? Consumer Reactions Psychology, 91(6), 1123–37. to Corporate Social Responsibility,” Journal of Marketing Gregoire, Yany, and Robert J. Fisher (2008), “Customer Betrayal Research, 38, 225–243. and Retaliation: When Your Best Customers Become Your Small, Deborah A., and Cynthia Cryder (2016), “Prosocial Worst Enemies,” Journal of the Academy of Marketing consumer behavior,” Current Opinion in Psychology, 10, 107- Science, 36 (June), 247–61. 111. Unanticipated Consequences of Constraining Behavior Chairs: Jackie Silverman, University of Delaware, USA Shalena Srna, University of Michigan, USA

Paper #1: When Limits Backfire: The Ironic Effect of Setting that has been budgeted for a specific purpose than money that has not Limits on Entertainment Consumption been budgeted because the loss is psychologically realized when the Shalena Srna, University of Michigan, USA money is initially mentally allocated. Finally, Paper 4 explores the Jackie Silverman, University of Delaware, USA effects of having the choice to pre-commit to a savings program. Spe- Jordan Etkin, Duke University, USA cifically, a field experiment finds that allowing employees to choose between immediate or delayed enrollment (vs. immediate enrollment Paper #2: Spending Gained Time only) did not affect enrollment rates but led employees to delay en- Selin Malkoc, Ohio State University, USA rollment, thus saving less overall. Gabriela Tonietto, Rutgers University, USA Together, these four papers highlight the importance of under- Paper #3: Loss Booking: Mental Accounting Facilitates standing how consumers manage their time and money, document- Consumer Spending ing ironic psychological processes and unintended negative conse- Chang-Yuan Lee, Boston University, USA quences of multiple popular resource management tools. In line with Carey Morewedge, Boston University, USA the conference theme, the differing experimental approaches and Paper #4: Save More Later? The Roles of Present Bias and research questions addressed within this session would provide a Perceived Urgency “rendez-vous” for researchers, policy-makers, and practitioners who Joseph Reiff, University of California, Los Angeles, USA are interested in a wide variety of topics, including time manage- John Beshears, Harvard University, USA ment, mental accounting, savings, goals, and how new technologies Hengchen Dai, University of California, Los Angeles, USA impact wellbeing and decision-making. Katherine Milkman, University of Pennsylvania, USA, USA Shlomo Benartzi, University of California, Los Angeles, USA When Limits Backfire: The Ironic Effect of Setting Limits on Entertainment Consumption SESSION OVERVIEW Consumers have limited time and money that they can spend EXTENDED ABSTRACT in an infinite number of ways. Deciding how to best spend one’s re- New technologies have made it easier than ever before to con- sources involves tradeoffs between short-term desires and long-term sume entertaining content. On average, people spend nearly 2.5 investments. To help people make better choices, individuals and hours per day on social media alone—over half of which occurs dur- firms frequently impose constraints on how consumers spend their ing the workday. To help people better manage their time, many tech- time and money (Cheema and Soman 2006; Gollwitzer and Sheeran nological devices (e.g., iPhones; Nintendo Switch) and popular apps 2006; Shefrin and Thaler 1988; Thaler and Bernartzi 2004). For ex- (e.g., Facebook, Instagram, YouTube) have introduced time-tracking ample, people often self-impose deadlines to reduce procrastination tools, providing consumers with the option to set costless, non-bind- or create budgets to encourage fiscal responsibility (e.g., Ariely and ing time limit reminders on entertaining activities. These reminders Wertenbroch 1999; Cheema and Soman 2006). This session seeks alert consumers to having spent a pre-determined amount of time to better identify and understand the unanticipated consequences of on an activity or platform each day (e.g., 45 minutes on Instagram). constraining such behaviors. Four papers answer questions about Presumably, people set time limits on entertaining, but unpro- how different types of constraints (i.e., time limits, schedules, bud- ductive, activities in order to encourage themselves to spend less gets, and pre-commitment devices) can have unanticipated and even time on those activities (and more time on productive work). Con- negative outcomes for how consumers spend their time and money. trary to this intended effect, we propose that such limits can backfire, The first two papers investigate how consumers seek and man- increasing, rather than decreasing, time spent on the activities they age time spent on “leisure,” exploring how constraining time spent are meant to constrain. We argue that this counterintuitive effect oc- through setting limits and scheduling can have surprising conse- curs because, much like when consumers set goals (e.g., Heath, Lar- quences. Paper 1 examines how a common time management tool— rick and Wu 1999; Wallace and Etkin 2018), setting limits changes setting an alert after a pre-determined amount of time spent (akin salient reference points. Without a time limit, consumers use their to the option on most smartphones)—affects how much time people starting point (i.e., spending no time on the activity) as the reference spend on the activity. Results show that setting such non-binding time point by which to judge their (dis)satisfaction with the time spent. limits on entertaining activities, like social media or games, counter- Once a time limit has been set, however, the focal reference shifts to intuitively increases the time spent on them. Paper 2 finds that, even this limit amount (e.g., 45 minutes), making any time spent under the though consumers desire more leisure time, when they actually gain limit amount feel more acceptable (i.e., less like a loss or a waste of time that could be spent on leisure activities, they show more inter- time) and overall discouraging restraint. Five pre-registered studies est in work. This is because gaining time (versus having free time) examining both external and self-set time limits on different enter- prompts individuals to question how to best allocate their time, lead- taining activities (browsing social media, playing an online game) ing them to consider – and address – their “to-do” list more. support these predictions. The next two papers examine counterintuitive effects of con- In Study 1a (N = 208), participants imagined that they were reg- straining another important resource – money – by exploring how ular users of a social media app and that they had spent 52 minutes budgeting money and pre-committing to future savings can backfire, on the app yesterday. Participants were randomized to either imagine increasing spending and reducing savings. Paper 3 finds that mental that they had set a limit on their phones, such that they would see a budgeting, which is intended to inhibit spending, actually encourages reminder of their time spent at 60 minutes, or not. We then asked par- it. Results show that consumers are more inclined to spend money ticipants to rate on -10-to-10 scales their satisfaction with how they

Advances in Consumer Research 1101 Volume 48, ©2020 1102 / Unanticipated Consequences of Constraining Behavior had spent their time yesterday (4 items: how happy they felt about their unpaid work, buying services like childcare and housekeeping the time they spent on the app, how much they felt they had lost/ (Whillans et al. 2017). These trends imply an increase in leisure con- gained time, how well they used their time, and how much they felt sumption. However, hours dedicated to socializing have decreased they had wasted their time [reverse-coded]). A one-way MANOVA over the last decade (ATUS 2017), potentially because when feel- found a main effect of condition across these four items (F(4, 203) ing busy, leisure activities are often jettisoned (Southerton 2003). = 6.27, p < .001). An independent t-test revealed that participants We study this paradox by examining how people consume gained who set a limit reported felt significantly better about how they spent time (due to a cancelation or plan change) and propose that despite their time (M = 0.01) than participants who did not set a limit (M = a desire for more leisure, consumers may choose to work when they -1.60; t(206) = 3.46, p < .001). Study 1b (N = 807) replicates this actually gain time. effect; participants felt better about the time they spent onsmart- Consumers desire free time, which often means time free from phone games (either 18 or 52 minutes) when they imagined having obligations or generally “leisure” (Unger and Kernan 1983). Howev- set a daily limit of 60 minutes (M = 2.97) versus when they did not er, we argue that despite being objectively free of obligations, gained imagine setting a limit (M = 0.91; t(805) = 7.06, p < .001). Together, time is not automatically classified as “free.” Instead, it is assigned to Studies 1a and 1b suggest that people do indeed shift their reference an “available” time account, which spurs consumers to question how point regarding how much time they should spend on entertainment to allocate that time. In doing so, consumers turn to their running in response to setting a limit. list of tasks requiring their time (i.e., to-do list), which are usually Our remaining studies directly test the effects of setting a limit work-oriented. As an outcome, the more salient work-like tasks take (versus not) on real consumption behavior (i.e., actual time spent on precedence and otherwise desired leisure activities are abandoned. an entertaining activity). In Study 2, participants (N = 199) were giv- Study 1 demonstrates this paradox of desiring leisure, but en a total of 5 minutes to spend on an entertaining activity (a bubble choosing work during gained time. We compared chosen tasks dur- shooter game) and a paid work task (transcribing text strings). They ing a gained time with those for an unspecified future date. Unlike could spend as much time as they wanted on the game before switch- the present or near future, the (unspecified) future is often decontex- ing to the work task. Half of participants read that, to help them man- tualized from our daily demands (Zauberman and Lynch 2005). As age their time, a reminder would appear after spending three min- such, we expected participants to treat an unspecified future time as utes on the game. Importantly, after seeing the reminder, they could free (and meant for leisure), while treating a particular gained hour continue to engage in the game for as long as they liked. The other as simply unaccounted (prompting a need to account for it). half of participants did not read about or receive such a limit. As pre- 175 undergraduates imagined either gaining a few hours in the dicted, an independent t-test on log-transformed time spent revealed next couple of weeks or later in the same day and indicated how they that participants in the limit condition spent more time on the game would spend this time (1=definitely leisure; 9=definitely work). We (M = 117.21 seconds) than those in the control (M = 94.77; t(197) = also measured participants’ perceived time famine and their desire 2.59, p = .010). for more work and leisure in their life. Unsurprisingly, participants Study 3 (N = 192) replicated Study 2 with a different entertain- reported experiencing time famine (M=4.79; p<.01 compared to the ing activity (scrolling through Pinterest), again finding that setting a scale midpoint) and wished they had more leisure (M=5.06) than limit increased the time spent compared to the control (Mlimit = 24.48 work (M=3.83, p<.01). Further, reported time famine was positively seconds vs. Mcontrol = 15.88; t(190) = 2.63, p = .009). correlated with desire for more leisure (r = .28, p < .01), but uncor- Study 4 (N = 201) tested if self-set limits would have a similar related with desire for more work (r = -.03, p = .63). Thus, it appears effect on behavior. It used the same design as Study 3, but allowed that participants crave more time to increase leisure consumption. participants to choose their own limit. In particular, participants were However, we found that when participants considered gaining a told that, to help them manage their time during the study, they could particular period of time, they preferred work-like tasks (M=7.27) set a non-binding time limit, which would appear as a reminder dur- compared to those who considered gaining an unspecified future ing the Pinterest activity. The limit could be set in 15-second incre- time (M=5.27, p<.01). Importantly, participants’ desire for more lei- ments for up to the full 5 minutes (i.e., the equivalent of not setting a sure did not predict their intended behavior in either condition (both limit; Mself-set limit = 83.96 seconds, SD = 92.65, Median = 60). As pre- p’s>.90). dicted, participants in the limit condition spent significantly longer In Study 2 we directly compared having 1.5 hours of free browsing Pinterest (M = 23.52 seconds) than those in the control (M time to gaining 1.5 hours. 153 MTurkers reported how they would = 14.84; t(199) = 2.03, p = .044). Importantly, this effect casts doubt consume this time (1=fun and enjoyable task; 9=work-related on a potential alternative explanation based on externally imposed task). As expected, we found that participants were more likely to limits signaling how long people should spend on the task. choose work-like tasks during gained time (M=3.17) than free time In sum, despite intending to reduce the time spent, setting limits (M=2.44, p<.05). can ironically increase time spent on unproductive entertaining ac- In Study 3, we examined whether differences in construal level tivities (over paid work). This occurs because limits change consum- (Trope and Liberman 2010) account for our results by manipulat- ers’ salient reference point when evaluating their use of time. Our ing temporal distance. 342 undergraduates imagined having/gaining findings contribute to emerging research on the unintended conse- 1.5 hours either that evening or the same evening in one week and quences of tracking behavior (Etkin 2016; Soman and Cheema 2004) indicated their task preference (1=catch up with friend/family/neigh- and how new technologies impact wellbeing and decision-making bor; 9=catch up on errands/housework/chores). We predicted that a (e.g. Melumad and Pham 2020). particular period of time gained in distant future is similar to time gained sooner as they both define a specific time period and require Spending Gained Time its accounting. As expected, participants who gained time preferred a work-like task (M = 5.46) compared to those who had time free EXTENDED ABSTRACT (M=4.71, p<.01). Importantly, there was no main effect nor interac- Working hours have steadily declined over the last century tion of temporal distance (both p’s>.2). (Huberman and Minns 2007), and consumers increasingly outsource Advances in Consumer Research (Volume 48) / 1103

The remaining studies test our proposed framework: gaining We posit loss booking an important case in which expectations time prompts individuals to account for this time, leading them to serve as the reference point by which consumers evaluate outcomes reference their to-dos, and to choose work. First, in Study 4, we add- (Kőszegi 2006). This suggests that the psychological loss associated ed a third condition in which participants were told that a unit of free with the expense occurs when the mental budget is created for antici- time was “unaccounted.” In line with the idea that free time is often pated expenses rather than at the time when the goods are purchased. equated with leisure, but that prompting consumers to think of how Consequently, the psychological impact of the loss associated with to account for time increases propensity to choose work, participants the purchase of goods within the mental budget is mitigated or di- in both the gained time (M=3.68) and unaccounted time conditions minished. Specifically, loss booking theory predicts that losses are (M=3.89) were more likely to choose work than the free time condi- psychologically realized when consumers budget for certain expens- tion (M=2.82, ps<.05). es (H1). Due to the prior realization of losses, consumers experience Importantly, if gaining time indeed increases prioritization of less pain of paying for expenses associated with existing mental bud- work, then the effect should be exaggerated for consumers who have gets (H2). Finally, the reduced pain of paying incurred when pur- a tendency to work first and save leisure for later. Testing this, in chasing goods associated with an existing mental budget facilitates

Study 5, 202 participants imagined having or gaining 3 hours and in- spending within that budget (H3). dicated their activity preference (1=fun and enjoyable task; 9=work- We report five experiments providing evidentiary support for related task) and their desire to prioritize work (e.g., I’d rather get our hypotheses. Experiment 1 illustrates the effect of mental bud- some work tasks out of the way). As before, participants who gained geting on actual holiday spending. Experiments 2-5 used vignette time were more likely to prefer a work-like task (M=2.83) than those designs. Experiment 2 provides evidence for “loss booking” (H1). who had time (M=2.17, p < .05). Importantly, this was moderated by Experiment 3 supports the hypothesis that due to the loss realiza- participants’ desire to prioritize work, where participants who pri- tion, the subsequent consumption of that money elicits less pain of oritize work to a greater extent (-.27 standard deviations from the paying than would the expenditure of unbudgeted money or money mean or higher) showed greater preference to work during gained budgeted towards other accounts (H2). Experiment 4 shows that this (vs. free) time. hypothesis holds when the budgeting process is implicit. Finally, Finally, in Study 6, we tested the proposed role of to-do lists. Experiment 5 shows that this reduced pain of paying facilitates the In addition to indicating how they would spend gained versus free purchase of expenses associated with the mental budget, even when time (1=definitely leisure/fun; 7= definitely work/productive), par- the purchase exceeds the budget limit (H3). ticipants (N=374) also provided their running to-do list for the day. Experiment 1 (N=328) was conducted in two phases on separate Analyses revealed a main effect of gained time, whereby participants days. In phase 1, we manipulated the presence of a mental budget by were more likely to choose work during gained (M=6.98) compared assigning half of participants to explicitly make a holiday shopping to free time (M=6.31, p=.004). There was also a significant interac- budget for the upcoming Black Friday weekend. Six days later (i.e., tion with the number of to-dos. Those with more tasks on their toin phase 2), participants reported how much money they actually do list (-.40 standard deviations from the mean or higher) showed spent. We found that participants in the budget condition spent sig- greater propensity to choose work during gained (vs. free) time. nificantly more money than did controls, who were not asked how In sum, despite their desire for more leisure, consumers may much they planned to spend. Experiment 2 (N=244) manipulated paradoxically prefer to work when they gain time. This is because whether participants made a clothing budget before they went shop- while gaining time creates free time, gained time is categorically dif- ping, and whether they purchased a t-shirt. We measured the extent ferent from free time. to which participants perceived themselves to have lost or gained money during the shopping trip. We found that the purchase of a shirt Loss Booking: Mental Budgeting Facilitates Consumer without a budget was evaluated as a loss (i.e., loss framing), whereas Spending the purchase of a shirt that did not exceed the budget was evaluated as a gain. In addition, whereas participants without a budget per- EXTENDED ABSTRACT ceived themselves to have neither gained nor lost money, those with Budgeting is a common money management practice. Approxi- a budget perceived themselves to have gained money if they did not mately 41% of Americans prepared a written or computerized budget make a purchase every month (U.S. Bank Possibility Index, 2016). It has a mental Experiment 3 (N=300) manipulated whether or not participants analog, mental budgeting, whereby mental labels associated with made a clothing budget (vs. food budget or no budget) before they money change how it is spent (Heath and Soll 1996; Thaler 1985). purchased a pair of jeans. We measured the extent to which partici- People create spending plans, for example, by earmarking funds into pants perceived the purchase of jeans as a monetary loss, and the various mental accounts (e.g., groceries, clothing). Once money is pain of paying associated with their purchase. We found that par- budgeted into mental accounts, it is no longer perfectly fungible, or ticipants were less likely to perceive purchasing the jeans as a loss substitutable, with other money. in the clothing budget condition than in the control and food budget Mental budgeting is often characterized as a self-control de- conditions. Moreover, the difference between the control and food vice that inhibits consumer spending (Heath and Soll 1996; Thaler budget conditions was not significant. A similar pattern of results 1985). In this research, however, we propose a loss booking theory was observed for the pain of paying. A mediation analysis revealed to reinterpret mental budgeting––that creating budgets for expenses the mediating impact of loss framing on pain of paying. facilitates rather than inhibits consumer spending. The intuition be- Experiment 4 (N=100) manipulated whether participants with- hind the concept of loss booking is that consumers psychologically drew cash at ATM (i.e., implicit budgeting) in advance of expendi- realize or “book” the loss of the money when it is budgeted into a tures (controls imagined that they had the same amount of cash in mental account. This implies that many fungibility violations occur their wallet). In addition to the two measures used in Experiment 3, because losses are realized for the expenses when the mental budget we measured the extent to which participants expected to spend the is “booked,” and would be realized again if that money were spent cash that they had. We found that participants in the withdrawal con- outside the budget. dition (a) exhibited a higher expected propensity to spend, (b) were 1104 / Unanticipated Consequences of Constraining Behavior less likely to perceive the expenditure as a loss, and (c) experienced creased the number of programs they would enroll in (95% CI of in- less pain of paying than participants assigned to the control condi- direct effect=[.02,.07]) but led people to infer that the employer less tion. We found the serial mediating impact of expected propensity to urgently recommended these programs, which reduced the number spend and loss framing on pain of paying. of enrolled programs (95% CI=[-.09,-.04]). These competing forces Experiment 5 (N=309) manipulated whether participants bud- could explain a null effect on program enrollment (p=0.40). Further- geted money to buy sunglasses ($90) in advance of their purchase, more, lower perceived urgency reduced the number of programs that and whether their price was under or over that budget (i.e., $69 vs. people chose to enroll in immediately (95% CI=[-.07,-.02]). $110). We teased apart the budgeting and reference price accounts in To reconcile our findings with the Save More Tomorrow explaining spending by including a reference price condition. More- (SMarT) intervention, we compared our choice sets with SMarT in over, we measured purchase likelihood. We found that in both price a preregistered study (N = 1,499 MTurkers). Participants imagined conditions, participants in the budget condition (a) were less likely receiving a mailing about their company’s retirement savings plan. to perceive the purchase of sunglasses as a loss, (b) experienced less They were randomly assigned to one of three conditions. In the im- pain of paying associated with its purchase, and (c) more likely to mediate commitment condition, participants only had the option to buy the sunglasses than participants in the control and reference enroll immediately. In the simultaneous pre-commitment condition price conditions. A mediation analysis revealed the serial mediating (mimicking the simultaneous pre-commitment condition in our field impact of loss framing and pain of paying on purchase likelihood experiment), they had the option to either enroll immediately or across prices. enroll in four months. In the sequential pre-commitment condition (mirroring the classic design of SMarT), they were first given the Save More Later? The Roles of Present Bias and option to enroll immediately, and then if they declined, they were of- Perceived Urgency fered the option to enroll in four months. Consistent with our theory, participants in the immediate commitment condition viewed their EXTENDED ABSTRACT company as conveying a stronger recommendation to save (M=5.10, Previous research has shown that choice architects can increase SD=1.42) than participants in the simultaneous pre-commitment con- farsighted behavior by offering the option to pre-commit (e.g., Milk- dition (M=4.77, SD=1.58), p<0.001). Moreover, participants in the man, Rogers, and Bazerman 2009)we analyze the film rental and sequential pre-commitment condition inferred a marginally stronger return patterns of a sample of online DVD rental customers over a recommendation to save than participants in the immediate com- period of four months. We predict and find that should DVDs (e.g., mitment condition (Msequential=5.27; Mimmediate commitment =5.10; p=0.06), documentaries. People are more likely to save for retirement, for in- which may offer a potential explanation for why SMarT is able to stance, when offered the option to start saving in the future, rather capitalize on the power of pre-commitment without backfiring. than today (Thaler and Benartzi 2004). We reexamine this idea and We tested our full theory in an incentive compatible experi- show when and why the strategy can backfire. ment (N = 2,399). Following a filler task, participants were offered In a field experiment, 5,196 employees at four US universities an optional, unpaid ten-minute financial well-being assessment. Par- were sent a mailing that encouraged them to join a savings plan. ticipants were randomly assigned to either be offered (1) the option Employees were randomly assigned to receive either: (1) the im- to take the assessment now (the immediate commitment condition), mediate commitment mailing, which offered employees an opportu- (2) the option to take the assessment either now or in 1 week (the nity to immediately enroll in a savings plan or (2) the simultaneous simultaneous pre-commitment condition), or (3) the option to take pre-commitment mailing, which gave employees the choice to either the assessment now followed by the option to take it in 1 week if immediately enroll or pre-commit to enroll months later. Contrary they declined (the sequential pre-commitment condition). Consistent to prior work, offering pre-commitment reduced the likelihood of with our theory, participants that were offered sequential pre-com- immediate enrollment by 21.4% (p=0.015) with no effect on whether mitment were more likely to enroll (57%) compared to participants or not people enroll at any time (p=0.25), causing people to save 6% offered simultaneous pre-commitment (48%) or immediate commit- less on average over the eight-month study period (p=.007). ment (32%). Participants offeredsimultaneous pre-commitment were We theorize that offering a pre-commitment option along with marginally less likely to enroll immediately (28%) than participants an immediate enrollment option (1) reduces the anticipated costs of offeredimmediate commitment (32%; p = .07). This may happen be- doing the farsighted behavior (consistent with prior research on pres- cause, while both sequential pre-commitment and simultaneous pre- ent bias) but simultaneously (2) leads people to infer that the choice commitment made the farsighted behavior (taking a financial-well- architect less urgently recommends the behavior. These opposing being assessment) seem less costly (Msequential=4.2 & Msimultaneous=4.2 forces together can explain why pre-commitment can yield a null VS. Mimmediate commitment =4.5), sequential pre-commitment increased the effect on people’s willingness to engage in the farsighted behavior. perception that the action is urgently recommended and simultane- Meanwhile, lack of perceived urgency may lead people to choose to ous pre-commitment decreased this perceived urgency (Msequential=4.2 delay the behavior; thus, offering pre-commitment may cause people VS. Mimmediate commitment =3.9 VS. Msimultaneous=3.6). Our mediation analy- to opt to do it later instead of now. ses confirmed that while anticipated costs always positively medi- We tested this theory in a preregistered study where 1,161 ated the relationship between receiving a pre-commitment option MTurkers imagined that they just joined a company and had to and likelihood of enrolling in the assessment, perceived urgency choose whether to enroll in any of three employer-sponsored benefits negatively mediated this relationship when pre-commitment was of- programs. Participants were randomly assigned to either the immedi- fered simultaneously and positively mediated when it was offered ate commitment condition where they chose which programs they sequentially. would enroll in immediately or the simultaneous pre-commitment In sum, we develop and test a new theory about how different condition where they had the option to enroll in each program either forms of pre-commitment nudges impact whether and when people immediately or in six months. Consistent with our theory, the re- engage in farsighted behavior. While prior research suggests that pre- sults suggest that adding a pre-commitment option (vs. only offering commitment nudges can increase farsighted behavior by reducing immediate enrollment) made enrolling feel less painful, which in- the anticipated costs of doing farsighted behavior, we document an Advances in Consumer Research (Volume 48) / 1105 overlooked mechanism; people draw inferences about the urgency Melumad, Shiri, and Michel Tuan Pham (2020), “The Smartphone with which the choice architect recommends the behavior based on as a Pacifying Technology,” Journal of Consumer Research, the sequence with which pre-commitment options are offered. More https://doi.org/10.1093/jcr/ucaa005. generally, the current research contributes to a growing literature Milkman, Katherine L., Todd Rogers, and Max H. Bazerman about decision makers’ inferences about the motives and implicit (2009), “Highbrow Films Gather Dust: Time-Inconsistent recommendations of choice architects (Krijnen, Tannenbaum, and Preferences and Online DVD Rentals,” Management Science, Fox 2017). Researchers and practitioners can anticipate and leverage 55(6), 1047-4059. these inferences to better design choice architecture that works as Shefrin, Hersh M. and Richard H. Thaler (1988),”The behavioral intended rather than backfiring. life‐cycle hypothesis,” Economic inquiry, 26(4), 609-643. Soman, Dilip, and Amar Cheema (2004), “When Goals are REFERENCES Counterproductive: The Effects of Violation of a Behavioral Ariely, Dan and Klaus Wertenbroch (2002), “Procrastination, Goal on Subsequent Performance,” Journal of Consumer deadlines, and performance: Self-control by precommitment,” Research, 31(1), 52-62. Psychological science, 13(3), 219-224. Bureau of Labor Southerton, Dale (2003). “’Squeezing Time’: Allocation practices, Statistics (2017), “American Time Use Survey”. coordinating networks and scheduling society,” Time and Cheema, Amar and Dilip Soman (2006), “Malleable mental Society, 12(1), 5-25. accounting: The effect of flexibility on the justification of Thaler, Richard H. (1985), “Mental Accounting and Consumer attractive spending and consumption decisions,” Journal of Choice,” Marketing Science, 4 (3), 199–214. Consumer Psychology, 16(1), 33-44. Thaler, Richard H. and Shlomo Benartzi (2004), “Save More Etkin, Jordan (2016), “The Hidden Cost of Personal Tomorrow: Using Behavioral Economics to Increase Quantification,”Journal of Consumer Research, 42(6), 967- Employee Saving,” Journal of Political Economy, 112(1). 984. Tonietto, Gabriela and Selin A. Malkoc (2016), “The calendar Gollwitzer, Peter M. and Paschal Sheeran (2006), “Implementation mindset: Scheduling takes the fun out and puts the work in,” intentions and goal achievement: A meta‐analysis of effects Journal of Marketing Research, 53(6), 922-936. and processes,” Advances in experimental social psychology, Trope, Yaacov and Nira Liberman (2010), “Construal level theory 38, 69-119. of psychological distance,” Psychological Review, 117(2), Heath, Chip, Richard P. Larrick, and George Wu (1999), “Goals as 440-463. Reference Points,” Cognitive Psychology, 38 (1), 79-109. Unger, Lynette S. and Jerome B. Kernan (1983), “On the meaning Heath, Chip and Jack B Soll (1996), “Mental Budgeting and of leisure: An investigation of the subjective experience,” Consumer Decisions,” Journal of Consumer Research, 23 (1), Journal of Consumer Research, 9(4), 381-392. 40–52. Wallace, Scott G., and Jordan Etkin (2018), “How Goal Specificity Huberman, Michael and Chris Minss (2007), “The times they are Shapes Motivation: A Reference Points Perspective,” Journal not changin’: Days and hours of work in Old and New Worlds, of Consumer Research, 44 (5), 1033-1051. 1870-2000,” Explorations in Economic History, 44, 538-467. Whillians, Ashley V., Elizabeth W. Dunn, Paul Smeets, Rene Kőszegi, Botond (2006), “Emotional Agency,” The Quarterly Bekkers, and Michael I. Norton (2017), Buying time promotes Journal of Economics, 121 (1), 121–55. happiness, PNAS, 114(32). Krijnen, Job M. T., David Tannenbaum, and Craig R. Fox (2017), Zauberman, Gal and John G. Lynch Jr. (2005), Resource slack and “Choice Architecture 2.0: Behavioral Policy as an Implicit propensity to discount delayed investments of time versus Social Interaction,” Behavioral Science & Policy, 3(2), i–18. money, Journal of Experimental Psychology: General, 134(1), 23-37. Shining A Light on Female Power: Women’s Experiences of Transcending Aging Chairs:Pilar Rojas Gaviria, Marketing Lecturer, University of Birmingham, UK Céline Del Bucchia, Associate Professor, Marketing, Audencia Business School, France

Paper #1: Reclaiming the Crone: Reimagining Old Age and to this special session by showing how the emerging themes associ- Feminine Power ated with the Crone invite us to reimagine old age and female power. Lorna Stevens, University of Bath, UK The second paper focuses on women’s narratives reflecting their Pauline Maclaran, University of Royal Holloway, UK experience of going through the menopause in the UK. Their work Olga Kravets, University of Royal Holloway, UK offers a theoretical reflection on how women can transcend their ‘unproductive’ body. Paper #2: ‘The Foreigner within us’: Catharsis amid Horror The third paper explores how the aid of material and technology and Confusion in Menopause can enhance value co-creation for elderly women with family mem- Pilar Rojas Gaviria, University of Birmingham, UK bers/friends and service providers. It illustrates how women stimu- Céline Del Bucchia, Audencia Business School, France late and leverage their interdependence in the context of the Elderly Camilla Quental, Audencia Business School, France Consumption Ensemble. Paper #3: Opportunities and Challenges for Aging Women in The fourth paper explores intergenerational transmission though CoCreating Value in the Elderly Consumption Ensemble the theoretical lens of wisdom. Their paper raises issues about how Lisa Peñaloza, Kedge Business School, Bordeaux, France grandmothers transcend aging thanks to the sharing of wisdom, and Carol Kelleher, University College, Cork, Ireland how this process helps them to build and maintain their relational Michelle Barnhart, Oregon State University, USA identity. Paper #4: Wisdom Examined via a Qualitative Investigation This special session aims to stimulate discussion and further of the Bathing Rituals of Aging Moroccan Women and their consumer research to shine light on the power of female aging. To- Granddaughters gether, the four papers raise two key questions: first, how can women Delphine Godefroit-Winkel, Toulouse Business School, enhance our power to transcend age by better understanding and uti- Casablanca Campus, Morocco lizing aspects of consumer culture, the market, and family dynam- Marie Schill, Université de Reims, Champagne Ardenne, ics? And second, how and what can women’s empowerment in aging France contribute to a more enlightened society? By gaining more insights Margaret K. Hogg, Lancaster University, UK into and leverage over the creative aspects of female aging, consumer researchers, policymakers and marketers are able to enhance SESSION OVERVIEW women’s well-being, negotiate brand positioning, develop communication campaigns, and thus advance the discourse of female aging in “It is a mistake to regard age as a downhill grade slide towards order to provide more positive and fairer representations. dissolution. The reverse is true. As one grows older, one climbs upwards with surprising strides.” - George Sand Reclaiming the Crone: Reimagining Old Age and Feminine Power While a large amount of research on aging consumers has focused on the paradigm of the privation that accompanies aging (Mo- EXTENDED ABSTRACT chis 2012), this special session focuses on how women transcend the stereotypes of aging and nurture their capacity to create and co-create “The rhythmic phases of the moon waxing and waning, light and alternative social imaginaries with respect to their feminine aging darkness, the crescent, sickle-shaped new moon, the full moon, process. The special session draws from the consumer research lit- and the waning, darkening moon – the maiden in the crescent, the erature, which investigates old age and new beginnings (Schau, Gilly mother in the full moon and, in the darkening moon, the wise old & Wolfinbarger, 2009) as well as gender issues (Steinfield, Coleman, woman whose light was within her. The Divine Trinity of the Mother Tuncay Zayer, Ourahmoune and Hein 2019). Goddess, as old as time itself, endlessly renewing itself – darkness, The four papers examine how women transcend aging and turn light, darkness - all aspects of the Mother Goddess to be embraced, it into a creative force. The papers complement each other in exam- expressing the unity of sky, earth, sea, and the rhythmic patterns of ining various facets of feminine aging, focusing on popular culture, nature and all living things.” (Baring and Cashford 1993: 63) the decline of the body, collective experiences in the context of intergenerational transmission, and the Elderly Consumption Ensemble Associated with the final stage in the cycle of life, The Crone (Barnhart and Peñaloza 2013) in a span of women’s life stages from was the third aspect of the Triple Goddess in ancient culture. Being middle life to old age. the wise old woman, the seer, the midwife, and the healer, she sym- Our aim in this session is to extend discussions on women, ag- bolised wisdom, knowledge and the underworld. In Irish mythology ing and consumption to counterbalance the traditional view of aging she was the Dark Morrigan, presiding over death in the battlefield, as a state of privation and decline, a moment of lost capacities (Szmi- taking back the bodies of the fallen ones into her eternal, primordial gin and Carrigan 2001). This belief has been to the disadvantage of embrace (Stevens and Maclaran 2007). In time the sacred associa- women in comparison to men who are more likely to benefit from tions of the Triple Goddess (Maiden, Mother, Crone) came to be de- positive associations attached to their aging process such as maturity sacralised and diabolized by male-centred religions. (Walker 1985). and charisma (Steinfield et al. 2019). They supplanted matri-centred societies and replaced the worship of The first paper focuses on cultural representations of the Crone nature and the Mother Goddess with the worship of culture and a as an empowering figure in popular culture. Their work contributes Male God and Son. The ‘crone’ came, in the Middle Ages, to be de-sacralised as an ugly, withered old woman, loathed and feared in

Advances in Consumer Research 1106 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1107 equal measure. She was witchlike, evil, cruel, cantankerous; a hag, lifetime of experience. ‘I feel like a snake, shedding its own skin,’ an old bag, an old bat (Collins Thesaurus); she was an object ‘usually I told a friend recently. On the surface I may look more or less the of contempt’ (Webster Dictionary). same. Inside, nothing is as it was.” (female writer’s testimonial, The contempt expressed for the post-menopausal woman, of- The Guardian, 2008_01_05) ten depicted as physically repulsive, argumentative and difficult, has echoed down through the centuries, and nowhere is this contempt One of the major assumptions of popular culture is that “a more visible than in Western popular culture, where women’s youth, woman is her body”, or more specifically, her body’s reproductive beauty and fecundity is worshipped (acknowledging the Maiden and capacity. “‘In the hyper visible landscape of popular culture, the the Mother), but older women (the Crone) are viewed with horror, body is recognized as the object of women’s labor: it is her asset, her fear and revulsion. product, her brand and her gateway to freedom and empowerment in This negative perception of the Crone, reflected historically in a neoliberal market economy.’ (Winch 2015, 233).” The centrality the Patriarchal suppression of Goddess Culture and in contemporary of woman’s role as a mothering agent has resulted in less attention society by negative attitudes towards aging women, has always been given to the development of women in midlife when reproductive challenged by feminists (Rountree 1997; Ronn and Daugaard 2005). capacities diminish and women start to become invisible in cultural This is aided by the flourishing of contemporary counter-cultural representations (Segal 2014). Moreover, when attention focuses on movements such as Neopaganism and Wicca (Griffin, 1995). The women’s middle life, the concern is often exclusively biological crone is now reinstated, re-imagined and reinvigorated in our ‘col- (Ballard, Elston, Gabe 2009, Gerguen 1990, p476), perpetuating the lective consciousness’ (Walker, 1985), as women grow increasingly assumption that women are their body. tired of the negative associations that go with being beyond the The hyper visibility of the reproductive body emphasizes the menopause (Bay 2004; Pearsall 2018) and seek to reframe how we fact that the menopause is seen as a decline. Corus and Saatcioglu think about the Crone and the aging woman. (2012) refer to the literature on the menopause as a loss, guided by The crone, often a vilified and marginalised figure, is thus a medical discourse: “Although menopause is a natural hormonal positive archetype and force in contemporary women’s culture, her transformation, it is sometimes treated as a ‘process of breakdown, fire fanned by the internet. She is celebrated as a source of wisdom, failure and decline’” (Martin 2001, p43, cited in Corus and Saatciog- knowledge, freedom and feminine power, as evidenced by the pro- lu 2012). In this mainstream medical discourse, the menopause is de- liferation of books, websites, forums, workshops, retreats, artworks, scribed as a problem of hormone deficiency, a pathological state that magazines and e-zines, performances, music, workbooks, and films, needs to be treated. “Terms such as ‘deterioration’, ‘wearing out’, all of which reclaim the Crone for our times. Drawing on these popu- ‘failing’, ‘no longer up to it’ and ‘getting slower’ are traditionally lar culture sources, we reveal key themes associated with the crone used to describe the age-related physiological changes within the that assist women in reclaiming her as an empowering figure, name- body” (Ballard et al. 2009, 279). Alternative stances, like the natural ly: freedom to choose; a disruptive cackle; and the censoring look. perspective, view the menopause as a natural physiological transfor- Then, using Butler’s theory of gender performativity (Butler 1990; mation that brings with it new roles and freedoms in women’s lives 1993), we discuss how these themes denote a range of resignify- (Corus and Saatcioglu 2012). ing practices that challenge existing ideals of femininity and disrupt We tend to focus on women’s biological aging and pay less at- norms that value a youthful - as opposed to aged - female body. tention to the cultural dimension of aging. This is what Susan Sontag (1972) called ‘The double Standard of aging’ (14): women are mar- ‘The Foreigner within us’: Catharsis amid Horror and ginalized and diminished when youthful looks depart (Segal 2014), Confusion in Menopause and are subject to growing gender inequality in aging (Steinfield, Coleman, Tuncay Zayer, Ourahmoune, and Hein 2019). However, EXTENDED ABSTRACT feminist writers have highlighted some powerful aspects of female aging (Segal 2014, Pearsall 2018, Beauvoir 1970), referring to the “The worst symptom is never mentioned: there is a post- real self versus the aging self. More recently, consumer researchers menopausal elephant in the bedroom and its name is Dry Vagina have called for further investigation into new understandings of the (…) A person can apply only so much lubrication. Even then, context of aging (Schau, Gilly, and Wolfinbarger 2009, Barnhart and penetration is excruciating. (…) Finally, I developed a vaginal Peñaloza 2013, Corus and Saatcioglu 2012). Our paper contributes infection (…) So what are all those other post-menopausal, to this call for alternative perceptions of aging by examining the con- sexually active women doing (referring to celebrities)? Have they text of women’s experiences of the menopause. miraculously escaped?” (reader’s testimonial, The Guardian, In this study, we seek alternative discourses about the meno- 2019_01_04). pause experience. How do women live and make sense of their transformation in relation to the menopause? How can these women, at the intersectionality between age and gender, re-empower “When biology no longer defines a body by its potential to themselves? We collected 1000 articles referring to the menopause reproduce, a woman will find further equality in her relationships.” and published in the Guardian newspaper between 2005 and 2020. (journalist’s testimonial, The Guardian, 2019_05_19) Among them, 150 articles included personal testimonials about the menopause experience written by regular readers and opinion lead- “Insomnia gone, full of energy, no more night sweats, only very ers (female journalists, artists, managers). occasional flushes and a new confidence and stronger sense of We conducted our analysis based on these 150 testimonials. self.” (reader’s testimonial, The Guardian 2019_08_25) They reveal three main themes: 1. Confusion and taboo, 2. Meno- pause horrors as experienced by the writers, 3. Catharsis. The study expands on the philosophical ideas of Julia Kristeva (1991) about “I catch glimpses of the enthusiastic child I once was, and being strangers to ourselves to further explore the menopause as a intimations of an older woman’s wisdom, hard-won through a lived experience. Kristeva suggests searching deep inside ourselves 1108 / Shining A Light on Female Power: Women’s Experiences of Transcending Aging for our more profound fears about those we call ‘the foreigners’. In tion ensemble (ECE hereafter, Barnhart et al. 2014; Barnhart and a similar way, old age is often approached with the same kind of Peñaloza 2013)—how might it enhance these social and market ties? fear that is felt with regard to foreigners. By welcoming the other This research builds on previous work on the ECE by develop- within us, in other words, aging and the menopause, by facing the ing theoretical and practical insights pertaining to the cocreation of confusion and sometimes the horror it arouses in women and society value (Karababa and Kjeldgaard 2014; Kelleher et al. 2019; Stor- at large, we may be able to get a better understanding of the strange- backa et al. 2016) by examining the experiences and concerns of ness of aging and its cathartic benefits. Our study goes beyond the elderly women. Elderly women’s value cocreation in consumption dichotomy between medicalization and a more ‘natural’ approach to in the ECE is of particular importance for their quality of life and for dealing with the menopause by representing women’s experiences the societies they inhabit. Notably, of the 49 million persons in the with the menopause as an embodied encounter with the strangeness U.S. over the age of 65, 55% are women (Roberts et al. 2018), with of an ‘unproductive’ body. an average life expectancy of 81 years compared to 76 years for the One interesting finding of the study has to do with the distri- 45% who are men (Population Reference Bureau 2019). In addition bution of voices among the different opinion pieces. Testimonials to living longer on average, and having played particular roles in from regular readers frequently refer to confusion and horror with families and households for those who have/did not have children, regard to the menopause, while its cathartic and empowering nature aging women have distinct life situations and experiences that shape is defended to a greater degree by the testimonials of female opinion and impact their consumption (Drolet et al. 2010). According to a leaders rather than by regular readers. recent census report on the elderly in the U.S. (Roberts et al. 2019), This paper contributes to the discussion of female aging in mid- elderly women have lower levels of education (approximately 1 in dle life. It extends prior research that opposed the mainstream medi- 5 women over the age of 65 have at least a bachelor’s degree as cal discourse on menopause with the feminist discourse (Corus and compared to 1 in 3 men of the same age), income (their respective Saatcioglu 2012) by showing the capacity of women to transcend median incomes are $41,200 versus $56,850 for men), and poverty their unproductive and declining body and reveal themselves beyond (11% of women live below the poverty level as compared to 7% of their body. In this process, becoming “strangers to themselves” helps men in this same age cohort), and exhibit higher rates of employment women to distance themselves from the patriarchal assumptions that (14% of women continue to work as compared to 22% of men). associate them uniquely with their bodies. In this context, the paper The context of this study is the U.S., in drawing from empiri- raises the question of how the empowerment testimonials of women cally grounded analysis of interviews exploring the daily life of ag- opinion leaders can support or clash with regular women in their ing American women in relation to family members/friends and paid journeys to discover the other within themselves. care providers in 10 ECEs, participant experience in researchers’ families, and representations of elderly women in popular culture. What’s Love Got to Do with It that Technology can’t Our findings detail the circumstances and patterns of relationships Handle? Opportunities and Challenges for Aging in the ECE, as aging women cocreate value with and against family Women in CoCreating Value in the Elderly Consumption members/friends and service providers, with the aid of material and Ensemble technology. To explain the figure: in the circle at the top left, the figure illustrates an existing division of consumption ‘labor’ activities EXTENDED ABSTRACT among the elderly person, E, family member/friend(s), F, and paid provider(s), P. While all persons within the ECE utilize technology, They need me. I mean, if they are capable to be alone, they will not we focus on its use by the elderly person in this study, as designated need me. –Maria, paid caregiver, age 55 by MT at the top, center of the diagram. Circumstances that alter this division of consumption labor activities include changes in the health, physical abilities, and living arrangements of the elderly per- There really is…a need for a book, the Dr. Spock parenting book for son, their proclivities, and relations with others, as well as changes in doing this, cause there’s just so much…You’ve been living your own the participation of other members of the ECE. In turn, these changes life and now you’re 55 years old, suddenly somebody has to inherit ripple through the ECE in altering patterns of collaboration (E and F Mom… –Kevin, age 60 begin to jointly pay bills, for example), alliances (E and P exclude F from an activity, for example), conflict (between E and F, or F and P, If my iPad were a woman, I’d marry her! Doris, age 76 for example), deferral (E defers to F to enhance their relation, F defers to the expertise of P, for example), and dis/engagement (E fires Persons 65 years of age and over represent approximately 18% P or F exits the ECE, as examples). of the population in Europe, as compared to 9% in the U.S. (Popula- Discussion elaborates how the women empower themselves by tion Reference Bureau 2019). In the U.S. persons over the age of leveraging technology in cocreating value with other agents in the 65 exhibit somewhat similar rates of home ownership (78%) and ECE, by enhancing their capabilities to monitor health and living internet access (76%), although patterns of living alone, with family arrangements, handle finances, procure provisions, maintain social members, and with nonfamily members tend to vary with the loss relationships with ECE and others, and remain active in their com- of sight, hearing, mobility, cognitive ability, and self-care that many munities by keeping abreast of/participating in local events/interest persons experience as they advance in age (Roberts, Ogunwole, groups. Implications concern women’s autonomy and vulnerabilities Blakeslee and Rabe 2018). The opening quotes illustrate the im- in enhancing such value cocreation through the use of technology portance of family, service providers, and technology, and previous in dynamic, ongoing, and emergent experience and interaction with research strongly and repeatedly shows that greater social contacts family members, friends, and service providers in the ECE and be- bring about higher and longer quality of life (Pettigrew and Moschis yond, and in favorably comparing their life situations and experienc- 2012; House, Landis, and Umbertson 1988). And yet, what is the es with those of others, both real life and mediated. The presentation impact of technology on value cocreation in the elderly consump- closes with promising avenues for further research. Advances in Consumer Research (Volume 48) / 1109 Wisdom Examined via a Qualitative Investigation of the Data was analyzed following recommendations for qualitative Bathing Rituals of Aging Moroccan Women and their data analysis (Moisander and Valtonen 2006). The authors analyzed Granddaughters the data first separately and then together (Sherry 2007). Iteration was used to form, revise, and develop understandings of the entire EXTENDED ABSTRACT data set. The data analysis focused on informants’ reflexive thoughts This study investigates how aging women (Barnhart and Pe- about bathing experiences and their relationships with their relatives. ñaloza 2013) exhibit consumer wisdom (Luchs and Mick 2018; After considering several theoretical frameworks, we were able to Luchs, Mick and Haws under review) in their families. It specifically identify how wisdom emerged in narratives. examines how wisdom is revealed during the public bathing practic- Our findings first uncover how grandmothers use their wisdom es and rituals of older Moroccan women with their granddaughters. in the building of their relational identities with their granddaugh- Wisdom in consumer behavior research refers to “the pursuit of ters. Specifically, it shows how grandmothers’ wisdom is manifested wellbeing for oneself and for others through mindful management - through sharing their views on lifestyle responsibility, purpose, of consumption-related choices and behaviors as realized through flexibility, perspective, prudent reasoning and transcendence (Luchs, the integrated application of intentionality, contemplation, emotional Mick and Haws under review) - as they build their relational identi- mastery, openness, and transcendence” (Luchs and Mick 2018, 371). ties with their granddaughters in the context of bathing practices. The five facets of consumer wisdom refer to specific dimensions and Second, this study unveils how grandmothers’ wisdom is in- traits – cognitive, affective, and behavioral characteristics and ten- formed by their granddaughters’ consumption practices. In particu- dencies. Recently, Luchs, Mick and Haws (under review) extended lar, grandmothers integrate contextual cues related to their bathing their framework and depict six dimensions to consumer wisdom in practices with their granddaughters to redefine their own sense of the specific context of purchase decisions: lifestyle responsibility, wisdom. This result complements existing work, which has essen- purpose, flexibility, perspective, prudent reasoning, and transcen- tially focused on how one’s personal consumption is used to develop dence. Lifestyle responsibility corresponds to how people resist wisdom (Luchs, Mick and Haws under review). temptation to achieve their lifestyle goals. Purpose relates to priori- Finally, this study highlights how granddaughters - younger tizing experiences that help build relations with others. Flexibility women - build their own sense of wisdom when they interact with is about sharing things with others. Perspective refers to the efforts their grandmothers, and how they learn from their grandmoth- made to grasp multiple perspectives. Prudent reasoning is about be- ers’ consumption experiences. This finding contributes to existing ing cautious before making a decision. Transcendence refers to ethi- knowledge on wisdom, which has mainly conceived wisdom build- cal, communal, and environmental values. ing as an individual process, which occurs in late adulthood (Hall How consumer wisdom manifests and is understood heavily 2010). depends on contextual settings (Hall 2010). In the marketing and Our study contributes to existing literature about consumer consumer behavior fields, research on wisdom has essentially fo- wisdom and relational identity in old age as related to bathing and cused on business managers (Mick, Bateman and Lutz 2009), shop- beauty rituals in consumption. It extends prior research on wisdom, pers (Mick, Spiller and Baglioni 2012), gatekeepers in local organi- and takes account of interpersonal relations (Luchs and Mick 2018) zations (Luchs and Mick 2018), or innovative members of e-tribes and contextual cues (Hall 2010). Grandmothers’ wisdom refers to the (Kozinets, Hemetsberger and Schau 2008). However, little is known collective pursuit of balanced relational identities and family mem- about how wisdom manifests in family settings and intergenerational bers’ wellbeing through shared consumption experiences. Grand- exchanges, and particularly the intersection between consumer wis- mothers’ wisdom is co-created in intergenerational exchanges. dom and aging women in intergenerational exchanges. Grandmothers embrace various individual, relational and col- REFERENCES lective identities (Epp and Price 2008), which are embedded in inter- Ballard, Elston, Gabe (2009), “Private and Public Ageing in the UK generational exchanges. Intergenerational exchanges were examined The Transition through the Menopause,” Current Sociology between adult children and their parents (Karanika and Hogg 2016) Vol. 57 No. 2 Monograph 1 and between grandmothers and their young grandchildren (Gode- Baring, Anne and Cashford, Jules (1993), The Myth of the Goddess: froit-Winkel, Schill and Hogg 2019). However, little is known about Evolution of an Image, London: Penguin. how grandmothers’ wisdom plays out in intergenerational exchanges Barnhart, M., Huff, A.D. & Cotte, J. (2014). Like a member particularly between grandmothers and their granddaughters. of the family: including and excluding paid caregivers in This study aims to fill this gap by capturing Moroccan grand- performances of family. Journal of Marketing Management, mothers’ and granddaughters’ narratives about bathing rituals and 30:15-16, 1680-1702. practices. In the Arab world, women traditionally go to the public Barnhart, Michelle and Lisa Peñaloza (2013), “Who are you calling bath, hammam, each week. While today many families have their old? Negotiating Old Age Identity in the Elderly Consumption own private bathroom in their homes, going to the public bath re- Ensemble,” Journal of Consumer Research, 39 (April), 1133- mains an important practice among both younger and older Moroc- 1153. can women. This study captures how both Moroccan grandmothers Bay, Ruth E. (2004) “Towards the Croning of Feminist and their teenage/adult granddaughters shape and ascribe meanings Gerontology”, Journal of Aging Studies, 18 (1), pp. 109-121. to a ritual that is steeped in history. Beauvoir de Simone (1970), La Vieillesse, Collection Blanche, To grasp a deeper understanding of how wisdom works between Gallimard, Paris. grandmothers and granddaughters in the context of bathing rituals Butler, Judith (1990). Gender Trouble: Feminism and the nd and practices, we conducted long interviews (McCracken 1988) with Subversion of Identity, (2 ed.). New York: Routledge. 12 Moroccan grandmothers in their homes, and with their grand- ______(1993). Bodies That Matter: On the Discursive Limits of daughters (age 16-24; varied socio-cultural backgrounds). We also ‘Sex’. New York: Routledge. used participant observation in public baths in Casablanca, Morocco. Interviews ranged from 45 minutes to 2 hours. 1110 / Shining A Light on Female Power: Women’s Experiences of Transcending Aging

Corus, Canan and Bige Saatcioglu (2012), The Social Construction Mick, David G., Thomas S. Bateman, and Richard L. Lutz (2009), of Womanhood in MiddleAge: Menopause As a Rite of “Wisdom, Exploring the Pinnacle of Human Virtues as a Passage. Advances in Consumer Research, eds. Zeynep Central Link from Micromarketing to Macromarketing,” Gürhan-Canli, Cele Otnes, and Rui (Juliet) Zhu, Duluth, MN: Journal of Macromarketing, 29, 2, 98-118. Association for Consumer Research, 40,18-21. Mick, David G., Stephen A. Spiller and Anthony J. Baglioni Drolet, Aimee, Norbert Schwartz, and Carolyn Yoon (2010), (2012), “A systematic Self-Observation Study of Consumers’ “The Aging Consumer: Perspectives from Psychology and Conceptions of Practical wisdom in Everyday Purchase Economics,” NY: Routledge. Events,” Journal of Business Research, 65, 7, 1051-1059. Epp, Amber M. and Linda L. Price (2008), “Family Identity: A Moisander, Johanna and Anu Valtonen (2006), Qualitative Market Framework of Identity Interplay in Consumption Practices,” Research: A Cultural Approach, London: Sage Publication. Journal of Consumer Research, 35 (1), 50–70. Moschis, George P. (2012), “Consumer Behavior in Later Gergen, Mary M. (1990), Women’s Development within the Life: Current Knowledge, Issues, and New Directions for Patriarchy. Psychology of Women Quarterly, 14, 471-493. Research,” Psychology & Marketing. Vol. 29, 2, 57-75. Godefroit-Winkel, Delphine, Marie Schill and Margaret Hogg Pearsall, Marilyn (ed.) (2018) The Other Within Us: Feminist (2019), “The Interplay of Emotions and Consumption in the Explorations of Women and Aging, Routledge, London. Relational Identity Trajectories of Grandmothers with their Pettigrew, S. and G. Moschis (2012), Consumer Wellbeing in Grandchildren,” European Journal of Marketing, 53 (2), 164- Later Life. In D.G. Mick, S. Pettigrew, C. Pechmann and J.L. 194. Ozanne, eds., Transformative consumer research: for personal Gopaldas, Ahir and Anton Siebert. (2018), Women over 40, and collective well-being, London: Routledge, p. 565-582. foreigners of color, and other missing persons in globalizing Population Reference Bureau (2019), Population Bulletin, “Aging mediascapes: understanding marketing images as mirrors of in the United States,” Washington, DC. https://www.prb.org/ intersectionality, Consumption Markets & Culture, 21:4, 323- aging-unitedstates-fact-sheet/ 346, DOI:10.1080/10253866.2018.1462170 Roberts, Andrew W., Stella U. Ogunwole, Laura Blakeslee, and Griffin, Wendy (1995), “The Embodied Goddess: Feminist Megan A. Rabe, (2018), “The Population 65 Years and Older Witchcraft and Female Divinity’, Sociology of Religion, 56 in the United States: 2016 American Community Survey (1), pp. 35-48. Reports, Washington, D.C. U.S. Printing Office: U.S. Bureau Hall, Stephen S. (2010), Wisdom, From Philosophy to of the Census, October 2018, ACS-38. https://www.census. Neuroscience, New York, NY: Vintage Books. gov/content/dam/Census/library/publications/2018/acs/ACS- House, J.S., K.R. Landis, and D. Umbertson (1988), “Social 38.pdf. Relationships and Health,” Science, 241: 540-545. Ronn, Vicki and Daugaard, Kaylene (2005), “The Crone Karababa, Eminegul and Dannie Kjeldgaard (2014), “Value in Conspiracy”, International Journal of the Humanities, (Sept), Marketing: Toward sociocultural perspectives,” Marketing 3 (1), pp. 43-50. Theory, 14(1) 119–127. Rountree, Kathryn (1997), “The New Witch of the West: Feminists Karanika, Katerina and Margaret K. Hogg (2016), “Consumption Reclaim the Crone”, Journal of Popular Culture, 30 (4), Through the Ambivalent Prism of Intergenerational Support”, (Spring), p. 211. European Journal of Marketing, 50, 3/4, 575-601. Schau, Hope Jensen, Gilly, Mary C., Wolfinbarger, Mary (2009), Kelleher, Carol, Dierdre O’Loughlin, Johanna Gummerus, and Lisa “Consumer Identity Renaissance: The Resurgence of Identity- Peñaloza (2019), “Viewing the Kaleidoscope: Orchestration of Inspired Consumption in Retirement”. Journal of Consumer Value Co-Creation,” Journal of Service Research, published Research. Vol. 36 Issue 2, p255-276. DOI: 10.1086/597162. online 7 November. Segal Line (2014), Out of time, the pleasures and perils of ageing, Kelleher, Carol and Lisa Peñaloza (2017), “Aging and Verso, London. Consumption” in Part 5, Consumption and Social Divisions, Sherry, John F. (2007), “Fielding Ethnographic Teams: Strategy, The Routledge Handbook on Consumption, Margit Keller, Implication and Evaluation,” in Handbook of Qualitative Bente Halkier, Terhi-Anna Wilska, Monica Truninger, editors, Methods in Marketing and Consumer Research, ed. Russell London: Routledge, pp. 326-338. Belk, Northampton, MA: Elgar, 261-276. Kozinets, Robert V., Andrea Hemetsberger and Hope J. Schau Steinfield, Laurel A., Catherine A. Coleman, Linda Tuncay Zayer, (2008), “The Wisdom of Consumer Crowds: Collective Nacima Ourahmoune, Wendy Hein (2019), “Power logics Innovation in the Age of Networked Marketing,” Journal of of consumers’ gendered (in)justices: reading reproductive Macromarketing, 28, 4, 339-354. health interventions through the transformative gender justice Kristeva Julia (1991), Strangers to Ourselves, Translated by Leon framework”,Consumption Markets & Culture, 22:4, 406-429, S. Roudiez, Columbia University Press. DOI: 10.1080/10253866.2018.1512250 Luchs, Michael G. and David G. Mick (2018), “Consumer Wisdom: Sontag Susan (1972), “The double standard of aging “, The a Theoretical Framework of Five Integrated Facets,” Journal Saturday Review, New York. of Consumer Psychology, 28, 3, 365-392. Stevens, L. and Maclaran, P. (2007), “Exploring the Celtic Luchs, Michael G., David G. Mick and Kelly L. Haws (under Narrative in Advertising: Goddess Culture and the Lexicon review), “Measuring and Understanding Consumer Wisdom of Perfumery”, Journal of Strategic Marketing, February, 15, for Personal Well-Being”. 29-39. McCracken, Grant (1988), The Long Interview, Newbury Park: Storbacka, Kaj, Roderick J. Brodie, Tilo Böhmann, Paul P. Sage. Maglio, and Suvi Nenonen (2016), “Actor Engagement as a Microfoundation for Value Co-Creation”, Journal of Business Research, 69 (8), 3008-3017. Advances in Consumer Research (Volume 48) / 1111

Szmigin, Isabelle and Marylin Carrigan, M. (2001), “Learning to Winch, Alison (2015), “Brand Intimacy, Female Friendship and love the older consumer”, Journal of Consumer Behaviour, 1: Digital Surveillance Networks,” New Formations, 84-85: 22-34. doi:10.1002/cb.51 228-45. Walker, Barbara G. (1985), The Crone: Woman of Age, Wisdom and Power, HarperCollins Publishers, New York. The Body and Equipment: Understanding Consumption Through Relationships Between Body and Equipment Chairs: Annetta Grant, Bucknell University, USA Robin Canniford, University of Melbourne, Australia

Paper #1: Why Do Consumers Put Their Bodies at Risk? A ity. This study examines the embodiment of material culture to show Reflexive Modernist Analysis of CrossFit’s Marketplace Culture how consumers assemble masculinity through an array of socio- Craig Thompson, University of Madison-Wisconsin, USA material practices. Findings show how the dualistic categories of Anil Isisag, Emlyon Business School, France mind/body and subject/object dissolve in the intricacies of embodied socio-material relations. Paper #2: The Socio- and Material-Temporal Routines of The fourth paper considers skill and knowledge as equipment Mothering: Why and How Physical Activity Often Fails to Fit through which consumers effect body practices that have restorative Fiona Spotswood, University of Bristol, UK benefits. In an ethnographic study of hobby fly fishing, we contrib- Paper #3: Assembling Embodiment: Body, Techniques and ute an extended theory of flow to show how consumers bodies, the Things environment, and fishing equipment create restorative experiences. Stephen Murphy, University of Essex, UK We use various contexts, theoretical, and methodological ap- Paper #4: Extending Flow: How Place, Materials, and Body proaches to body-equipment relationships to reveal new perspectives Create Restorative Consumption in Nature on the role of the body in consumer experience, consumer respon- Annetta Grant, Bucknell University, USA sibilization, and gendered understandings of practices. This session Avi Shankar, University of Bath, UK embraces the suggested conference themes of technology (equip- Robin Canniford, University of Melbourne, Australia ment) and obesity. The session is especially relevant for researchers interested in embodiment, temporality, and practice and assemblage SESSION OVERVIEW theories. The body is the primary conduit to consumers’ sense and understanding of the world (Canniford et al 2019; Joy and Sherry 2003). Why Do Consumers Put Their Bodies at Risk? A The body is a site at which consumers enact self-identity (Roux and Reflexive Modernist Analysis of CrossFit’s Marketplace Belk 2019; Patterson and Schroeder 2010), reflexivity (Husemann Culture and Eckhardt 2019; Thompson and Üstüner 2015), as well as pleasure and pain (Goulding et al. 2009; Scott, Cayla, Cova 2017). In all EXTENDED ABSTRACT cases, however, the consumer’s body is surrounded and augmented Consumer researchers and sports sociologists have taken an in- with forms of physical and discursive equipment. By equipment, terest in understanding why consumers choose to literally put their we mean sets of devices used to achieve practical ends (Rabinow bodies on the line, subjecting themselves to extreme physical stress 2009). Equipment can span multiple scales, from the apparatus of the and pain, and to risk injury (and even death) in the service of auto- individual body (Zanette and Scaraboto 2019), through meso-level telic leisure activities (Atkinson 2011; Celsi, Rose, and Leigh 1993; assemblages of body and consumption goods (Murphy et al 2018), Dawson 2017; Scott, Cayla, and Cova 2017; Saville 2008; Schneider, to technologies that link consumers into large-scale dispositifs (Lup- Butryn, Furst, and Masucci 2007). These studies have been framed ton 2015). In this special session, we extend current understandings by two distinct interpretive frameworks, which we refer to as the of embodiment with new perspectives on the relationships between existential and neoliberal models, each highlighting a different set of consumers’ bodies and equipment. We explore material and non- explanatory socio-cultural conditions. material concepts of equipment to ask: What does the relationship The existential model builds on the work of David Le Breton between body and equipment reveal about consumption? And more (2000) and argues that such risk-taking activities provide individu- particularly how does considering different facets of equipment en- als opportunities for spiritual quests and authenticating experiences lighten an understanding of embodiment? that they find lacking in the rationalized flow of modern life and its The first paper considers how consumers mobilize and shape bureaucratic work orders (Celsi et al. 1993; Scott et al. 2017; also see their bodies in preparation for contingencies and imminent threats Arnould and Price 2000; Caillois 2001). In comparison, the neolib- through the equipment common to the global phenomenon of Cross- eral model maintains that physically challenging activities are part Fit. Based on a two-year ethnography, this study analyses CrossFit’s and parcel of a new disciplinary regime of neoliberal governmental- marketplace culture as a meso-level inflection of reflexive modernity (Foucault 2008) whereby individuals regard social life as a sphere ization. The study’s findings offer an alternative to extant existential of cutthroat competition in which only the fittest (and most self-dis- and neoliberal explanations of consumer’s embodied risk-taking. ciplined) can attain success (Cederström and Spicer 2015; Datta and The second paper considers the equipment of time, and time as Chakraborty 2017). equipment. Through an analysis of diary data and depth interviews Previous examinations of CrossFit’s marketplace culture— with mothers from deprived urban areas, this study illuminates the which affords a communally shared experience of physical- chal entangled practices and temporalities of mothering. The resulting lenge, risk taking, and calculated endurance of pain—have rather configurations shape the (im)possibilities of physical activity and expectedly followed an existential (Dawson 2017; Herz 2014; Pe- mothering. A body equipped with time is better off, but the equip- kkanen, Närvänen and Tuominen 2017) or neoliberal lens (Heywood ment of time is both gendered and classed, raising considerations for 2016; James and Gill 2017; Nash 2017). While the narratives of the rising levels of obesity. CrossFit enthusiasts we interviewed exhibit motifs that are compat- In the third paper, the materials necessary for hobby motorcycle ible with both existential and neoliberal interpretive frames, we pro- repair constitute the equipment for expressing embodied masculin- pose that there is more to the story than an enactment of neoliberal

Advances in Consumer Research 1112 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1113 responsibilization (Giesler and Veresiu 2014) or a quest to authen- accounts, it manifests an always contingent endeavor for being pre- ticate the self through transcendent experiences (Arnould and Price pared to effectively respond to challenges and contingencies amidst 1993, 2000). the radical uncertainty induced by systemic risks. Another important macro influence on these risk-taking consumption practices is the reflexive modernist orientation that has The Socio- and Material-temporal Routines of arisen from the structural imposition of global systemic risks (Beck Mothering: 1999, 2006; Beck, Giddens, and Lash 1994). According to this Why and How Physical Activity Often Fails to Fit framework, systemic, human-generated risks are the unintended and unanticipated consequences of the modernist project and its abiding EXTENDED ABSTRACT goals of creating wealth and prosperity. Modern economies produce The transition to parenthood is a critical transformative experi- risks just as they produce economic gains, and these risks are dif- ence and an inflection point for obesity (Saxbe et al., 2019). Be- ficult to manage and distribute through probability assessment—as coming a parent often leads to reduced physical activity (PA) levels, exemplified by climate change, large-scale industrial accidents, ter- but this effects mothers more than fathers, both in volume and type rorist attacks, and outbreaks of infectious diseases. As a cultural re- (Rhodes et al., 2014). Studies report that whereas some mothers ad- sponse to these conditions, the reflexive modernist ethos comprises just physical activity participation strategies to fit around children of a reflexive awareness of systemic risks, a socially pervasive state (McGannon et al., 2018), many attempt unsatisfactorily to integrate of doubt about the ability of experts to accurately predict and con- PA into family time (Hamilton and White, 2010). Lower socioeco- trol such risks, a profound uncertainty about which authorities to nomic-economic status mothers are the least likely to be active trust, and a paradoxical quest for being prepared for a wide range (Bellow-Riecken et al., 2008). Supporting mothers from deprived of imaginable contingencies in hopes that such preparations might areas to become and maintain healthy levels of physical activity is prove useful to a certain degree in a time of crisis. When CrossFit therefore an important challenge for social marketers. practices are analyzed in relation to this reflexive modernist frame, Qualitative health research tends to explain mothers’ low PA we identify a set of socio-cultural motivations that are ideologically levels as the product of social and cultural discourses which mothers and experientially distinct from those highlighted by the existential draw on. ‘Mother’ is understood as women’s ‘true calling’, with an and neoliberal models. emphasis on intensive presence and prioritising caring above all else Founded in 2000 by former gymnast Greg Glassman, Cross- (Molander and Hartmann, 2018). PA is therefore constantly subor- Fit hinges on the idea that constantly varied, high intensity, whole dinated through daily computations made in the shadow of socially body workouts could induce a state of all-round functional fitness prescribed expectations. Relatedly, mothers’ low PA levels is ex- unattainable through conventional gyms. With this principle in mind, plained by illuminating the limited leisure time afforded to mothers Glassman developed high-intensity workout routines that integrated as they assume fuller caring roles than male partners (Hamilton and movements from gymnastics, calisthenics, and Olympic weightlift- White, 2010), which compresses and fragments time for themselves. ing, all involving multiple repetitions in rapid-fire succession. Glass- Mothers’ ‘free time’ is noted to be highly dependent on the support man’s unorthodox approach to building functional fitness through of others (Wearing, 1990). Time is conceptualised as a resource that maniacal physical exertion attracted an expanding clientele of for- mothers have little of. mer athletes seeking to recreate bygone experiences of competitive Existing studies tend to overemphasise the capacity and respon- intensity and, importantly, first responders and military personnel sibility of individuals to realise change through resistance to soci- who face a broad spectrum of physical challenges in the conduct of etal expectations (McGannon and Schinke, 2012). Through a novel their occupations (Crockett and Butryn 2018). In press interviews, practice theory lens, this study shifts focus to the mundane experi- Glassman matter-of-factly discussed latent risks that CrossFitters ences of mothering, examining everyday temporalities and mundane knowingly take in order to garner the benefits of its hyperintense yet expertly operated materialities. The connecting theme of ‘time’ training regimen: ““It can kill you,” he said. “I’ve always been com- emerges as vital for understanding the ‘demands’ of the system of pletely honest about that”” (Quoted in Cooperman 2005). practices that dominates the organisation of mothers’ everyday lives Based on a two-year long ethnography of CrossFit’s market- and often competes out physical activity. place culture—consisting of participant observation, semi-structured We undertook and analysed 15 depth interviews with mothers interviews, and examinations of media discourses and branded con- in Bristol. Mothers in lower socio-economic positions were selected tent—, we analyze CrossFit as a meso-level articulation of reflexive because they have less control over how their paid and non-paid time modernization. We demonstrate that CrossFitters’ understanding of (Southerton, 2006). Half the mothers kept a diary for a week prior embodied risk taking is shaped by the normative goal of preparing to the interviews. Interviews explored everyday household routines, oneself for unexpected contingencies and imminent threats. In doing use of material things and explored experiences and reflections of so, we bridge the analytical gap between studies addressing consum- PA. Mothers ranged between 20 and 50 years and the sample in- ers’ proactive risk taking and those focusing on the feelings of anxi- cluded single, adoptive, married and cohabiting mothers and those ety and uncertainty induced by the structural imposition of systemic whose children had special needs. Five participated in regular PA. risks. We highlight that such risk-taking challenges can themselves Transcripts were analysed using thematic analysis and three key be market-mediated responses to the conditions of reflexive modern- themes were identified. ization, whereby consumers proactively take calculated bodily risks Mothering can be understood as a multiple integrative practice out of a desire to expand their capacities to effectively respond to inseparable from a larger household system of practices involving the unpredictable manifestations of systemic risks. Finally, we show multiple practitioners: children and fathers but also schools, work- that the notion of personal responsibility—which has predominantly places and childcare providers. Mothering is highly competitive, been studied as a defining feature of neoliberal governmentality by even predatory. As one respondent explained, “We don’t really do consumer researchers (Giesler and Veresiu 2014)— serves different anything [now we have kids]…” PA for all our respondents took a ideological ends in a reflexive modernist frame. Rather than taking low priority, and the synchronisation of physical activity alongside a neoliberal, entrepreneurial orientation as represented in previous 1114 / The Body and Equipment: Understanding Consumption Through Relationships Between Body and Equipment mothering practices was found to be highly unsatisfactory and “not sion (Goulding et al., 2009; Scott et al., 2017) and emotive bodily really proper exercise”, or impossible. expressions (Hewer and Hamilton, 2010). Recent studies broaden Three intersecting themes add theoretical power to our under- the scope of multi-sensory enquiry, by establishing a basis for con- standing of the predatory nature of mothering and how this impli- sumer research that attends to the olfactory (Canniford et al., 2018) cates PA participation. Firstly, mothering practices have a dominant and sonic (Patterson and Larsen, 2018). Notwithstanding these sig- temporal rhythm through their entanglement with external interlock- nificant contributions, a deep understanding of embodiment cannot ing space-time configurations. For example travel timetables, dis- be reached through recourse to the senses alone. Embodiment also tances from work and childcare/school opening hours all shaped the foregrounds body techniques, habituation, and the social, physical rhythm of mothering. These space-time configurations demand skil- and mindful cultivation of skilful dispositions (Mauss, 1973; Cross- ful use of materials such as managing the weekly wash, which often ley 2007; Murphy et al., 2019). Recent CCT studies illuminate the happened in a prescribed 90 minute slot; or calculating how long a skilful nature of a host of consumption activities that range from bus journey would take and if there was time for an errand or visit freeskiing and paintballing (Woerman and Rokka, 2015), Cosplay between school hours. (Seregina and Weijo, 2017), craft beer connoisseurship (Maciel and Secondly, cutting across the external space-timescapes are the Wallendorf, 2017) and high-speed motorcycling (Murphy and Pat- temporal rhythms of everyday caring. These were dominated for our terson, 2019). Although consumption objects are clearly central to participants by the need for careful synchronisation of multiple ac- the performance of these skilful activities, these studies do not fore- tivities and by temporal ‘hot spots’. Mothers were required to orches- grounded materiality conceptually, and as a consequence we know trate practices so they occurred in the right order, with the skilled use little about how things feature in the development of skills and the of particular equipment (like toothbrushes or cooking equipment), at embodied dispositions that pertain to them. To this end, the present the right time of day and with understood degrees of periodicity. Hot study extends Dant’s (2004) assemblage theory to examine the intri- spots occur because many of these practices colonise (Shove, 2012) cacies of socio-material relations between motorcycle repairers and particular points of the day such as children’s bedtime, and are also their motorcycles. intersected by external, institutional spatial-temporal rhythms. PA Re-theorizing Assemblage: An Embodied Perspective needs to compete to recruit mothers. As Jane explains, her walk was Tim Dant’s extensive work on materiality (1999; 2004a; 2004b) delayed because “he poo-ed all over the carpet and that takes half an brings the issue of embodiment and its role in the formation of human hour out of your day before you can go anywhere”. Mothering prac- non-human assemblages into critical light. Dant (2004) argues that tices must often be accomplished in a particular order. the embodied experience of being-in-the-world only makes sense Thirdly, infusing the household system of practice were collec- because we have a history of sensory experience. We know how it tive expectations towards ‘time spent’ that shape PA participation. In feels to be in the world, and we intuitively know how to interact in line with others (Silva 2002), we found an accepted lack of ‘time to the world because of this experience. This embodied way of being- yourself’ for mothers and a low prioritisation of PA for mothers but in-the-world is not something that we are born with, it perpetually not fathers. The way practices compete and intersect raises questions develops over time through interaction. This process crystallises as about value and authenticity. Carrie exemplifies her frustration at the bodily memory and because of this being-in-the-world becomes in- elevated ‘value’ of her partner’s PA in comparison with her own, telligible (Merleau-Ponty, 1962). Throughout our lives we engage explaining that “he just takes off and goes cycling”. Later, she admits further in the material world, embodying new skills and techniques, how difficult the household level expectations about her own PA are thereby bringing the material world more and more into our every- to change: “it’s hard-wired as to how our daily structure works… we day lives (Dant 1999, 2004). By drawing attention to the intricate are just stuck in a daily routine I guess”. nature of socio-material relations, assemblage theory invites us to Only when there was compatibility between the external space- consider how things become active not only in shaping perceptual time configurations, everyday temporal rhythms of mothering prac- capacities but also the development of skilful embodied techniques. tices and the temporal expectations that infuse the household system By examining the case of the repairer/motorcycle assemblage, this of practices did our respondents participate in regular PA. As such, research aims to develop a theoretical understanding of how skilful our findings suggest that responsibilising mothers to participate in embodied relations with things are formed and also how these rela- PA through ‘sporty’ identity creation or resistance is unlikely to suc- tions become constitutive of embodiment. ceed and is ethically problematic. Rather, social change relating to mothers’ PA should acknowledge socio-cultural, material and above A Brief Summary of Findings all temporal characteristics of the practices which both demand and The repairer’s ability to relate to the material world of the mo- shape the everyday organisation of mothering. Insodoing, our analy- torcycle is acquired and strengthened in skill (Borgmann, 1987). sis contributes to the burgeoning critical social marketing agenda As a central socio-material dynamic of assemblage formation, skill (Gordon, 2011; 2018) which draws on social practice theory to il- development closes the distance between the repairer and the mo- luminate the socio-cultural shaping of everyday life (Gordon, 2018) torcycle. Sharpened skills bring people and things together to the ex- as a basis for conceptualising intervention. tent that they can assemble as one. This dynamic was exemplified in instances where the participants describe using tools with such ease Assembling Embodiment: Body, Techniques and Things and dexterity that they become like an extension of hand (Mellström, 2002). Being always takes place in the world, and because of this EXTENDED ABSTRACT our capacities for skill development and habituation are inextricably By positioning embodiment to the forefront of enquiry, a na- relational (Dant 2004; Merleau-Ponty, 1962). scent corpus of CCT research illuminates the sensuous and affec- Crawford (2010) describes motorcycle repair as a stochastic tive nature of transformative consumption experiences. Extending art, which necessitates a disposition that is open and attentive to the analysis beyond the focus on discourse and meaning which preoc- machine. To embody this disposition, perception must be trained in cupied the ‘interpretive turn’, these studies animate the lived experi- relations with the motorcycle. This study illuminates how novices ence of embodiment as it surfaces through intense sensory immer- are taught to perceive the motorcycle’s obscurities by developing Advances in Consumer Research (Volume 48) / 1115 an aural appreciation of its rhythms. By mindfully listening to the riences in nature, using limited leisure time and financial resources, sounds of engines, the capacity for acoustically informed diagnosis are restored and rejuvenated, able to return ‘healed’ and better able gradually accumulates. Similarly, through immersion in these activi- to deal with life in a time-stricken and ever-connected world (Can- ties, the body becomes attuned to diagnosing problems by identify- niford and Shankar 2013; Husemann and Eckhardt 2019). While we ing resonant sounds, smells and feels (Mellström, 2003). This path- may know why consumers seek out these experiences, and that they way to skill development is illustrative of the recursivity of these are restored and rejuvenated by them, how such experiences are re- socio-material relations. The motorcycle demands a finely attuned storative and rejuvenating remains less well understood. attentiveness that is receptive to its signals. But the repairer/motor- To answer this question, we use Csikszentmihalyi’s (1975) cycle assemblage, also illuminates a recursivity to socio-material concept of flow as a guiding theory. Flow, characterized as a state relations which complicates any notion of agentic subjects continu- of total absorption, diverts people’s attention from the mundane to ously acting upon inanimate objects. focus on the task at hand (Csikszentmihalyi 1975). As a guiding By tuning into the socio-material dynamics of assemblage for- framework, it sensitizes our work to the experience of temporarily mation, we see that repairers are oftentimes drawn out of themselves losing a sense of place and time. We also draw upon actor-network and into an intimate relation with the motorcycle, whereby they are theory (Law 1992) to examine how these experiences come together forced to question what it needs. In these instances, ‘technology is no as an assemblage of place, materials (equipment), and body (mind longer a means by which our mastery of the world is extended’ but and skills). We engaged in an ethnographic study of fly fishing. Fly rather it becomes ‘an affront to our usual self-absorption’ (Crawford, fishing offers a unique perspective of consumers who have a deep 2010: 16). Skilled repairers must be able to see beyond the embodied engagement with the activity—which includes the study of entomol- techniques of their craft, to see how skills can be grounded in rela- ogy, river conditions, equipment, the skills necessary to cast, and tions with the material demands of the situation. It is only with this adapt to varying weather conditions. We collected data based on the outward orientation towards fixing things that the repairer can truly ethnographic immersion of two authors of this project, participant understand their nature. It is through these socio-material relations observation, and phenomenological interviews with 21 fly fishers. In that the repairer incorporates the daily movements and actions of order to more closely capture the in-situ experience, we also engaged repair culture. in video recording participants who then reflected upon their experi- When we segment the world into imaginary categories and hi- ences further in interviews. erarchies in order to get things done, we in-effect produce dualisms Our findings show that the restorative and rejuvenative nature (Canniford and Shankar, 2015). When we resist dualisms, by alter- of these experiences relies on a deep engagement with the context, ing the ontological assumptions that guide our research, we open sometimes spanning over a lifetime or decades. We show the pro- new up new pathways to see how previously overlooked elements cess of restorative consumption occurs through an extended model constitute phenomena. Rather than slipping into the naturalised on- of flow that includes: (1) preparing for flow, (2) flowing, (3) remem- tological stance which separates mind/body and subject/object, this bering flow. First, preparing for flow involves a ritualized material, paper illuminates the socio-material processes through which mind/ embodied, and cognitive practice. Second, flowing is the phenom- body and subject/object assemble to co-constitute embodiment. In enological experience in which consumers enter a non-reflexive state the routine, habitual practices of motorcycle repair sensory percep- where embodied practice prevails. Finally, Csikszentmihalyi (1990, tions and gestural dexterity become mindfully honed to interact with 229) argues that flow happens in the “here and now”; however, we the motorcycle and its world. By illuminating the bodies aptitude find that remembering flow is a highly ritualized and integral part for knowledgeability and the material objects capacities to impose of the flow experience. The act of remembering what flow feels like demands upon the would-be-subject, these socio-material relations comes to shape consumers’ understanding of flow and future experi- blur the boundary that typically separates mind from body and sub- ences of flow. ject from object. We see cogently that the motorcycle repairer could Our research contributes an understanding of the process by not come into being without this embodied emplacement in the ma- which consumption environments and practices create rejuvena- terial world. tive experiences. First, we show that rejuvenative experiences are fleeting, and difficult to achieve and maintain. Beyond Csikszent- Extending Flow: How Place, Materials, and Body Create mihalyi’s (1975) descriptive need for an intersection between skill Restorative Consumption in Nature and challenge, our work captures the phenomenological experience of the entire process of flow. Second, consumer researchers broadly EXTENDED ABSTRACT refer to flow as akin to a broad range of experiences—transportation Consumer research documents a broad range of consumer ex- experiences (Phillips and McQuarrie 2010; Wang and Calder 2006), periences that offer restorative benefits. From extraordinary experi- edgework (Canniford and Shankar 2013), immersion (Phillips and ences in nature (Arnould and Price 1993; Canniford and Shankar McQuarrie 2010), transcendental (Schouten, McAlexander, and 2013; Celsi, Rose, and Leigh 1993), to painful (Scott, Cayla, and Koegin 2007). We bring clarity to distinguish flow from other flow- Cova 2017), or conflict-laden experiences (Tumbat and Belk 2011), like experiences that, after Turner (1969), we call flowoid. Third, drug-induced nights out (Goulding et al. 2009), pilgrimages for the self-awareness is usually considered a virtue. Literature extols the deeply devoted (Higgins and Hamilton 2019; Husemann and Eck- benefits of the reflexive consumer agentically navigating social class hardt 2019), or through natural products (Thompson 2004), such (Thompson, Henry, Bardhi 2018), however, in our study, consum- experiences are understood to ameliorate the demands of today’s ers are involved in a deep engagement to reflexively engage in non- society, the burdens of reflexivity, and the resulting “saturated self” reflexivity. In juxtaposition with extraordinary experiences literature (Husemann and Eckhardt 2019; Scott, Cayla, and Cova 2017, 24). that shows that consumers need an escape to achieve the restorative We draw into question what many of these consumption experi- benefits of flow, we suggest that it is through a deep immersion that ences share in common: they take place in nature. A common trope consumers reap these benefits. Lastly, we draw to attention the im- in consumer research is that nature is imbued with healing qualities portance of the availability of a place to flow. Sustaining natural en- (Canniford and Shankar 2013). Consumers who seek out such expe- 1116 / The Body and Equipment: Understanding Consumption Through Relationships Between Body and Equipment vironments that facilitate such flow experiences, and making them Dant, T. (1999) ‘Material culture in the social world’, McGraw-Hill readily available to populations is important. Education (UK). Dant, T. (2004a) ‘The Driver-car’, Theory, Culture & Society REFERENCES 21(4–5): 61–79. Arnould, Eric J. and Linda L. Price (1993), “River Magic: Dant, T. (2004b) ‘Materiality and society’, McGraw-Hill Education Extraordinary Experience and the Extended Service (UK). Encounter,” Journal of Consumer Research, 20 (June), 24–45. Dawson, Marcelle C. (2017), “CrossFit: Fitness Cult or Reinventive ______(2000), “Authenticating Acts and Authoritative Institution?,” International Review for the Sociology of Sport, Performances: Questing for Self and Community,” in The Why 52(3), 361–79. of Consumption: Contemporary Perspectives on Consumer Foucault, Michel (2008), The Birth of Biopolitics: Lectures at Motives, Goals and Desires, ed. S. Ratneshwar, David Glen the College de France, ed. Michael Senellart, New York: Mick, and Cynthia Huffman, New York: Routledge, 140–63. Palgrave. Atkinson, Michael (2008), “Triathlon, Suffering and Exciting Giesler, Markus and Ela Veresiu (2014), “Creating the Responsible Significance,”Leisure Studies, 27(2), 165–80. Consumer: Moralistic Governance Regimes and Consumer Beck, Ulrich (1999), World Risk Society, Malden, MA: Polity Press Subjectivity,” Journal of Consumer Research, 41(3), 840–57. and Blackwell Publishers. Gordon, R. (2011), Critical social marketing: Definition, application ______(2006), Cosmopolitan Vision, Malden, MA: Polity Press. and domain. Journal of Social Marketing, 1, 82–99. Beck, Ulrich, Anthony Giddens, and Scott Lash (1994), Reflexive Gordon, R. (2018), Critical Social Marketing: Reflections, Modernization: Politics, Tradition and Aesthetics in the introspections and future directions, chapter 5, in ed. Modern Social Order, Stanford, CA: Stanford University Tadajewski, M., Higgins, M., Denegri-Knott, J. and Varman, Press. R. (2018), Critical Marketing, Routledge. Bellow-Riecken, KH. and Rhodes, R.E. (2008), A birth of Goulding, C., Shankar, A., Elliott, R. and Canniford, R. (2009) inactivity? A review of physical activity and parenthood, ‘The marketplace management of illicit pleasure’, Journal of Preventative Medicine, 46(2): 99-110 Consumer Research 35(5): 759-771. Borgmann, A. (1987) ‘Technology and the character of Hamilton, K. and White, K.M. (2010), Understanding parental contemporary life: A philosophical inquiry’, Chicago: physical activity: Meanings, habits and social role influence, University of Chicago Press. Psychology of Sport and Exercise, 11, 275-285. Caillois, Roger (2001), Man, Play and Games, Urbana and Herz, J.C. (2014), Learning to Breathe Fire: The Rise of CrossFit Chicago, IL: University of Illinois Press. and the Primal Future of Fitness, New York: Three Rivers Canniford, Robin, and Avi Shankar (2013), “Purifying Practices: Press. How Consumers Assemble Romantic Experiences of Nature,” Hewer, P. and Hamilton, K. (2010) ‘On emotions and salsa: Journal of Consumer Research, 39 (5), 1051-1069. Some thoughts on dancing to rethink consumers’, Journal of Canniford, Robin and Shankar, Avi (2015) “Post-Dualistic Consumer Behaviour 9(2): 113-125. Consumer Research: Nature-Cultures and Cyborg Heywood, Leslie (2016), “‘We’re in This Together:’ Neoliberalism Consumption,” in Assembling Consumption, pp. 147-164. and the Disruption of the Coach/Athlete Hierarchy in London: Routledge. CrossFit,” Sports Coaching Review, 5(1), 116–29. Canniford, R., Riach, K., and Hill, T. (2018) ‘Nosenography: How Higgins, Leighanne, and Kathy Hamilton (2019), “Therapeutic smell constitutes meaning, identity and temporal experience in Servicescapes and Market-Mediated Performances of spatial assemblages’, Marketing Theory 18(2): 234–248. Emotional Suffering,”Journal of Consumer Research, 45(6), Cederström, Carl and Andre Spicer (2015), The Wellness Syndrome, 1230-1253. Malden, MA: Polity Press. Husemann, Katharina C., and Giana M. Eckhardt (2019), Celsi, Richard L., Randall L. Rose, and Thomas W. Leigh (1993), “Consumer Deceleration,” Journal of Consumer Research, 45 “An Exploration of High-Risk Leisure Consumption through (6), 1142-1163. Skydiving,” Journal of Consumer Research, 20(1), 1–23. James, Eric P. and Rebecca Gill (2017), “Neoliberalism and the Cooperman, Stephanie (2015), “Getting Fit, Even If It Kills You,” Communicative Labor of CrossFit,” Communication & Sport, The New York Times, December 22. 6(6), 1–25. Crawford, M. (2010) The Case for Working with Your Hands: Or Joy, Annamma, and John F. Sherry Jr. (2003), “Speaking of Art why office work is bad for us and fixing things feels good. as Embodied Imagination: A Multisensory Approach to London: Penguin . Understanding Aesthetic Experience,” Journal of Consumer Crockett, Matt C. and Ted Butryn (2018), “Chasing Rx: A Spatial Research, 30 (2), 259-282. Ethnography of the CrossFit Gym,” Sociology of Sport Law, John (1992), “Notes on the Theory of the Actor-Network: Journal, 35(2), 98–107. Ordering, Strategy, and Heterogeneity,” Systems Practice, 5 Csikszentmihalyi, Mihaly (1975), Beyond Boredom and Anxiety. (4), 379-393. San Francisco: Jossey-Bass Publishers. Le Breton, David (2000), “Playing Symbolically with Death in Czikszentmihalyi, Mihaly (1990), Flow: The psychology of optimal Extreme Sports,” Body & Society, 6(1), 1–11. experience. New York: Harper & Row. Lupton, D. (2015). Quantified sex: a critical analysis of sexual Datta, Anisha and Indranil Chakraborty (2017), “Are You and reproductive self-tracking using apps. Culture, health & Neoliberal Fit? The Politics of Consumption under sexuality, 17(4), 440-453. Neoliberalism,” in The SAGE Handbook of Consumer Culture, Maciel, A. and Wallendorf, M. (2017) ‘Taste Engineering: ed. Olga Kravets, Pauline Maclaran, Steven Miles, and Alladi An Extended Consumer Model of Cultural Competence Venkatesh, Thousand Oaks, CA: Sage, 453–77. Constitution’, Journal of Consumer Research 43(5): 726-746. Advances in Consumer Research (Volume 48) / 1117

Mauss, M. (1973) ‘Techniques of the Body’, Economy and Society Schouten, John W., James H. McAlexander, and Harold F. Koenig 2(1): 70–88. (2007), “Transcendent Customer Experience and Brand McGannon, K.R. and Schinke, R.J. (2013), “My first choice is Community,” Journal of the Academy of Marketing Science, to work out at work; then I don’t feel bad about my kids”: a 35 (3), 357-368. discursive psychological analysis of motherhood and physical Scott, Rebecca, Julien Cayla, and Bernard Cova (2017), “Selling activity participation, Psychology of Sport and Exercise, 14, Pain to the Saturated Self,” Journal of Consumer Research, 179-188. 44(1), 22–43. Mellström, U. (2003) ‘Masculinity, Power and Technology: A Seregina, A. and Weijo, H. (2016) ‘Play at Any Cost: How Malaysian Ethnography, Hampshire: Ashgate. Cosplayers Produce and Sustain Their Ludic Communal Merleau-Ponty, M. (1962) Phenomenology of Perception. London: Consumption Experiences’, Journal of Consumer Research Routledge. 44(1): 139–159. Molander, S. and Hartmann, B. (2018) Emotion and practice: Shove, E. (2012) Habits and their creatures. In: Warde A and Mothering, Cooking, and Teleoaffective Episodes,Marketing Southerton D (eds), The Habits of Consumption. Helsinki: Theory, 18, 3, 371-390 Helsinki Collegium for Advanced Studies. Murphy, S., Patterson, M. and O’Malley , L. (2018) ‘Learning Silva E.B. (2002), Routine matters: Narratives of everyday life in how: Body techniques and the consumption of experience’, families. In: Crow Gand Heath S (eds) Social Conceptions Marketing Theory 19(4): 425-445. of Time: Structure and Process in Work and Everyday Life. Nash, Meredith (2018), “‘Let’s Work on Your Weaknesses’: Basingstoke, Hampshire: Palgrave Macmillan, pp. 179–194. Australian CrossFit Coaching, Masculinity and Neoliberal Southerton, D. (2006), Analysing the temporal organisation of Framings of ‘Health’ and ‘Fitness,’” Sport in Society, 21(9), daily life: social constraints, practices and their allocation, 1432–53. Sociology, 40(3) Patterson, M. and Larsen, G. (2018) ‘Listening to consumption: Thompson, Craig J. (2004), “Marketplace Mythology and Towards a sonic turn in consumer research’, Marketing Theory Discourses of Power,” Journal of Consumer Research, 31 (1), 19(2): 1-23. 162-180. Patterson, M., & Schroeder, J. (2010). Borderlines: Skin, tattoos Thompson, C. J., & Üstüner, T. (2015). Women skating on the edge: and consumer culture theory. Marketing Theory, 10(3), 253- Marketplace performances as ideological edgework. Journal 267. of Consumer Research, 42(2), 235-265. Pekkanen, Anna, Elina Närvänen, and Pekka Tuominen (2018), Thompson, Craig J., Paul C. Henry, and Fleura Bardhi (2018), “Elements of Rituality in Consumer Tribes: The Case of “Theorizing Reactive Reflexivity: Lifestyle Displacement CrossFit,” Journal of Customer Behaviour, 16(4), 353–70. and Discordant Performances of Taste,” Journal of Consumer Phillips, Barbara J., and Edward F. McQuarrie (2010), “Narrative Research, 45 (3), 571-594. and Persuasion in Fashion Advertising,” Journal of Consumer Tumbat, Gülnur, and Russell W. Belk (2011), “Marketplace Research, 37 (3), 368-392. Tensions in Extraordinary Experiences,” Journal of Consumer Rabinow, Paul (2009), Anthropos today: Reflections on modern Research, 38 (1), 42-61. equipment. Princeton University Press. Turner, Victor (1969), “Liminality and Communitas,” The ritual Rhodes, RE., Blanchard CM., Benoit, C. et al. (2014), Physical process: Structure and anti-structure, 94, 130. activity and sedentary behaviour across 12 months in cohort Wang, Jing, and Bobby J. Calder (2006), “Media Transportation samples of couples without children, expecting their first and Advertising,” Journal of Consumer Research, 33 (2), child, and expecting their second child, journal of behavioural 151-162. medicine, 37(3), 533-542. Wearing, B. (1999), Beyond the Ideology of Motherhood: Leisure Roux, Dominique, and Russell Belk (2019), “The Body as as Resistance, ANZJS, 26(1), March, 36-58 (Another) Place: Producing Embodied Heterotopias Through Woermann, N. and Rokka, J. (2015) ‘Timeflow: How Consumption Tattooing,” Journal of Consumer Research, 46 (3), 483-507. Practices Shape Consumers’ Temporal Experiences’, Journal Saville, Stephen John (2008), “Playing with Fear: Parkour and the of Consumer Research 41(6): 1486–1508. Mobility of Emotion,” Social & Cultural Geography, 9(8), Zanette, M. C., & Scaraboto, D. (2019) “‘To Spanx or not to 891–914. Spanx’: How objects that carry contradictory institutional Saxbe, D., Rossin-Slater, M. and Goldenberg, D. (2018), The logics trigger identity conflict for consumers,”Journal of transition to parenthood as a critical window for public health, Business Research, 105, 443-453. American Psychologist, 73(9) Schneider, Terri, Ted Butryn, David Furst, and Matthew Masucci (2007), “A Qualitative Examination of Risk among Elite Adventure Racers,” Journal of Sport Behavior, 30(3), 330–57. Reimagining Pro-Social Behaviors through Micro-, Meso-, and Macro-Level Perspectives of Charitable Giving Charis: Lez Trujillo Torres, University of Illinois at Chicago, USA Akon E. Ekpo, Loyola University Chicago, USA

Paper #1: The Influence of Public Discourse on Charitable behaviors. Through a series of investigations of charitable giving, Giving this special session seeks to fill this gap and addresses the following: Lez Trujillo Torres, University of Illinois at Chicago, USA • What role(s) does the marketplace play in enacting pro- Yingting Wen, ESSEC Business School, France social behaviors? Akon E. Ekpo, Loyola University Chicago, USA • How does public discourse shape the types of pro-social Benét DeBerry-Spence, University of Illinois at Chicago, USA causes we value and thus contribute to? • How does pro-social and market logics shape meanings of Paper #2: Bodies as Gardens, Bodies as Machines: Charitable pro-social behaviors and thus decisions to enact them? Giving as Coping Rebecca Scott, Cardiff University, UK According to Giving USA (2019), total charitable giving rose Tonya Williams Bradford, University of California – Irvine, 0.7% over a total of $424.74 billion contributed in 2017. Interest- USA ingly, while giving by individuals decreased by 1.1% (for a total estimate of $292.09 billion), giving by foundations and corporations Paper #3: A Divine Duty: Commercial Surrogacy in India and increased by 7.3% and 5.4% respectively, leaving giving by bequests Kin Altruism flat. Though individual giving has decreased, individual giving re- Sujit Raghunathrao Jagadale, Xavier University Bhubaneswar, mains the highest form of charitable giving, accounting for 68% of India the total giving. Thus, charitable giving is an ideal context to exam- Lez Trujillo Torres, University of Illinois at Chicago, USA ine pro-social behaviors across multiple levels, journeying from the Laetitia Mimoun, Cass Business School, City, University of micro-level: A Divine Duty: Commercial Surrogacy in India and Kin London, UK Altruism (individual). To the mezo-level: Bodies as Gardens, Bodies Paper #4: When Pro-social and Market Logics Collide: the as Machines: Charitable Giving as Coping (family). And finally, to Sensemaking Journeys of Egg Donors in the American Infertility the macro-level: The Influence of Public Discourse on Charitable Field” Giving (societal) and When Pro-social and Market Logics Collide: Anna E. Hartman, University of Melbourne, Australia the Sensemaking Journeys of Egg Donors in the American Infertility Eileen Fischer, York University, Canada Field (market). This session will be of interest to pro-social, gift-giving, and SESSION OVERVIEW healthcare marketing scholars. Increasingly, scholars are interested in understanding the motivating factors and consequences of consumers engaging in pro-social The Influence of Public Discourse on Charitable Giving consumer behavior. In fact, a recent special issue of the Journal of the Association for Consumer Research points to an 848% increase EXTENDED ABSTRACT in such studies from 2000 – 2018 (White, Habib, and Dahl 2020). This research investigates the influence of public discourse Pro-social behavior refers to behaviors that consumers engage in that on charitable giving and how this evolves over time. The context may come at some cost to the self for the greater good or benefit for of our study is the United States charitable marketplace, the larg- others (Small and Cryder 2016; White et al. 2020); and can include est philanthropic market in the world, between the years of 1991 to behaviors such as: charitable giving, volunteering, altruistic behav- 2015. We contribute to a growing literature devoted to understanding iors, ethical purchasing, engagement in cause-related activities, and how individuals, through monetary contributions to nonprofit orga- advocacy or activism. For the most part, motivations for pro-social nizations, address pressing societal issues, such as poverty, hunger, consumer behaviors have dealt with individual and situational fac- public health, animal welfare, and socioeconomic and environmental tors mainly from a micro-level perspective. From an individual per- justice. The significance of these issues, alongside the rise in chari- spective, factors such as maintaining a positive image (Aaker and table giving to the nonprofit sector, makes the focus of this research Akutsu 2009), self-concept (Grant and Dutton 2012), moral self- important. To this point, in 2018 alone, the nonprofit sector hit a worth (Jordan, Mullen, and Murnighan 2011; Sachdeva, Iliev, and new record with $410 billion U.S. dollars in contributions distributed Medin 2009), or acting in one’s self-interest through charitable giv- across approximately 10 million nonprofit organizations worldwide ing may create positive feelings that encourage pro-social behaviors (Giving USA 2018; Salamon and Newhouse 2019). (Chang and Hung 2018). From a situational perspective, factors such By large, extant research encompasses two major influences as whether one may socially benefit from public displays of pro- on charitable giving. The first influence deals with individual do- social behavior (Ariely, Bracha, and Meier 2009; White and Peloza nor characteristics such as altruistic identity, morality, and personal 2009), situational norms surrounding engagement in pro-social be- values of the donor (e.g., Boenigk and Helmig 2013; Nilsson, Er- havior (Frey and Meier 2004; Shang, Reed, and Croson 2008); and landsson, and Västfjäll 2016; van Dijk, Van Herk, and Prins 2019). attempts to gain a sense of belonging (Lee and Shrum 2012; Twenge The second influence deals with the strategic action of and within et al. 2007) all contribute to whether and how someone engages in nonprofit organizations and includes the effectiveness of rhetorical pro-social behaviors. communication appeals, perceived characteristics of the organiza- While these studies provide insight into why individuals choose tion, and the entrepreneurship of institutional actors (e.g., Botner, to engage in pro-social behaviors, there remains a limited under- Mishra, and Mishra 2015; Gautier, Pache, and Chowdhury 2016; standing of the role of meso- and macro-level factors vis-à-vis these Sargeant, Ford, and West 2006). Surprisingly, little is known about

Advances in Consumer Research 1118 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1119 how macro factors, such as public discourse, which both reflect and Bodies as Gardens, Bodies as Machines: Charitable influence historic, regulatory, socioeconomic, and political issues, Giving as Coping impact charitable giving over long periods of time. In addressing this gap, we show how charitable behavior emerges and evolves as EXTENDED ABSTRACT public discourse also changes. And, consequently we provide a more Individuals source solutions to problems in the marketplace. comprehensive understanding of the dynamism associated with These problems range from basic provisioning and entertainment charitable behavior. to broader issues such as managing significant life transitions (e.g., Our research methods involved a multi-method approach. We marriage, death) and addressing societal issues (e.g., homeless- collected charitable giving patterns of the top 400 charities in the ness, health crisis). We investigate charitable giving as a means of US (1991-2015) and related fundraising expenses. This data served addressing societal issues, with a focus on the shortage of organ to operationalize actual historical charitable behavior and the rank- donation for transplantation. ing of organizations based on total contributions received in a given Charitable giving is recognized as a means for individuals to year. We also utilized three archival data sources of public discourse. provide help, hope, and life to others in need (Sherry 1983). The First, we used a consumer opinion dataset that provide an aggregated psychological literature examines roles for individual differences picture of what issues or charitable causes consumers value. Second, and empathy to inspire charitable behaviors (Lee and Bradford we used archival data from popular news media including newspa- 2015), and sociocultural perspectives examine such gifts in relation pers, magazines, TV, radio, podcasts, to identify, quantitatively and to their impact on society (Bradford 2013; Strathern 2012; Titmuss qualitatively, what issues and topics the U.S. society values. Third, 1997). While benefits of charitable giving accrue to those in need, we collected information about the top 400 charitable organizations, donors may be motivated to participate to attain any number of including advertisements in high-circulation news media to assess benefits such as enhancing the well-being for others, or attaining the strength of and/or interaction between public discourse and orga- benefits (e.g., tax, career advances) for themselves (Bendapudi, nizational strategic actions on charitable giving. Singh, and Bendapudi 1996). The literature provides insights into The quantitative analysis of datasets includes correlations, t- how donors may partner with organizations to support those in tests, and automated content analysis. The qualitative analysis of need. Missing is an understanding of how donors may employ news media data and charitable data (e.g., textual like articles, tex- charitable giving as a means of coping in response to loss. tual/visual like advertisements) follows a hermeneutical approach Consumer researchers examine various forms of loss and (Thompson 1997) to interpret patterns of meanings in these data. We roles for consumption related to loss. Individuals experience loss integrate both approaches by following Sanders’ (1982) recommen- in various forms throughout life—jobs, relationships with people dation to identify and reflect on emerging patterns, and to derive key or brands, or the transference of possessions from one owner to themes from the team’s reflections. We use the program MaxQDA to another. Prior research finds that voluntary losses may be managed transcribe analytical memos and analyze both qualitative and quan- as individuals prepare for disposition of their wealth or special titative data. possessions (Bradford 2009; Price, Arnould, and Curasi 2000; Our preliminary results indicate that there is not a significant Winterich, Reczek, and Irwin 2017). And though less prevalent, correlation between fundraising expenses and donations received there are some considerations in the literature of involuntary loss by nonprofit organizations, from 1991 to 2015, across 18 categories such as with homelessness (Hill and Stamey 1990), illness (Wong of charities (e.g., social services, children & youth, health, public and King 2008), or death (Bonsu and Belk 2003). What is the rela- affairs, international, arts & culture, education, and environment tionship between loss and charitable giving? The present research & animal welfare). Rather, we find that three interrelated types of examines the traumatic experiences of loved ones’ sudden loss. public discourse played a role in the rise of donations received by More specifically, we examine the experience of deceased family nonprofit organizations in the US: consumer opinions of pressing (DDF) members who participate in charitable giving by donating societal problems, earned media by nonprofit organizations, and ad- their loved one’s organs to others in need. More specifically, our vertisements by non-profit organizations in high-circulation news research examines how experiences of charitable giving may serve media outlets. For instance, we find a significant correlation between as a means of coping when individuals are faced with loss. pressing societal problems in the mind of consumers and donations The relationship between the marketplace and coping find that received by organizations focused on those issues in 1991-1999. In the market may be both a source of and a relief from stress. For addition, over time we find that consistently nonprofits in the social example, scholars examine how consumers cope with stress in the services and college education categories have received combined marketplace, such as with unsatisfactory experiences (Duhachek the highest amounts of earned media and advertised across high- 2005), brand termination (Russell and Schau 2015), difficult deci- circulation news media. sions (Luce, Bettman, and Payne 1997), or bias (Bone, Christensen, Our research provides two key contributions. First, we expand and Williams 2014). Further, scholars find consumers employ the the literatures on prosocial behavior and value by providing evidence marketplace as a means of coping to manage life transitions, such of public discourse-driven charitable behavior and their resulting hi- as divorce (Yap & Kapitan, 2017), or relieve temporary stress, as erarchies of value (e.g., pressing societal issues, charitable causes, through retail therapy (Atalay & Meloy, 2011; Townsend & Sood and non-profit organizations). In doing so, we extend the types of 2012). It is evident the marketplace provides opportunities for con- antecedents and consequences of prosocial behavior examined thus sumers to cope with life stressors. While prior research considers far in consumer research. Second, we contextualize patterns of indi- common life stressors and roles for the marketplace in consumer vidual charitable behavior and embed them in macro historical con- coping, the present research examines sociocultural influences texts, thus addressing calls for contextualization of micro consumer to explain how individuals employ the marketplace to cope with behavior (Askegaard and Linnet 2011). trauma. We conduct an ethnographic study with deceased organ donor family members to interrogate experiences of involuntary loss. We find three themes: transforming loss into gifts of comfort; crafting 1120 / Reimagining Pro-Social Behaviors through Micro-, Meso-, and Macro-Level Perspectives of Charitable Giving new futures from loss; and, the marketplace and the sacred. De- service providers and politicians, and systematically collected and ceased donor families contemplate the life of their loved one, and the analyzed articles on the topic published in four leading daily news- potential of some part of the body living even as the essence of their papers. We rely on prolonged engagement, purposive sampling, and loved one is dead. For those families, sacralization is evident as they thick description to address the issue of trustworthiness (Belk, Fisch- recognize continuity of life through a loved one: er, and Kozinets 2013). Our initial findings suggest that surrogates construe their mar- Good, really positive that something good comes out of something ket participation as two forms of kin altruism (Ashton et al. 1998; so tragic, it doesn’t seem a waste. In some way it’s better than if Lee, Winterich, and Ross Jr 2014; Small and Simonsohn 2008) someone just dies, they’re cremated or buried and that’s it. He’s still downplaying the pecuniary aspects related to it. Kin altruism refers out there, part of him is still out there. When you’ve come from the to the “tendency to feel empathy and attachment toward others. This family that we have, we’ve all worked helping other people in our is based on the deduction that one is more likely to behave altruisti- lives, it’s a natural thing to do to, go on a do it when you’re dead. cally toward one’s kin if one considers costs and benefits to those Totally unselfish. ~ Jade kin, and not only costs and benefits to oneself, when deciding among behavioral alternatives” (Ashton et al. 1998). On one hand, surro- Deceased donor family members begin to consider parts of the gates adopt a discourse that invokes existing and popular sociocul- body and how body parts may provide life to others. In so doing, tural and religious symbols, which aims to remove social-cultural the body itself is recognized as sacred as a whole as well as each of differences between them and non-surrogates in India. For instance, its parts due to its life-giving forces. We extend prior understand- several participants equate themselves with Yashoda, a foster mother ings of consumer coping with an explanation of how charitable of mythological Hindu God Lord Krishna. In doing so, they empha- giving facilitates the transformation of loss from that of suffering size the responsibility put on them by their divinity to help the needy into new futures. and deficient. On the other hand, surrogates create kin-like connections with the commissioning mothers, often with the aid of market- A Divine Duty: Commercial Surrogacy in India and Kin place intermediaries such as health clinics and for-profit agencies. Altruism By adopting empathetic and deep sisterhood connections, surrogates aim to remedy the deficiencies (i.e., infertility) of commissioning EXTENDED ABSTRACT mothers, which enable them to justify and manage their own bodily, The emergence and global expansion of Assisted Reproduc- social, and cultural sacrifices. Through these discursive constructive Technologies (ART) continues to challenge deeply held values tions, commercial surrogates attempt to challenge and to rectify the and beliefs regarding reproduction, parenthood and family across increasingly disdainful societal attitudes toward gestational carriers the world (Sobande, Mimoun, and Trujillo Torres 2019; Takhar and and influence policymakers. Rika Houston 2019). While certain types of ART such as in vitro Our work contributes to understanding how constructions of fertilization have become more accepted(Fischer, Otnes, and Tuncay altruism are used to justify the participation in a market exchange 2007), commercial surrogacy remains a highly controversial (Tada- process in dramatically under-researched context (Belk 2016; Tada- jewski 2016), and regulated practice characterized by perceptions of jewski 2016), which is characterized by vastly different sociocul- exploitative commodification and (dis)empowerment of gestational tural, economic, political, and legal values and systems than the carriers (Fenton-Glynn 2019). This industry, now estimated at 6 bil- typical free market contexts seen in the literature. Furthermore, our lion dollars (USD), is subject to moral, ethical, legal, and political work addresses the justification of actors engaging in an increasingly debates, especially in countries with large numbers of socioeconomi- regulated and stigmatized market exchange, while most of the prior cally vulnerable women (Varman et al., 2016). literature considers market actors’ practices in contexts which are In public discourse surrounding international reproductive tour- destigmatizing (e.g., Sandikçi and Ger 2010; Scaraboto and Fischer, ism, commercial surrogacy is often construed as dirty work (Pande 2013). By doing so, our work highlights how market participation in 2010) and surrogates are derided and silenced for participating in stigmatized market exchange is emphasized as a prosocial behavior this exchange process. Moreover, commercial surrogates operate helping deficient consumers. This work also broadens the growing as commodified consumers and stigmatized producers in a market literature on assisted reproductive technologies markets by show- system where the exchange boundaries between these two roles are ing how constructions of charitable giving intersects with the emer- blurred. That is, they consume conventional health services and act gence, growth, and eventual decline of an international biotechnol- as value-producing actors within socio-cultural contexts that harbor ogy market. negative attitudes towards surrogate’s participation. Because commercial surrogacy, especially in an emerging market context like When Pro-social and Market Logics Collide: the India, is conceived as a stigmatized exchange, market actors cannot Sensemaking Journeys of Egg Donors in the American rely on mainstream discourses to legitimize their participation in this exchange. Particularly, in the Indian context, the public discourse Infertility Field associated with gestational surrogacy is increasingly influenced by EXTENDED ABSTRACT ongoing regulatory actions to ban surrogacy and curb the growth of In fields characterized by institutional complexity, where mul- this industry. Thus, our project investigates how surrogates, as con- tiple and often contradictory logics are at play, the prescriptions and sumers and co-producers, construct and justify their participation in proscriptions of differing logics can lead to tensions in the market- an increasingly regulated and stigmatized exchange process. place (e.g., Ertimur and Coskuner-Balli 2015; Scaraboto and Fischer To address our research question, we conducted an examina- 2013) . These tensions are likely to be particularly acute in contexts tion of gestational surrogate women in the heart of the international, characterized by both market and pro-social logics (e.g., Beverland commercial surrogacy industry in Gujarat, India. Data were collect- and Luxton 2005; Hartman and Coslor 2019). While considerable ed primarily through nineteen in-depth interviews with commercial research has studied how formal organizations such as social en- surrogates and their families over a period of eighteen months. To terprises navigate competing market and prosocial logics (e.g., Jay triangulate this initial dataset, we also interviewed eight local health Advances in Consumer Research (Volume 48) / 1121

2013; McMullen and Bergman 2017) less research has examined done this? How could you have given your babies away?’ But they how individual actors who are not members of formal organizations weren’t babies, you know. It wasn’t like I was putting them up for manage the multiple logics that confront them as they interact with adoption, these weren’t mine. These were just things that we pro- other market actors and engage in exchanges. duce each month.” Another stated. “It was like having an old piece In this study, we examine a particular category of actors who of furniture in your house that you don’t need and somebody said, engage in exchanges in a context characterized by competing pro- ‘Hey, I’ll give you $8,000 for it.’ I was planning on getting rid of it social and market logics: we study women who sell their ova or eggs anyway.” This market-logic influenced way of making sense of their to buyers in the U.S. infertility industry. We do so to understand how own egg donation was compounded for those who became “known they make sense of their role in the marketplace, and how their sen- donors,” a term used to describe those whose previously donated semaking influences their actions. This paper thus contributes to a eggs had reliably resulted in the birth of healthy babies. These indi- better understanding of the implications of logic multiplicity in gen- viduals could command higher compensation levels; choosing to do eral, and of conflicting pro-social and market logics in particular. so would be difficult to reconcile with a pro-social logic. Since the mid 1980’s, the demand for donated human eggs has Our data analysis suggests, however, that not all egg donors steadily increased as a solution to certain types of infertility diagno- came to rely purely on market logics to rationalize their behaviors. ses. Egg donors are usually young, college-aged women recruited by In particular, those donors who exerted their agency to connect with market intermediaries (donor egg agencies) or fertility clinics to be the recipients who purchased their eggs, and particularly with their matched with potential recipients. Historically, donors would work biological offspring, often tended to circle back to a reliance on both directly with matched recipients to provide eggs through a fresh pro-social and market logics to make sense of their past behaviors cycle and embryo transfer to the recipient. More recently, due to in- and plan their future actions. As a tentative theoretical insight, we novations in Assisted Reproduction Technologies (ART), donors can posit that when actors can engage in meaningful social relations in also provide eggs without a matched intended recipient, where eggs contexts of the type described here, they are more likely to find pro- are retrieved for frozen egg banking to be used at a later time. While social logics and market logics similarly salient. In contrast, those there is no data available on how many women have donated their whose exchanges commence and remain anonymous are less likely eggs, nor how many cycles have been performed on donors, we do to draw on pro-social logics to rationalize their actions and more apt know that in 2016, 24,300 ART cycles were performed for recipients to rely solely on market logics for sensemaking. using donor eggs (CDC, 2018). The prosocial logic in this context is reflected in terminology that frames women who provide eggs as REFERENCES ‘donors’ and ‘angels’ who ‘give the gift of life’ (Almeling 2007). Aaker, Jennifer L and Satoshi Akutsu (2009), “Why Do People At the same time, the market logic in this context is everywhere in Give? The Role of Identity in Giving,” Journal of consumer evidence, and is reflected in discourse regarding how egg donors can psychology, 19 (3), 267-70. ‘make $10,000’ and ‘pay off their bills/tuition’ (Hartman and Coslor Almeling, Rene (2007), “Selling Genes, Selling Gender: Egg 2019). Agencies, Sperm Banks, and the Medical Market in Genetic Depth interviews were conducted with 11 women who had Material,” American Sociological Review, 72 (3), 319-40. donated their eggs in the US medical marketplace and received Ariely, Dan, Anat Bracha, and Stephan Meier (2009), “Doing Good compensation. Additionally, online observations were conducted in or Doing Well? Image Motivation and Monetary Incentives in online forums where egg donors and potential recipients discuss sa- Behaving Prosocially,” American Economic Review, 99 (1), lient issues such as compensation, motivation and matching. Data 544-55. were analyzed in an iterative fashion, that involved both initial open Ashton, Michael C, Sampo V Paunonen, Edward Helmes, and coding and multiple stages of higher order coding, guided by the Douglas N Jackson (1998), “Kin Altruism, Reciprocal research goals of this paper and abetted by the enabling lens of in- Altruism, and the Big Five Personality Factors,” Evolution and stitutional theory. Human Behavior, 19 (4), 243-55. In the initial stages of considering egg donation, when partici- Askegaard, Søren and Jeppe Trolle Linnet (2011), “Towards an pants were first engaging with recruitment advertising or egg donor Epistemology of Consumer Culture Theory: Phenomenology agencies, many drew on both prosocial and market logics to make and the Context of Context,” Marketing Theory, 11 (4), 381- sense of the possibility of becoming a donor. While some partici- 404. pants stressed altruism as a primary rationale consistent with a pro- Belk, Russell (2016), “Anthropomorphism and Anthropocentrism,” social logic and referred to financial compensation as “a nice bonus,” ACR North American Advances. others explained their initial motivation as primarily economic. Most Belk, Russell W., Eileen Fischer, and Robert V. Kozinets (2013), informants invoked both logics to rationalize their initial involve- Qualitative Consumer & Marketing Research, London: SAGE. ment in the industry. Bendapudi, Neeli, Surendra N. Singh, and Venkat Bendapudi As the participants moved from considering egg donation to (1996), “Enhancing Helping Behavior: An Integrative actually engaging in the practice, balanced sensemaking typically Framework for Promotion Planning,” Journal of Marketing, gave way to an a nearly exclusive focus on the market logic. At least 60 (3), 33-49. partially in response to interactions with medical personnel that were Beverland, Michael and Sandra Luxton (2005), “Managing experienced as highly depersonalizing, many described objectify- Integrated Marketing Communication (Imc) through ing themselves, specifically as a provider of eggs responding to a Strategic Decoupling: How Luxury Wine Firms Retain Brand market demand. While we know that some young women perceive Leadership While Appearing to Be Wedded to the Past,” their eggs as potential babies, (Rauscher et al. 2017), most of our Journal of Advertising, 34 (4), 103-16. participants sought to distance themselves from any emotional attachment to their the eggs. They tended to view their eggs as expend- able “things” that they could monetize. For example, one participant recalled a conversation where she was asked: “‘How could you have 1122 / Reimagining Pro-Social Behaviors through Micro-, Meso-, and Macro-Level Perspectives of Charitable Giving

Boenigk, Silke and Bernd Helmig (2013), “Why Do Donors Hill, Ronald Paul and Mark Stamey (1990), “The Homeless in Donate? Examining the Effects of Organizational America: An Examination of Possessions and Consumption Identification and Identity Salience on the Relationships Behaviors,” Journal of Consumer Research, 17 (3), 303-21. among Satisfaction, Loyalty, and Donation Behavior,” Journal Jay, Jason (2013), “Navigating Paradox as a Mechanism of Change of Service Research, 16 (4), 533-48. and Innovation in Hybrid Organizations,” Academy of Bone, Sterling A., Glenn L. Christensen, and Jerome D. Williams management journal, 56 (1), 137-59. (2014), “Rejected, Shackled, and Alone: The Impact of Jordan, Jennifer, Elizabeth Mullen, and J Keith Murnighan (2011), Systemic Restricted Choice on Minority Consumers’ “Striving for the Moral Self: The Effects of Recalling Past Construction of Self,” Journal of Consumer Research, 41 (2), Moral Actions on Future Moral Behavior,” Personality and 451-74. Social Psychology Bulletin, 37 (5), 701-13. Botner, Keith A, Arul Mishra, and Himanshu Mishra (2015), Lee, Angela Y. and Tonya Williams Bradford (2015), “The Effects “What’s in a Message? The Longitudinal Influence of a of Self-Construal Fit on Motivation, Attitudes, and Charitable Supportive Versus Combative Orientation on the Performance Giving,” in Handbook of Culture and Consumer Behavior, ed. of Nonprofits,” Journal of Marketing Research, 52 (1), 39-55. Sharon Ng and Angela Y. Lee, New York: Oxford, 135-49. Bradford, Tonya Williams (2013), “From Twins to Strangers: Lee, Jaehoon and LJ Shrum (2012), “Conspicuous Consumption Considerations of Paired Kidney Donation across Gift and Versus Charitable Behavior in Response to Social Exclusion: Market Economies,” in Marketing and the Common Good, ed. A Differential Needs Explanation,”Journal of Consumer John F. Sherry, Jr. and Patrick E. Murphy, 253-66. Research, 39 (3), 530-44. Centers for Disease Control and Prevention, American Society for Lee, Saerom, Karen Page Winterich, and William T Ross Jr (2014), Reproductive Medicine, and Society for Assisted Reproductive “I’m Moral, but I Won’t Help You: The Distinct Roles of Technology (2018), “2016 Assisted Reproductive Technology Empathy and Justice in Donations,” Journal of Consumer National Summary Report,” ed. US Dept of Health and Research, 41 (3), 678-96. Human Services, Atlanta (GA): US Dept of Health and Human Luce, Mary Frances, James R. Bettman, and John W. Payne (1997), Services. “Choice Processing in Emotionally Difficult Decisions,” Chang, Hannah H and Iris W Hung (2018), “Mirror, Mirror on Journal of Experimental Psychology: Learning, Memory and the Retail Wall: Self-Focused Attention Promotes Reliance Cognition, 23 (2), 384-405. on Feelings in Consumer Decisions,” Journal of Marketing McMullen, Jeffery S and Brian J Bergman (2017), “Social Research, 55 (4), 586-99. Entrepreneurship and the Development Paradox of Prosocial Duhachek, Adam (2005), “Coping: A Multidimensional, Motivation: A Cautionary Tale,” Strategic Entrepreneurship Hierarchical Framework of Responses to Stressful Journal, 11 (3), 243-70. Consumption Episodes,” Journal of Consumer Research, 32 Nilsson, Artur, Arvid Erlandsson, and Daniel Västfjäll (2016), “The (1), 41-53. Congruency between Moral Foundations and Intentions to Ertimur, Burçak and Gokcen Coskuner-Balli (2015), “Navigating Donate, Self-Reported Donations, and Actual Donations to the Institutional Logics of Markets: Implications for Strategic Charity,” Journal of Research in Personality, 65, 22-29. Brand Management,” Journal of Marketing, 79 (2), 40-61. Pande, Amrita (2010), “” At Least I Am Not Sleeping with Fenton-Glynn, C (2019), “Surrogacy: Why the World Needs Rules Anyone”: Resisting the Stigma of Commercial Surrogacy in for «Selling» Babies,” BBC News, 26. India,” Feminist studies, 36 (2), 292-312. Fischer, Eileen, Cele C Otnes, and Linda Tuncay (2007), “Pursuing Rauscher, Emily A, Stephanie L Young, Wesley T Durham, and Parenthood: Integrating Cultural and Cognitive Perspectives Joshua B %J Health communication Barbour (2017), ““I’d on Persistent Goal Striving,” Journal of Consumer Research, Know That My Child Was out There”: Egg Donation, the 34 (4), 425-40. Institutionalized “Ideal” Family, and Health Care Decision Frey, Bruno S and Stephan Meier (2004), “Social Comparisons Making,” 32 (5), 550-59. and Pro-Social Behavior: Testing” Conditional Cooperation” Russell, Cristel Antonia and Hope Jensen Schau (2015), in a Field Experiment,” American Economic Review, 94 (5), “Consumers’ Experience of Brand Withdrawal: Unfolding 1717-22. Consumption Bereavement Theory,” in Strong Brands, Strong Gautier, Arthur, Anne-Claire Pache, and Imran Chowdhury Relationships., New York, NY, US: Routledge/Taylor & (2016), “Bound to Fail yet Successful: Understanding Francis Group, 263-81. Institutionalization Processes of Contested Practices,” in Sachdeva, Sonya, Rumen Iliev, and Douglas L Medin (2009), Academy of Management Proceedings, Vol. 2016: Academy of “Sinning Saints and Saintly Sinners: The Paradox of Moral Management Briarcliff Manor, NY 10510, 13220. Self-Regulation,” Psychological science, 20 (4), 523-28. Giving USA (2018), “Giving USA 2018: The Annual Report on Salamon, Lester M. and Chelsea L Newhouse (2019), “Nonprofit Philanthropy for the Year 2017.” Economic Data Bulletin #47,” in The 2019 Nonprofit --- (2019), “Giving USA 2019: The Annual Report on Philanthropy Employment Report, ed. Johns Hopkins University Center for for the Year 2018.” Civil Society Studies. Grant, Adam and Jane Dutton (2012), “Beneficiary or Benefactor: Sanders, Patricia (1982), “Phenomenology: A New Way of Viewing Are People More Prosocial When They Reflect on Receiving Organizational Research,” Academy of management review, 7 or Giving?,” Psychological science, 23 (9), 1033-39. (3), 353-60. Hartman, Anna E. and Erica Coslor (2019), “Earning Vs Giving: Sargeant, Adrian, John B Ford, and Douglas C West (2006), Plural Logics of Reproductive Commodification,”Journal of “Perceptual Determinants of Nonprofit Giving Behavior,” Business Research (105), 405-19. Journal of Business Research, 59 (2), 155-65. Advances in Consumer Research (Volume 48) / 1123

Scaraboto, Daiane and Eileen Fischer (2013), “Frustrated Takhar, Jennifer and H Rika Houston (2019), “Forty Years of Fatshionistas: An Institutional Theory Perspective on Assisted Reproductive Technologies (Arts): The Evolution of a Consumer Quests for Greater Choice in Mainstream Markets,” Marketplace Icon,” Consumption Markets & Culture, 1-11. Journal of Consumer Research, 39 (6), 1234-57. Thompson, Craig J (1997), “Interpreting Consumers: A Shang, Jen, Americus Reed, and Rachel Croson (2008), “Identity Hermeneutical Framework for Deriving Marketing Insights Congruency Effects on Donations,”Journal of Marketing from the Texts of Consumers’ Consumption Stories,” Journal Research, 45 (3), 351-61. of Marketing Research, 34 (4), 438-55. Sherry, John F., Jr. (1983), “Gift Giving in Anthropological Titmuss, Richard M. (1997), The Gift Relationship: From Human Perspective,” Journal of Consumer Research, 10 (2), 157-68. Blood to Social Policy, New York: The New Press. Small, Deborah A and Cynthia Cryder (2016), “Prosocial Consumer Twenge, Jean M, Roy F Baumeister, C Nathan DeWall, Natalie J Behavior,” Current Opinion in Psychology, 10, 107-11. Ciarocco, and J Michael Bartels (2007), “Social Exclusion Small, Deborah A and Uri Simonsohn (2008), “Friends of Victims: Decreases Prosocial Behavior,” Journal of personality and Personal Experience and Prosocial Behavior,” Journal of social psychology, 92 (1), 56. Consumer Research, 35 (3), 532-42. van Dijk, Mathilde, Hester Van Herk, and Remco Prins (2019), Sobande, Francesca, Laetitia Mimoun, and Lez Trujillo Torres “Choosing Your Charity: The Importance of Value Congruence (2019), “Soldiers and Superheroes Needed! Masculine in Two-Stage Donation Choices,” Journal of Business Archetypes and Constrained Bodily Commodification Research, 105, 283-92. in the Sperm Donation Market,” Marketing Theory, White, Katherine, Rishad Habib, and Darren W Dahl (2020), “A 1470593119847250. Review and Framework for Thinking About the Drivers of Strathern, Marilyn (2012), “Gifts Money Cannot Buy,” Social Prosocial Consumer Behavior,” Journal of the Association for Anthropology, 20 (4), 397-410. Consumer Research, 5 (1), 000-00. Tadajewski, Mark (2016), “Relevance, Responsibility, Critical White, Katherine and John Peloza (2009), “Self-Benefit Versus Performativity, Testimony and Positive Marketing: Other-Benefit Marketing Appeals: Their Effectiveness in Contributing to Marketing Theory, Thought and Practice,” Generating Charitable Support,” Journal of Marketing, 73 (4), Journal of marketing management, 32 (17-18), 1513-36. 109-24. Carrots over Candy: A Multi-Method Examination of the Social Factors That Affect Healthy Food Choices Chairs: Janina Steinmetz, City University of London, UK Rima Touré-Tillery, Northwestern University, USA Blake DiCosola, Oxford University, UK

Paper #1: Feeling Judged? The Presence of Outgroup Members The third and fourth paper look at consumers’ active role in Promotes Virtuous Choices shaping the social environments of their food choices. In more de- Rima Touré-Tillery, Northwestern University, USA tail, the third paper argues that consumers encourage others to match Janina Steinmetz, City University of London, UK their consumption behavior when making indulgent (vs. healthy) Blake DiCosola, Oxford University, UK choices to justify their own indulgence. Consumers feel worse about indulging when their social environment eats healthily. Therefore, Paper #2: Conspicuous Self-Control: When Status Motives consumers try to influence others to indulge so they can, too. The Lead Consumers to Signal Restraint fourth paper investigates consumers’ beliefs about being influenced Aziza Caimile Jones, Rutgers University, USA by their social environment to make healthy choices. Consumers ap- Kristina Durante, Rutgers University, USA prove of others trying to influence their food choices by nudging, but Aparna A Labroo, Northwestern University, USA only when the nudges are perceived to be effective at aiding healthier Paper #3: Have Your Cake and Make Her Eat It too: Influencing choices. Thus, consumers endorse social influence when they believe One’s Social Influence to Justify Indulgence that it helps them to eat healthily. Stephanie Lin, INSEAD, Singapore Taken together, we show that consumers take an active role in Christian Wheeler, Stanford University, USA influencing others’ impression of them based on their food choices, Sherrie Ying Ying Ying Xue, Singapore Management influence others based on their desires, and approve of being influ- University, Singapore enced if it helps their healthy eating. These behaviors are based on Paper #4: Effectiveness and Acceptance of Healthy Eating consumers’ beliefs about healthy choices being desirable but diffi- Nudges cult to make without a little help from others. These findings have Romain Cadario, Rotterdam School of Management | Erasmus important theoretical implications for research on social influences University, The Netherlands and food choices, as well as practical implications for marketer and Pierre Chandon, INSEAD, France public policy makers.

SESSION OVERVIEW Feeling Judged? The Presence of Outgroup Members An increasing share of the population in industrialized countries Promotes Virtuous Choices is obese, with severe public health consequences. Although consumers generally have the goal to eat healthily, they often struggle to EXTENDED ABSTRACT follow through (Hofmann et al. 2012). Helping consumers to make Consumers routinely choose between products of virtue (e.g., healthier food choices is thus an important avenue for marketing re- healthy food) and those of vice (e.g., indulgent food). These choices search to ultimately foster consumer well-being. take place not in a social vacuum, but often in the presence of ob- This special session examines how consumers navigate the so- servers, for example in supermarkets or cafeterias. Whereas much cial influence that other people exert on their food choices and what research has shown that consumers choose virtue over vice in the beliefs underlie their behavior. We use a multi-method approach presence of observers compared to when alone (e.g., Griskevicius, to investigate which social situations aid versus hinder consumers’ Tybur and van den Bergh 2010), what is not yet known is whether healthy choices, and how consumers think about social influence characteristics of the observers affect consumers’ choices. on their food choices. Thereby, we add to the understanding of con- Social diversity is increasing in neighborhoods, schools, and sumption not only as need satisfaction, but also as a way to seek workplaces (Putnam 2007). As a consequence, consumers regularly recognition and belonging (Berger and Rand 2008). encounter outgroup others who belong to different social groups The first and second paper highlight that consumers use healthy (e.g., ethnic, academic, or professional outgroups). In such instances, food choices to manage the impression that others have of them. Spe- consumers might fear negative judgment, because they themselves cifically, the first paper shows that consumers make healthier food judge outgroup members more negatively than ingroup members choices in the presence of outgroup (vs. ingroup) others. This effect (Brewer 1979). As consumption not only fulfills personal needs for holds across a variety of outgroups (e.g., racial, professional, or aca- specific products, but also social needs for recognition and accep- demic outgroups). Consumers anticipate that outgroups others judge tance (Berger and Heath 2007), consumers might choose products to them more negatively than ingroup others, and they use healthy food alleviate the negative judgments they expect from outgroup observ- choices to counter these anticipated negative judgments, as healthy ers. eating is believed to be a generally praiseworthy behavior (Stein and In this research, we investigate whether the presence of an in- Nemeroff 1995). The second paper tests whether consumers believe group vs. outgroup observer influences consumer choices of virtu- that they can use healthy food choices as a means to signal social sta- ous versus indulgent foods. We predict that consumers will make tus to others. Because restraint is associated with many positive life healthier choices in the presence of outgroup (vs. ingroup) others, outcomes, restraint itself might signal status. Consumers who chroni- mediated by the anticipation of being judged negatively by out- cally or situationally desire high social status make healthier food group members. In seven studies, we showed the effect of observ- choices, due to the belief that high status entails exerting restraint. ers’ group-membership on choice virtuousness using various types of group memberships (race/ethnicity, school- and work-affiliation),

Advances in Consumer Research 1124 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1125 food choices, and participant populations (university-students and ities in a state (our proxy for the presence of outgroup observers) and general-adult populations). the relative number of Google searches for pretested healthy versus Study 1 tested the causal link between an observer’s group unhealthy foods in that state (our proxy for food choices). We found membership and participants’ food choices. Three hundred thirty-six a significant interaction of food type × percentage of minorities (b = participants were approached by a black or a white research assistant 15.63 (6.85), t(49) = 2.28, p = .027). As expected, as the percentage (RA) to complete a filler survey. Afterwards, participants chose be- of minorities in that state increased, the relative number of searches tween an indulgent KitKat bar and a healthier fruit snack (our depen- for healthy food (vs. unhealthy food) also increased. dent measure). We coded as ingroup (vs. outgroup) any responses Our finding that observer group-memberships affects consum- for which the RA and the participant were of the same (vs. different) ers’ indulgent vs. healthy choices has several implications. On a race. A greater proportion of participants in the outgroup condition theoretical level, research has not yet shown that people use food chose the healthier raisins (44.16%) than in the ingroup condition choices to influence others’ perceptions. This finding is important be- (29.41%; b = .64 (.33), z = 1.94, p = .053. cause it highlights that the social environment impacts food choices Studies 2 and 3 replicated the effect of observer group-member- in more complex ways than previously thought. On a practical level, ship on food choice for another type of group membership: univer- marketers should consider the diversity of the environment in which sity affiliation. In Study 2, 180 student participants were approached they market healthy vs. unhealthy products. Thus, social diversity by an RA dressed either in their university’s t-shirt and cap (ingroup matters when marketing products and segmenting consumers. condition) or in another local university’s t-shirt and cap (outgroup condition) to complete a filler survey. Afterwards, participants chose Conspicuous Self-Control: When Status Motives Lead between indulgent M&Ms and healthier raisins (our dependent mea- Consumers to Signal Restraint sure). A greater proportion of participants in the outgroup condition chose the healthier raisins (31.25%) than in the ingroup condition EXTENDED ABSTRACT (12.16%; b = 1.19 (.45), z = 2.66, p = .008. In Study 3, 154 student Extant research suggests that status signaling motivations lead participants completed an online study they believed originated from consumers to behave indulgently. They are more willing to purchase their own university (ingroup) or a foreign university (outgroup). conspicuous luxury products over less expensive options to convey For our dependent measure, participants chose between a health- wealth and signal high status (Rucker and Galinsky 2008). We posit minded vs. an indulgent restaurant gift-card. A greater proportion of that when consumers are unable to convey wealth, they respond to participants in the outgroup condition (80.77%) chose the healthier status signaling desires by attempting to convey self-control because restaurant than in the ingroup condition (65.79%; b = .78 (.38), z = of the need to appear goal-oriented. Goal-oriented individuals are 2.08, p = .037. more likely to succeed in school and in their careers (Locke and Study 4 tested the role of expected interpersonal-judgment in Latham 2006), both of which can lead to high status. Self-control the effect of observer group-membership on food choice. Using a may indicate that a consumer is goal-oriented because it involves causal-chain mediation, two separate studies examined this proposed sacrificing an immediate desire for a long-term result. Thus, status psychological process (expected interpersonal-judgment) as both an signaling consumers may attempt to conspicuously engage in self- effect of the independent variable (observer group-membership) control as a way to signal that they are highly goal-oriented, and and as a predictor of the dependent variable (food choice). In Study ultimately high status. 4a, 206 participants imagined making an indulgent food choice in Six studies examined whether consumers display self-control the presence of an ingroup (vs. outgroup) observer, and indicated to signal high status because such behaviors signal that they are their expected interpersonal-judgment from this person. Participants goal-oriented. In studies 1-4, we held the product prices constant expected the outgroup member would judge them more negatively in all the choices that participants made. We conducted Study 1 in a (M = 4.87, SD = 2.03) than the ingroup member (M = 5.52, SD = low-income area of a large metropolitan city. Two experimenters ap- 1.87, t(204) = 2.39, p = .018). In Study 4b, 192 participants imag- proached individuals in public and gave each participant one of two ined choosing between indulgent and healthy food options in the flyers, one that stimulated a status-signaling motivation or one that presence of an observer (with unspecified group membership), and did not. After reading the flyer, participants chose between receiv- indicated their expected interpersonal-judgment from this person, as ing a self-control signaling product (banana) or a different product well as their food choice. The more harshly participants expected (M&Ms). Status signaling motives (vs. control) led participants to the observer to judge them, the more likely they were to choose the select the banana (i.e., the self-control product; p = .04). healthier option, β = .235, SE = 0.113, t(197) = 2.076, p = .039. Study 2 replicated the effect in a controlled environment. We Study 5 tested the moderating roles of internal-attribution ten- randomly assigned participants to a status signaling or control con- dencies. Three hundred twenty-three student participants completed dition. After the status manipulation, they imagined that they were an online survey they believed originated from their own university with two other people while at home and had the choice between (ingroup) or another local university (outgroup). Participants indi- sharing a self-control related video (self-control signaling product) cated their preference for healthy and indulgent subscription food- or an enjoyable video. Status motives (vs. control) led consumers to clubs, and their general tendency to make internal attributions for select the self-control video (p < .01). their own actions using Touré-Tillery and Light’s (2018) self-diag- If status signaling motives lead consumers to conspicuously nosticity scale. We found a significant interaction of observer group- choose products that signal self-control, then consumers should membership × internal-attribution tendency × food type (F(1, 298) attempt to make such products observable to others (Sundie et al. = 4.76, p = .030). As we expected, internal-attribution tendencies 2011). Study 3 examined this by employing a 3 (status signaling moderated the effect of observer group-membership on food choice motivation: high, low, vs. control) × 2 (self-control signaling prod- such that it replicated at lower—but not at higher—levels of internal- uct: yes vs. no) between-subjects design in which preference for a attribution tendencies. transparent versus an opaque container was our dependent variable In Study 6, we used (US) state-level search data from Google of interest. We primed participants with either a high, low, or con- Trends to examine the relationship between the percentage of minor- trol (no prime) desire to signal status and randomly assigned par- 1126 / Carrots over Candy: A Multi-Method Examination of the Social Factors That Affect Healthy Food Choices ticipants to a healthy meal (i.e., self-control signaling product) or an to indulge more when others are indulging (Burger et al. 2010; Her- unhealthy meal condition. Participants imagined that they were with man, Roth, and Polivy 2003; Prentice and Miller 1993), and feel less a group of people and were purchasing lunch from a local restaurant guilty if another person engages in small indulges with them than if that served healthy (i.e., self-control signaling product) or unhealthy they indulge and the other person abstains (Lowe and Haws 2014). meals. They picked between 2 containers for their meals – one that If others’ bad behavior makes people feel better about their own was transparent or another that was opaque. The analysis revealed transgressions, might people manipulate those norms directly? We only a significant interaction (p = .04). Status signaling participants hypothesize that people encourage others to match their consump- preferred the transparent container more for a healthy, self-control tion behavior when they indulge (vs. not indulge) in order to justify signaling meal compared to an unhealthy meal (p = .009), an effect their indulgence. not found in the other two conditions. In study 1a, participants (100 MTurk) rated how fun and indul- Study 4 sought to examine whether a desire to signal that one gent 40 activities were (e.g. eating a slice of cake, buying grocer- is goal-oriented mediated the effect. We randomly assigned partici- ies) and how justified they would feel doing those activities alone pants to a status signaling or control condition. After completing and with a friend. People felt more justified doing indulgent activi- the manipulation, they imagined shopping with several people and ties with friends than alone, and this difference increased with rat- indicated their preference between receiving a self-control applica- ings of indulgence t(3996)=16.45, p<.001. In study 1b, participants tion or a gaming application. They also reported their need to signal (100 MTurk) rated how fun and indulgent the same activities were, that they were goal-oriented. Status signaling participants were more and how likely they were to encourage their friends to do those ac- likely to select the self-control application (p < .05). Desire to signal tivities with them. Indulgence predicted encouragement of friends that they were goal-oriented mediated the effect. to match their behavior, controlling for how fun the activity was, While the previous studies held the product prices constant, t(3980)=9.33, p<.001. Study 5 aimed to show that the effect weakens when consumers Study 2 tested whether people are more likely to encourage oth- choose between products that vary largely in price, and therefore, ers to match their indulgent (vs. non-indulgent) behavior to justify allow the consumer to signal wealth instead of self-control. We ran- their behavior. Participants (295 MTurk) imagined that they ordered domly assigned participants to a status signaling or control condi- a beer or lemonade with lunch, and that their friend asked what they tion. After completing the manipulation, they imagined shopping should order. Participants indicated their recommendation and how with several people and indicated their preference between purchas- justified they would feel if their friend ordered beer and how justi- ing a tablet with preloaded self-control applications or a tablet with fied they would feel if their friend ordered lemonade. Participants preloaded games. We told half the participants that both tablets cost encouraged their friend to match their consumption more when they $150. We told the rest of the participants that the gaming tablet cost ordered beer than lemonade, t(192)=–6.99, p<.001. This was not $350 and the self-control tablet cost $150. The results revealed a sig- driven by participants always encouraging beer; those in the beer nificant interaction (p < .001) in which status signaling participants condition encourage beer more than those in the lemonade condi- were more likely to select the self-control tablet, but only when it tion, t(192)=4.39, p<.001. (Across studies, matching and indulgence was equal price with the games tablet. are examined separately when relevant; we expect both more match- Study 6 aimed to demonstrate that even in situations where ing and more indulgence encouragement in the indulgence vs. non- status signaling consumers are tempted to purchase more expensive indulgence conditions). Mediation indicated that people encouraged products over less expensive options, they can be motivated to pur- matching more for beer than lemonade because they anticipated chase the less expensive option if they remember that saving behav- matching to make them feel more justified when drinking beer (vs. ior is a signal of self-control. We randomly assigned participants to a lemonade), CI95[-.21, -.02]. status signaling or control condition. After completing the manipula- Study 3a tested the underlying reason for this anticipated jus- tion, they imagined shopping with several people and indicated their tification. Female participants (N=200 MTurk) imagined ordering preference between purchasing one of two cell phones. One phone either cake or fruit salad, and indicated what they would encourage was $200 more than the other. We reminded half the participants that their friend to order. They responded to mediator items representing saving behavior signals self-control. The results revealed a signifi- three proposed mechanisms: (1) having a partner-in-crime (e.g., “If cant interaction (p < .05) in which status signaling participants were my friend followed my recommendation…I would feel more justi- more likely to select the less expensive phone, but only when we fied because at least we would both be engaging in equally bad be- reminded them that saving behavior signals self-control. havior,” (2) lowering perceived consequences (e.g., “…I would feel Our findings suggest that status-signaling motives can enhance like my choice was less ‘bad’ or ‘wrong’,”), and (3) changing the preferences for products associated with self-control because con- situational norm (e.g., “…it would feel like I made my choice for a sumers equate high status with a strong goal orientation. From a prac- celebration”). Those who chose cake recommended that their friend tical perspective, our findings suggest that marketers can encourage match their behavior more than those who chose fruit, t(192)=-5.33, consumers who desire to signal status to engage in self-control and p<.001 (and recommended cake more, t(192)=–4.39, p<.001). Only purchase self-control related products. the partner-in-crime measure mediated the effect, CI95[.022, .530]. Study 3b manipulated whether ordering cake was justified or not: Have your cake and make her eat it too: Influencing cake either exceeded (unjustified) or was within (justified) their alone’s social influence to justify indulgence lotted “diet points.” Participants (N=200 female MTurk) who had unjustifiably (vs. justifiably) ordered cake were more likely to en- EXTENDED ABSTRACT courage that their friend match their behavior by ordering cake, Because indulging violates personal standards of behavior t(197)=6.01, p<.001. Both partner-in-crime (CI95[.001, .244]) and (Dahl, Honea, & Manchanda, 2003), people have a variety of ways situational norms (CI95[.059, .417]) mediated this effect. to justify their actions (Khan and Dhar 2006; Kivetz and Zheng Study 4 explored the role of gender in the lab. Same-gendered 2006; Xu and Schwarz 2009). One way to justify one’s transgres- friend dyads were directed to two separate rooms, each containing sions is by comparing one’s behavior to others’. People are likely a face-down playing card (one red, one black, randomly assigned). Advances in Consumer Research (Volume 48) / 1127

The experimenter explained that if the card was red, they would eat ers’ welfare “as judged by consumers themselves » (Sunstein 2018), a Snickers bar (indulgence), and if it was black, they would eat sal- this assumption is rarely tested. Certainly, it seems cavalier to as- tine crackers (non-indulgence). Participants privately checked their sume that consumers will welcome any nudge, especially those that card. The experimenter then said she “forgot” to choose a card for restrain their perceived sense of agency. the other participant, and asked participants to choose a card for Therefore, the goal of this study is to help decision makers their friend from a box. The box contained eight face-down cards select the best healthy eating nudge by measuring consumers’ per- and two face-up cards (one red, one black). Participants chose a ceptions—including their acceptance—of the seven types of nudges card matching their own more when they were assigned to indulge identified in the literature. In a second step, these perceptions are (B=.38, SE=.15, p=.012), with no gender interaction. When examin- compared to the actual effectiveness of the same nudges reported in ing choice of indulgence, a pre-registered gender by snack interac- Cadario and Chandon (2019). tion emerged, B=.29, SE=.15, p=.052. Women were marginally more likely to choose indulgence when they were assigned to indulge than Methods not, B=.32, SE=.19, p=.085; men did not show this difference, B=– We surveyed 118 American citizens about their acceptance of .25, SE=.22, p=.27. different types of healthy eating nudges. To investigate the drivers Study 5 tested whether people indulge more when they can first of their acceptance, we asked them about their perceptions of the ef- facilitate indulgence in others. Participants (150 female students) fectiveness of each nudge and of the beneficiaries of the nudge (good expected to be assigned to snack or not while watching a video. In for health, good for business, or both). Participants were presented, the “self-only” condition the participant served herself mini peanut in random order, the scenarios for seven healthy eating nudges. butter cups from a large bowl. In the “both” condition, the participant We asked participants to answer two questions: “Do you ap- was also asked to serve another participant, ostensibly in another prove or disapprove of the following policy?” (Approve/disapprove) room. As expected, those in the “both” condition served themselves and “Do you think that this policy will make people eat better?” (Yes more than those in the “self-only” condition, t(103)=3.00, p=.003. it will/No it will not). Then, we compared these perceptions to the To ensure that participants in the “both” condition were not being average effect sizes about these seven nudges reported in Cadario generous and subsequently anchoring their own serving on a larger and Chandon (2019). amount, there was also an “other-only” condition in which participants served only another participant (ostensibly because the stimuli Results were in the wrong room). They unexpectedly served the other partic- The actual effectiveness of these seven nudges was inversely ipant more in the “other-only” condition than they did in the “both” related to their mean approval rating (r=-.57) as well as to their per- condition, t(100)=3.45, p<.001. This suggests that people were not ceived effectiveness (r=-.49). For example, only 43% of respondents simply anchoring on a generous amount in the “both” condition, as approved the most effective intervention – portion and package size they were even more generous in the “other only” condition. reductions. Thus, consumers play a role in shaping their social contexts To examine the drivers of nudge approval, we regressed ap- when they wish to justify their own indulgence. This contributes to proval on the effectiveness of the nudge (the standardized mean dif- research on social influence, dyadic consumption, and justification ference reported in the meta-analysis), its perceived effectiveness, literatures, while having marketing implications (e.g., when to offer two binary variables capturing the effects of the perceived benefi- “bring a friend” deals vs. BOGO deals). ciary of the nudge, and the individual characteristics. First, we found that approval was positively associated with Effectiveness and Acceptance of Healthy Eating Nudges perceived effectiveness, as one would expect, but negatively associated with actual effectiveness. Second, we found that interventions EXTENDED ABSTRACT perceived as a “win-win” for business and health had higher approv- A growing number of governments, food producers and retail- al than interventions perceived as benefiting either health or busi- ers, are considering implementing nudges promoting healthier eat- ness, and that there were no differences in approval between each of ing. A nudge can be defined as “any aspect of the choice architecture these respectively that alters people’s behavior in a predictable way (1) without forbid- ding any options, or (2) significantly changing their economic incen- Discussion tives. Putting fruit at eye level counts as a nudge; banning junk food The average approval rate of the seven healthy eating nudges does not” (Thaler and Sunstein 2008). was only 56%, a lower rate than we would expect based on prior Not all healthy eating nudges are equal. A recent meta-analysis results suggesting that the United States is a “pro-nudge” country (Cadario and Chandon 2019) found that their effectiveness increases (Sunstein et al. 2017). More importantly, there were large variations as their focus shifts from influencing what people know (cognitive in approval across nudges. Consistent with prior research, there was nudges), how they feel (affective nudges), or what they do (behavior higher approval for deliberative (“system 2”) than for automatic nudges). This study further distinguishes between two or three sub- (“system 1”) nudges (Felsen et al. 2013; Jung and Mellers 2016). types of nudges for each category, leading to seven types of nudges. Nudges with the highest approval (85%) were those that simply in- Cognitive nudges include “descriptive nutritional labeling,” “evalu- form consumers, such as descriptive nutrition labels. The addition of ative nutritional labeling,” and “visibility enhancements”. Affective interpretive symbols (such as color-coding) to shift from description nudges consist of “healthy eating calls” and “hedonic enhance- to prescription reduced approval levels from 85% to 76%. ments”. Behavioral nudges include “convenience enhancements” Our results extend this research by showing that the most effec- and “size enhancements.” tive healthy eating nudges receive significantly lower approval than Selecting the optimal healthy eating nudge is not just a ques- the rest. The average approval rate of the two most effective nudges, tion of selecting the most effective one. Governments and companies convenience enhancements and portion and package size reductions, must also consider whether these nudges are likely to be accepted by was 43%, indicating that when asked to choose between healthy eat- consumers. Although all nudges are supposed to improve consuming nudges, there is a tradeoff between approval and effectiveness. 1128 / Carrots over Candy: A Multi-Method Examination of the Social Factors That Affect Healthy Food Choices

What, then, drives the approval of healthy eating nudges? Our Kivetz, Ran and Yuhuang Zheng (2006), “Determinants of analyses rule out that people simply reject nudges that they deem Justification and Self-Control,”Journal of Experimental to be the most effective – for example because they do not want Psychology: General, 135(4), 572–87. to be influenced. In fact, approval ratings increased with perceived Locke, Edwin A., and Gary P. Latham (2006), “New Directions in effectiveness. This suggests that the problem is not that people dis- Goal-Setting Theory,” Current Directions in Psychological like being nudged but that they are poor judges of which nudges are Science, 15 (5), 265-68. effective. Lowe, Michael L. and Kelly L. Haws (2014), “(Im)Moral Support: The Social Outcomes of Parallel Self-Control Decisions,” REFERENCES Journal of Consumer Research, 41(2), 489–505. Berger, Jonah, and Chip Heath (2007), “Where consumers diverge Prentice, Deborah A. and Dale T. Miller (1993), “Pluralistic from others: Identity signaling and product domains,” Journal Ignorance and Alcohol Use on Campus: Some Consequences of Consumer Research, 34(2), 121-134. of Misperceiving the Social Norm,” Journal of Personality Berger, Jonah, and Lindsay Rand. 2008. “Shifting Signals to Help and Social Psychology, 64(2), 243–56. Health: Using Identity Signaling to Reduce Risky Health Putnam, Robert D. (2007), “E pluribus unum: Diversity and Behaviors.” Journal of Consumer Research 35 (3): 509–18. community in the twenty‐first century the 2006 Johan Skytte Brewer, Marilynn B. (1979), “In-group bias in the minimal Prize Lecture,” Scandinavian Political Studies, 30(2), 137- intergroup situation: A cognitive-motivational analysis,” 174. Psychological Bulletin, 86(2), 307-324. Rucker, Derek D. and Adam D. Galinsky (2008), “Desire to Burger, Jerry M., Heather Bell, Kristen Harvey, Jessica Johnson, Acquire: Powerlessness and Compensatory Consumption,” Claire Stewart, Kelly Dorian, and Marni Swedroe (2010), Journal of Consumer Research, 35 (2), 257-67. “Nutritious or Delicious? The Effect of Descriptive Norm Stein, Richard I, and Carol J Nemeroff (1995), “Moral Overtones Information on Food Choice,” Journal of Social and Clinical of Food: Judgments of Others Based on What They Eat,” Psychology, 29(2), 228–42. Personality and Social Psychology Bulletin, 21(5), 480–90. Cadario, Romain and Pierre Chandon (2019), “Which Healthy Sundie, Jill M., Douglas T. Kenrick, Vladas Griskevicius, Joshua Eating Nudges Work Best? A Meta-Analysis of Field M. Tybur, Kathleen D. Vohs, and Daniel J. Beal (2011), Experiments,” Marketing Science, in press. “Peacocks, Porsches, and Thorstein Veblen: Conspicuous Felsen, Gidon, Noah Castelo, and Peter B Reiner (2013), Consumption as a Sexual Signaling System,” Journal of “Decisional enhancement and autonomy: public attitudes Personality and Social Psychology, 100(4), 664-80. towards overt and covert nudges,” Judgment and Decision Sunstein, Cass R, Lucia A Reisch, and Julius Rauber (2017), Making, 8 (3), 202-13. “Behavioral insights all over the world? Public attitudes Griskevicius, Vladas, Joshua M. Tybur, and Bram Van den Bergh toward nudging in a multi-country study,” in Working Paper. (2010), “Going green to be seen: status, reputation, and Sunstein, Cass R. (2018), ““Better off, as judged by themselves”: conspicuous conservation,” Journal of Personality and Social a comment on evaluating nudges,” International Review of Psychology, 98(3), 392-404. Economics, 65(1), 1-8. Herman, C. Peter, Deborah A. Roth, and Janet Polivy (2003), Thaler, Richard H. and Cass R. Sunstein (2008), Nudge: Improving “Effects of the Presence of Others on Food Intake: A Decisions about Health, Wealth, and Happiness. New York: Normative Interpretation,” Psychological Bulletin, 129(6), Penguin Books. 873–86. Touré-Tillery, Maferima, and Alysson E. Light (2018), “No self to Hofmann, Wilhelm, et al. (2012), “Everyday temptations: an spare: How the cognitive structure of the self influences moral experience sampling study of desire, conflict, and self- behavior,” Organizational Behavior and Human Decision control,” Journal of Personality and Social Psychology, Processes, 147, 48-64. 102(6), 1318-1335. Xu, Jing and Norbert Schwarz (2009), “Do We Really Need a Jung, Janice Y and Barbara A Mellers (2016), “American attitudes Reason to Indulge?,” Journal of Marketing Research, 46(1), toward nudges,” Judgment and Decision making, 11 (1), 62- 25–36. 74. Khan, Uzma and Ravi Dhar (2006), “Licensing Effect in Consumer Choice,” Journal of Marketing Research, 43(2), 259–66. Obstacles and Opportunities to Consumer Engagement in Medical Decision Making Chair: Mohamed A. Hussein, Stanford University, USA

Paper #1: Shopping in the Dark – Barriers to Determining the and Menon further investigate how self-diagnosis is influenced by Price of American Healthcare contextual cues and find that ambiguous symptoms are perceived to Peter A. Ubel, Duke University, USA be less diagnostic and weighted less. Lastly, Lieberman et al . take the perspective of policy mak- Paper #2: How Search Strategy Shapes Self-Diagnosis ers and examine ways to encourage consumers to complete medi- Mohamed A. Hussein, Stanford University, USA cal screening. In two large field experiments, the authors leverage Szu-chi Huang, Stanford University, USA financial incentives and deadlines to encourage preventive screening. Paper #3: Web Wizard or Internet Addict? How to Help Together these papers raise (and provide answers to) the follow- Consumers Assess Risk ing questions: Priya Raghubir, New York University Stern Business School, • What are some of the hurdles to consumers’ involvement USA in medical decision making (e.g., inability to estimate out- Geeta Menon, New York University Stern Business School, of-pocket expenses and susceptibility to contextual cues)? USA • What are the psychological drivers that lead consumers to I-Ling Ling, Kaohsiung Medical University, Taiwan seek diagnosis (e.g., through self-diagnosing online and Paper #4: Testing the Role of Motivation and Procrastination in through taking a screening test), and what are the opportu- Colorectal Cancer Screening nities to encourage more accurate diagnoses? Alicea Lieberman, Rady School of Management - University This session provides timely insights into medical decision of California, San Diego, USA making, a critical topic given consumers’ increasing engagement Ayelet Gneezy, Rady School of Management - University of with medical decisions. This session echoes some of the confer- California, San Diego, USA ence’s themes, such as the use of multi-method approaches (e.g., fo- Emily Berry, Moncrief Cancer Institute - University of Texas cus groups, interviews, lab experiments, and field experiments) and Southwestern, USA understanding consumers’ interactions with technology (e.g., search Keith Argenbright, Moncrief Cancer Institute - University of engines) in the context of medical decision making. We believe that Texas Southwestern, USA this proposed session should be of interest to a broad audience of Samir Gupta, Department of Internal Medicine - University of scholars and practitioners working on healthcare and medical deci- California, San Diego, USA sions, technology, extrinsic and intrinsic motivation, and context effects. SESSION OVERVIEW Because of the rapid proliferation of medical information, con- Shopping in the Dark – Barriers to Determining the Price sumers are becoming more engaged in decisions about their health. of American Healthcare For example, consumers increasingly track their health-related behaviors (e.g., number of steps, calorie intake, sleep quality) through EXTENDED ABSTRACT wearable devices (e.g., Fitbits); go online to diagnose themselves and The United States is rapidly shifting toward a high out-of- learn about possible treatment options; and leave online reviews re- pocket healthcare system. People are increasingly enrolling in health flecting on their experiences with healthcare providers. insurance plans that expose them to costs when they receive health- Importantly, medical decisions often involve multiple stake- care services, either in the form of high deductibles (the amount they holders: physicians, consumers/patients, insurance providers, and pay out-of-pocket before insurance kicks in), co-pays (flat fees for policy makers. This session aims to explore these different perspec- medical services), or coinsurance (exposure to a certain percent of tives, and highlight four specific ways in which social, cognitive, and healthcare expenses). motivational variables could hamper or facilitate consumers’ medical In previous research, my colleagues and I have describe the decision making. These papers incorporate a diverse array of empiri- kind of challenges patients and physicians face trying to determine cal methods including focus groups, semi-structured interviews, on- out-of-pocket expenses in time to inform medical decisions – e.g. line studies with mock search engines, surveys with 1200 physicians, figuring out the relative cost of alternative medications for a patients’ and large-scale field experiments. illness to help determine which medicine the physician should pre- The first paper examines the obstacle of estimating out-of- scribe (Hunter et al. 2017; Ubel 2019; Ubel 2016). pocket costs. Consumers often cannot determine how much they I will discuss two studies which are currently underway, both of need to pay for different treatment options, which complicates their which will be completed by the time of the conference. Both studies decisions. Ubel tests physicians’ ability to estimate patients’ out-of- shed new light on this important topic. pocket costs and finds that about one-third of physicians incorrectly The first study involves a series of focus groups with physi- estimate these costs. In another study, Ubel uses focus groups with cians from four medical specialties – rheumatology, neurology, on- physicians to solicit solutions to overcome this challenge. cology, and primary care. We have completed five focus groups and The second and third papers take the perspective of consumers will have completed five more by the time of the conference. In the and examine factors that can bias their self-diagnosis, i.e., consum- focus groups, we present physicians with examples of patients strug- ers’ assessment of the extent to which they have a medical condi- gling to figure out the cost of their care. We draw these examples tion. Hussein and Huang examine how consumers’ search strategy from direct observation of clinic appointments and from interviews (symptom-focused vs. disease-focused) affect the process and out- of participating patients. We use the examples to prompt physicians come of self-diagnosis and find that symptom (vs. disease)-focused to discuss their own challenges trying to ascertain patients’ out-of- search leads to higher accuracy of self-diagnosis. Raghubir, Ling,

Advances in Consumer Research 1129 Volume 48, ©2020 1130 / Obstacles and Opportunities to Consumer Engagement in Medical Decision Making pocket expenses. More importantly, we then probe physicians to sug- We investigated how consumers’ search strategy affects self-di- gest ways to overcome these challenges. agnosis. First, we documented search strategies consumers naturally The solutions we have explored so far in the focus groups have use when self-diagnosing online (Study 1). We found that consum- been wide-ranging. They include: incorporating costs into electronic ers used two main types of search strategies: disease-focused and health records; employing financial navigators to help patients de- symptom-focused. Second, we investigated the consequences of termine the costs of potential care; developing tip sheets for patients these two search strategies on self-diagnosis (Study 2). We found with specific health conditions to suggest ways of reducing their that participants randomly assigned to use a symptom-focused (vs. out-of-pocket expenses – e.g. “co-pay assistance programs available disease-focused) strategy reported more accurate self-diagnoses. for patients with lupus;” and screening patients in waiting rooms, to We then turned to examining the psychological mechanism driving identify those struggling with out-of-pocket expenditures. the increased accuracy. We found that participants in the symptom- The focus groups have also yielded rich insights into barriers focused (vs. disease-focused) condition generated more abstract (vs. for implementing potential solutions, with the main challenge being concrete) hypotheses during their information search (Study 3), and misalignment of financial incentives: “You can hire financial coun- that those abstract (vs. concrete) hypotheses prevented consumers selors to make the physicians’ jobs easier, and to improve patients’ from being biased by disease-specific information and allowed them lives, but if you reduce income for the Anonymous Hospital Organi- to focus more on disease-agnostic information (Study 4), which in- zation, administrators won’t hire them.” creased their accuracy. In the second study, we survey a representative sample of 1200 In Study 1, participants (N = 120) were asked to imagine they US physicians and assess how well they can ascertain patients’ out- went on a nine-day Safari in rural Africa and that they suddenly fell of-pocket expenditures when they have sufficient information to do ill. Participants read that they experienced three symptoms (e.g., so. We present physicians with a hypothetical patient: back pain) and that travelers who visited the safari park sometimes Now, imagine the following: contracted one of four diseases (e.g., Malaria). Next, participants • You are about to prescribe the oral biologic DMARD to- were asked to use Google to better understand their health situation. facitinib to your patient, Ms. Gray. We used a software called Pipe, which allowed us to record par- • One month’s supply of tofacitinib costs $1,000. She asks ticipants’ screen as they conducted a Google search. We analyzed how much she’ll have to pay out of pocket. the search terms participants used on Google and found that 47.06% • Luckily, you have access to her private insurance informa- of participants used symptom-focused search terms, such as “fever tion. Tofacitinib is a tier 4 drug, and her insurance runs chills and backpain in Africa,” while 45.38% of participants used from January to December. disease-focused search terms, such as “Trachoma in rural Africa.” In Study 2, we investigated whether these two search strategies We ask them four questions about out-of-pocket expenses. For influenced the accuracy of self-diagnosis. We used the same safari example, in one question, we tell the physicians it is January 3, and paradigm described in Study 1, and randomly assigned participants ask them to advise the patient on how much she will have to pay (N = 400) to either a symptom-focused or a disease-focused search out-of-pocket for the drug this month. The correct answer is $1000, strategy. Participants in the symptom-focused (vs. disease-focused) because the patient has not met her $2000 deductible yet, so the full condition were asked to choose the symptoms (vs. diseases) they price of the drug is her responsibility to bear. In another question, we would like to focus their Google search on. After screen-recording asked them to imagine it is April, and the patient is still taking the their Google search using the same screen-recording client (Pipe), drug. We asked them to estimate how much the drug will cost her participants reported the probability that they had each of eight dis- that month. In that case, the patient has met her deductible, but is still eases: two diseases closely matched the symptoms in the scenario, responsible for half the price of the drug, meaning she will pay $500 two diseases were mentioned in the scenario but did not match the that month for the medication. symptoms, and four diseases were neither mentioned in the scenario We have conducted the first of two waves of data collection, nor matched the symptoms. We pre-registered all analyses. We found with a response rate so far of 40%. Our preliminary results reveal that participants in the symptom-focused (vs. disease-focused) con- widespread misunderstanding of out-of-pocket expenditures. For dition were more accurate in their self-diagnosis: they reported both example, for the two questions described above, 35%, and 41% of higher probabilities for having diseases that matched their symptoms responding physicians so far answered those questions incorrectly. (b = 7.58, z = 3.31, p = .001) and lower probabilities for diseases that These two studies provide a window into a dysfunctional did not match their symptoms (b = 4.84, z = -2.70 p = .007). market – in which people face high out-of-pocket expenses for the In Studies 3 and 4, we examined how search strategy affect- medical care but cannot determine their costs in time to inform their ed consumers’ hypothesis generation and testing (Klayman & Ha, healthcare choices. Moreover, when they turn to their physicians, 1987). To capture the key hypothesis motivating their search, par- they often cannot receive good information about their costs. ticipants in Study 3 (N = 200) engaged in a thought listing exercise How Search Strategy Shapes Self-Diagnosis (Posavac et al., 2010), in which they shared the questions they hoped their Google search would answer. These questions were coded for EXTENDED ABSTRACT the total number of thoughts, whether a concrete hypothesis was One of the key reasons consumers search for health information mentioned (e.g., “do I have Malaria?”), and whether an abstract hy- online is to self-diagnose (Fox & Duggan 2013). Self-diagnosis re- pothesis was mentioned (e.g., “which diseases exhibit symptoms of fers to consumers’ assessment of the extent to which they developed chills in general?”). We found that, despite reporting the same total a medical condition. In 2013, more than 35 percent of Americans number of thoughts (b = .27, t(198) = 1.01, p = .31), participants went online to self-diagnose (Fox & Duggan 2013), and in 2017 this in the symptom-focused (vs. disease-focused) condition were less percentage increased to 57 percent (Consumer Health Online 2017). likely to report a concrete hypothesis (logistic b = .28, z = -4.16, p Importantly, 77 percent of online self-diagnosis starts on a search < .001) and more likely to report an abstract hypothesis (logistic b = engine, like Google (Fox & Duggan 2013). 4.62, z = 4.84, p < .001). Together, these results suggest that search Advances in Consumer Research (Volume 48) / 1131 strategy influences the hypotheses consumers generate and bring to symptoms are elicited prior to risk estimates; and (c) including a the search engine. “none-of-the-above” option on the symptom list affects judgments In Study 4, we further tested whether the difference in hypoth- of how common they are. eses documented in Study 3 changed the type of information partici- Experiment 1 (n=75) showed that risk estimates for self pants subsequently chose. We predicted that symptom-focused par- (M=2.82) were significantly lower than the average undergraduate ticipants would pay more attention to information that can address (M=4.01, p<.001), but no different than one’s best friend (M=2.84; their abstract hypotheses (e.g., “Common causes of fever”) while F(2, 146)=31.80, p<.001). A regression of self-risk estimates on diseases-focused participants would pay more attention to informa- internet usage was significant (F(1, 69)=11.39, p<.001, Ra2=.129; tion that can better address their concrete hypotheses (e.g., “Malaria B=.046, t=3.38, p<.001). – Symptoms and Causes”). We pre-registered these predictions and Experiment 2 (n=109) estimated the effect of whether the IAD related analyses. To test this prediction, we used a yoked design in symptoms checklist had a “None-of-the-Above” option. When the which the hypotheses organically generated in Study 3 were present- option was present, fewer symptoms were identified (M=2.62 vs. ed to a new set of participants (N = 400), such that each participant in 3.52; F(1, 107)=8.14, p<.005), and risk estimates were directionally Study 4 read the hypothesis generated by a randomly selected partic- lower (M=56.96 vs. M=61.94; F(1, 107)=1.90, p=0.171), while the ipant from Study 3. Participants were then asked to select up to three self-positivity bias was robust reflecting lower self- versus other- search results that would best answer the central questions raised by judgments (M=55.83 vs. M=63.07; F(1, 107)=14.74, p<.001). A the hypothesis they were assigned. Participants chose between 10 moderated mediation (process 5) model showed that the effect of search results, five of which were disease-specific (i.e., the title con- self-reports of internet use affected risk judgments indirectly via retained the names of specific diseases, such as Malaria) and the other ports of behaviors associated with addiction, while the direct effect five were disease-agnostic (i.e., the title made no reference to a spe- was moderated by how internet use is elicited: in no “none-of-the- cific disease, such as “Types of Fevers in Travelers Returning from above” condition, there was no effect, but in the “none-of-the-above” Abroad”). We found that symptom-focused (vs. disease-focused) present condition, there was a direct effect. participants chose fewer disease-specific information sources (b = Experiment 3 (n=146 US undergrads) examines the effects of .31, t(180) = 3.13, p = .002) and more disease-agnostic information the contextual cue of the range of response alternatives (Schwarz sources (b = .27, t(180) = 2.74, p = .007). Together, these results sug- et al., 1985). Participants indicated internet use on low-, or high- gest that search strategy influenced both the hypothesis generation frequency response alternatives, such that the percentage of partici- and subsequent evidence selection needed for self-diagnosis. pants checking the last category in the “low” condition is compa- In summary, we investigated how online search strategy affects rable to the last four categories in the “high” condition. Those in the the process and outcome of self-diagnosis. Importantly, employing “low” condition should infer that they use the internet more than the a symptom-focused strategy resulted in greater self-diagnosis accu- average person, identify more symptoms, and estimate higher risk racy. This enhanced accuracy appears to be driven by consumers’ (Schwarz et al. 1985). Participants were then assigned to one of two adoption of more abstract hypotheses which attracts them away from order conditions (symptoms-before-risk, risk-before-symptoms) pri- disease-specific information and towards disease-agnostic informa- or to indicating the likelihood of having IAD (0-100) for themselves tion. and others, resulting in a 2 (response-alternatives) x 2 (order) x 4 (target: self, average person taking the study, average undergraduate Web Wizard or Internet Addict? How to Help student, and average person) mixed design. Consumers Assess Risk The range-of-response-alternatives affected the participants’ report of using the internet for >5 hours (Low=48.7%, High=83.8%, EXTENDED ABSTRACT (1)=19.70, p<.001). A mixed ANOVA on risk estimates revealed a A key issue with health risk assessments is one of self-diag- main effect of target (F(3, 426)=30.39, p<.001, =.176), and a target nosis. The surge in DTC advertising for conditions ranging from x order interaction (F(3, 426)=3.70, p=.012, =.025). The interaction depression to diabetes, typically involve a range of symptoms that showed that perceptions of self-risk (but not risk of the three other consumers are asked to identify in an attempt to self-diagnose. But targets) were lower only when symptoms were elicited prior to risk many of these are ambiguous. This paper examines how consum- estimates (M=44.61 vs. 54.01). ers assess the risk of internet addiction, which has a number of am- Overall, self-risk perceptions were significantly lower biguous symptoms. The prevalence and growth of internet addiction (M=49.31) than estimated risk for the average person taking the disorder (IAD) could pose a risk to consumer welfare, as well as study (M=55.745, t145=4.01, p<.001), the undergraduate (M=61.76, the industries that depend on the internet, from personal computers t145=7.31, p<.001), but no different from the average person (IBM, Apple), software (Microsoft, Adobe), and hardware providers (M=49.155). When the symptoms were elicited first, risk esti- (HP), to retailers (Amazon, Land’s End), search engines (Google) mates were based on both unambiguous and ambiguous symptoms and portals (Yahoo). (Bs=11.78 and 4.74, ts=3.39 and 2.33, ps<.01 respectively). When Five studies replicate and extend the robust self-positivity in risk estimates were elicited first, risk estimates were based on report- judgments of IAD. Studies 1 and 2 examine self-positivity effects used internet usage and ambiguous symptoms (Bs=20.68 and 12.10, ing open-ended estimates of number of hours of internet use. Studies ts=2.36 and 4.81, ps<.01, respectively). Study 4 replicated Study 3’s 3-5 examine the effect of the range-of-response-alternatives used to results using a different sample (n=167 Indian MBAs). elicit these estimates (Schwarz et al., 1985). A person is considered Study 5 (n=721 MTurkers) combined the designs of Studies 2 “addicted” when they have 5/8 symptoms of IAD. However, five and 4, manipulating elicitation at three levels (hi, lo, control open of these symptoms are more ambiguous than the remaining three. ended) x 2 (order) x 2 (None of the above: present/ absent), to show Overall, results show contextual cues affect risk judgments due to that these results are robust. Overall, results suggest that the greater symptom ambiguity: (a) the range of response alternatives used to the ambiguity of a symptom, the more prone it is to context effects, elicit internet usage affects reports of how many hours of the internet and the less likely it is to be assimilated into risk judgments. a person uses a day; (b) judgments of self-risk are lower when IAD 1132 / Obstacles and Opportunities to Consumer Engagement in Medical Decision Making Testing the Role of Motivation and Procrastination in participants who completed FIT within three weeks of invitation. Colorectal Cancer Screening Results indicated that both non-incentivized deadlines and deadlines coupled with a declining financial incentive (small or large) signifi- EXTENDED ABSTRACT cantly increased completion relative to standard outreach (ps < .005). Colorectal cancer (CRC) is the second leading cause of can- Further, completion in the short-deadline condition was higher than cer death in the U.S. (NIH, 2018). Although screening reduces CRC in the long-deadline condition (Short: 9.7% vs. Long: 7.2%; (1, N incidence and mortality (USPSTF, 2016), uptake is suboptimal, = 3,086) = 6.42, p = 0.011, ϕ = .05). Most notably, imposing a short especially among underserved populations (Gupta, 2013). Mailed non-incentivized deadline was just as effective as a deadline coupled outreach—inviting individuals to complete a fecal immunochemical with a financial incentive: completion in the short-deadline contest (FIT) in the comfort of their own home—offers a non-invasive dition (9.7%) was not statistically different than completion in the and more cost-effective approach. Yet, completion rates remain low small-declining(9.1%; (1, N = 3,076) = 0.33, p = 0.568, ϕ = .01) or (Singal et al., 2016). Using insights from behavioral science, we large-declining (12.0%; (1, N = 3,089) = 4.20, p = 0.04, ϕ = .04) con- predicted that both a lack of intrinsic motivation and procrastina- ditions. Thus, attenuating procrastination by imposing deadlines— tion tendencies may be contributing to low FIT completion. We test with or without financial incentives—increased FIT completion. these hypotheses in two large field experiments (N=16,336). Study Two large field experiments demonstrate that while fixed in- 1 aimed to increase screening by enhancing extrinsic motivation centives did not increase CRC screening relative to standard out- through financial incentives. Study 2 tested whether targeting pro- reach, deadlines did. Moreover, deadlines were equally effective at crastination—both with and without financial incentives—affected increasing FIT completion whether they were coupled with declining screening rates. We hypothesized that both financial incentives and financial incentives or were non-incentivized. These results impli- deadlines would increase screening completion relative to standard cate procrastination as a key barrier to FIT completion and suggest outreach. that offering modest incentives—at least in this context—amounts to Incentives are increasingly used in an attempt to influence paying for nothing. This paper provides insight into the psychology health behaviors (Saunders et al., 2018). However, while incentives underlying low cancer screening rates and suggests key consider- have been shown to motivate some behaviors (DellaVigna, 2016; ations in the use of behavioral science interventions to encourage Stone, 2002), their effectiveness is contextually dependent (Gneezy, health-behavior change. 2001), and whether they lead to long-term changes is uncertain (Charness & Gneezy, 2009). Thus, Study 1 tested the effectiveness REFERENCES of offering repeated financial incentives for FIT completion annually Akerlof, G.A. (1991), “Procrastination and obedience,” American for 3 years. Participants (N=8,565) received either standard mailed Economic Review. 81, 1-19. FIT outreach (N=6,565), outreach plus $5 (N=1,000), or outreach Ariely, D. & Wertenbroch, K. (2002), “Procrastination, deadlines, plus $10 (N=1,000) for FIT completion. Participants who completed and performance: Self-control by precommitment.” Psychol the test were re-invited using the same incentive the following year, Sci. 13, 219-224. for three years and then invited in Year 4 without any incentive. Charness, G. & Gneezy, U. (2009), “Incentives to exercise.” Contrary to predictions, completion was no different whether par- Econometrica, 77, 909-931. ticipants received an incentive, or not, in Year 1 (p = .59), Year 2 (p = Consumer Health Online 2017 (2017), “Consumer Health Online .75), or Year 3 (p = .08). There was also no difference in completion 2017”, dotHealth LLC www.get.health/research. across groups in Year 4, after incentives were discontinued (p = .08). DellaVigna, S. & Pope, D. (2017), “What motivates effort? Thus, offering modest incentives of $5 and $10 did not increase FIT evidence and expert forecasts” The Review of Economic completion relative to standard outreach. Study 2 therefore attempts Studies. 85, 1029-1069. to increase FIT completion by strengthening the financial incentives Fishbein, M., & Ajzen, I. (1975), Belief, Attitude, Intention, and intervention while also targeting a different psychological barrier— Behavior: An Introduction to Theory and Research. Reading, procrastination. MA: Addison-Wesley. Procrastination—delaying a task to a future day—is ubiquitous Fox, S. and Duggan, M. (2013), “Health Online 2013.” Washington, and often underlies failure to complete tasks (Akerlof, 1991; Tice & DC: Pew Research Center. Baumeister, 1997; Zauberman & Lynch, 2005). We posit that this Gupta, S. et al. (2013), “Comparative effectiveness of fecal deleterious tendency to procrastinate (Shu & Gneezy, 2010) may be immunochemical test outreach, colonoscopy outreach, and contributing to low FIT completion rates. Deadlines—particularly usual care for boosting colorectal cancer screening among the short deadlines—have been shown to attenuate procrastination (Ari- underserved: a randomized clinical trial,” JAMA Intern Med. ely & Wertenbroch, 2002; Shu & Gneezy, 2010). Thus, in Study 2, 173, 1725-1732. we test whether imposing deadlines increases FIT completion. Fur- Hsee CK (1996), “The evaluability hypothesis: an explanation for ther, while isolated incentives did not change behavior in Study 1, preference reversals between joint and separate evaluations of one explanation may be that there was no reference against which alternatives,” Organ Behav Hum Decis Process, 67, 247-57. participants could evaluate the incentive (Hsee, 1996). Thus, in addi- Hunter WG, Zafar SY, Hesson A, Davis JK, Kirby C, Barnett JA, tion to deadlines, we add a declining incentive component such that Ubel PA (2017), “Discussing Health Care Expenses in the participants are able to compare two incentives relative to one an- Oncology Clinic: Analysis of Cost Conversations in Outpatient other. Participants (N=7,771) were randomly assigned to receive one Encounters,” Journal of Oncology Practice. 13(11). 944-56. of five invitations: (1) control (standard invitation); (2) long deadline John, L.K. et al (2011), “Financial incentives for extended weight (three weeks); (3) short deadline (one week); (4) small declining in- loss: a randomized, controlled trial,” J Gen Intern Med. 26, centive ($10 for completion within one week or $5 for completion 621-626. within three weeks); or, (5) large declining incentive ($20 for com- Klayman, J. and Ha, Y.W. (1987), “Confirmation, disconfirmation, pletion within one week or $10 for completion within three weeks). and information in hypothesistesting,” Psychological Review. The prespecified primary dependent variable is the proportion of 94. 211–28. Advances in Consumer Research (Volume 48) / 1133

Menon, Geeta, Priya Raghubir, and Norbert Schwarz (1995), Singal, A.G. et al. (2016), “Outreach invitations for FIT and “Behavioral Frequency Judgments: An Accessibility- colonoscopy improve colorectal cancer screening rates: A Diagnosticity Framework,” Journal of Consumer Research, 22 randomized controlled trial in a safety‐net health system,” (2), 212-228. Cancer. 122, 456-463. ______, ______and ______(1997), “How Much Will I Stone, E.G., et al, (2002), “Interventions that increase use of adult Spend? Factors Affecting Consumers’ Estimates of Future immunization and cancer screening services: a meta-analysis,” Expenses,” Journal of Consumer Psychology, 6 (2), 141-164. Ann Intern Med. 136, 641-651(2002). O’Donoghue, T. & Rabin, M (1999), “Doing it now or later,” Am Sun, Yalin, Yan Zhang, Jacek Gwizdka, and Ciaran B Trace (2019), Econ Rev. 89, 103-124. “Consumer Evaluation of the Quality of Online Health Posavac, S., Kardes, F., & Brakus, J. (2010), “Focus induced tunnel Information: Systematic Literature Review of Relevant vision in managerial judgment and decision making: The Criteria and Indicators.” Journal of Medical Internet Research peril and the antidote,” Organizational Behavior and Human 21, no. 5. Decision Processes, 113, 102–111. Surveillance, Epidemiology, and End Results Program—Cancer Raghubir, Priya and Geeta Menon (1998), “AIDS and Me, Never stat facts: colorectal cancer (NIH, 2018); seer.cancer.gov/ the Twain Shall Meet: Factors Affecting Judgments of Risk,” statfacts/html/colorect.html Journal of Consumer Research, 25(1), 52-63. Tice, D.M. and Baumeister, R.F. (1997), “Longitudinal study of ______and ______(2005), “Depressed or Just Blue? The procrastination, performance, stress, and health: The costs and Persuasive Effects of a Self-Diagnosis Inventory”Journal of benefits of dawdling,”Psychol Sci. 8, 454-458. Applied Social Psychology, 35(12), 2535-2559. Ubel PA. (2019), Sick to Debt: How Smarter Markets Lead to Schwarz, Norbert, Hans-J. Hippler, Brigitte Deutsch, and Fritz Better Care, Yale University Press. New Haven & London. Strack (1985), “Response Scales: Effects of Category Range Ubel PA, Zhang CJ, Hesson A, Davis JK, Kirby C, Barnett J, on Reported Behavior and Comparative Judgments,” Public Hunter WG. (2016), “Study Of Physician And Patient Opinion Quarterly, 49(Fall), 388-395. Communication Identifies Missed Opportunities To Help Saunders, R., Vulimiri, M., Japinga, M., Bleser, W. & Wong, C Reduce Patients’ Out-Of-Pocket Spending,” Health Affairs. (2019), “Are carrots good for your health?” Current evidence 35(4): 654–661. on health behavior incentives in the Medicaid program. Vol. 1 Zauberman, G., & Lynch Jr, J. G (2005), “Resource slack and (Duke Margolis Center for Health Policy). propensity to discount delayed investments of time versus Screening for Colorectal Cancer US Preventive Services Task Force money,” J Exp Psych Gen. 134, 23-37. Recommendation Statement (2016), JAMA. 315, 2564–2575. Shu, S.B. & Gneezy, A. (2010), “Procrastination of enjoyable experiences,” J Mark Res. 47, 933-944. The Social and Behavioral Consequences of Browsing Social Media Chairs: In-Hye Kang, University of North Carolina Greensboro, USA Yuechen Wu, Johns Hopkins University, USA

Paper #1: Social Media Users are Penalized for Lacking Self- about the past (vs. present) less favorably because they perceive the Control bragging as more intentional. As a result, consumers become less In-Hye Kang, University of North Carolina Greensboro, USA likely to engage with the mentioned brand when the bragging is Yuechen Wu, Johns Hopkins University, USA about past (vs. present) experiences. Neeru Paharia, Georgetown University, USA Moving to the corporate social media posting, Rifkin, Chan, and Kahn examine how encountering postings about brand events Paper #2: I Care About Why You Share: Inferences about can affect consumers’ future engagement with the brand. Theau- Sharing Motives Influence Observers’ Engagement in Similar thors find that seeing social media photos of brand events can trigger Experiences FOMO for consumers who missed the events, increasing these con- Matthew J. Hall, Oregon State University, USA sumers’ intentions to engage with the brand. Daniel M. Zane, Lehigh University, USA Taken together, the four papers examine how browsing social Paper #3: The Time-Dependent Effects on Bragging media postings can influence consumers’ various perceptions and Francesca Valsesia, University of Washington, USA behaviors. This session is highly relevant to one of the themes of Jared Watson, New York University, USA ACR 2020, “Technology,” as the four papers generate timely wisdom Paper #4: The Role of Social Media-Induced FOMO in on how social media—a relatively new online platform made pos- Strengthening Brand Communities sible by advances in the mobile and Internet technology—impacts Jacqueline R. Rifkin, University of Missouri-Kansas City, consumer behavior. This session will appeal to a broad audience, USA including researchers interested in social media, people perception, Cindy Chan, University of Toronto, Canada inference making, temporal frame, and branding. Barbara E. Kahn, University of Pennsylvania, USA Social Media Users are Penalized for Lacking Self- SESSION OVERVIEW Control Consumers today spend a significant amount of time on social media platforms such as Facebook and Instagram. In 2018, Inter- EXTENDED ABSTRACT net users spent about 1.25 hours per day on social media platforms Reflecting the popularity of consumers’ social media use, a (Statista 2019). As such, a growing amount of research has started growing amount of research has examined the motivations for using to examine consumers’ motivations for posting on social media social media (Buechel and Berger 2018; Toubia and Stephen 2013) (Buechel and Berger 2018; Toubia and Stephen 2013) and how post- and the consequences of social media use (Barash et al. 2018; Wil- ing about one’s experiences on social media can influence their well- cox and Stephen 2013). Prior literature, however, has not explored being, including enjoyment of an experience (Barash et al. 2018). what inferences observers make about social media users. This is an Yet, little research has examined how browsing social media and important gap because consumers often use social media in publicly more specifically the content consumers encounter while browsing visible situations, such as while waiting for coffee (Herhold 2018). social media can influence consumers’ judgments and behavior. For We argue that visibly using social media can trigger negative infer- example, what does browsing social media in publicly visible situ- ences of increased addiction and lower self-control by observers, ations signal to others? How does seeing other consumers’ postings leading to unfavorable downstream consequences. about their experiences influence consumers’ own intention to con- Prior research suggests that social media use can be addictive sume similar experiences? How does encountering a brand’s post- (Kuss and Griffiths 2011). In our pretest (n=100), when asked to ing about its offline brand event influence consumers’ intentions to evaluate the addictiveness of several activities people do on their engage with the brand? smartphones, people believed that browsing social media was more This session sheds light on these questions. Kang, Wu, and addictive than reading a magazine, reading news, watching a movie Paharia examine how observers judge consumers who browse so- or even playing a game (ps ≤ .001). Therefore, we expect that ob- cial media in publicly visible situations. The authors find that visibly servers will infer that a person using social media is more addicted using social media can trigger negative inferences of increased ad- compared to a consumer engaging in other activities, leading to the diction and lower self-control by observers, leading to unfavorable inference of lower self-control. We further propose that the inference downstream consequences, such as being less likely to be chosen as of low self-control will lead to numerous negative reactions, such as a service provider or as a partner to work on a task requiring self- being less likely to be chosen as a service provider. control. We identify two theoretical boundary conditions. First, if a tar- The next two papers examine how browsing other consumers’ get consumer uses social media only for a small fraction of his free postings impact one’s future consumption. Hall and Zane find that time, the inference of addiction should be less, attenuating the nega- when consumers infer that others posted their experiences on social tive inferences on self-control. Second, individuals who have posi- media for extrinsically-motivated (vs. intrinsically-motivated) rea- tive attitudes towards social media may associate using social media sons, they show lower desire to engage in similar experiences, be- more strongly with positive attributes rather than negative attributes cause they anticipate that other people will infer extrinsic motives if such as addiction, again attenuating the negative inferences on self- they engage in similar experiences. control. Turning towards interventions, we expect that consumers Valsesia and Watson examine how consumers differentially are not spontaneously aware of the negative inferences about social evaluate others’ self-bragging postings depending on the temporal media users. Thus, we hypothesize that prompting consumers to take frame. The authors find that consumers evaluate others’ bragging

Advances in Consumer Research 1134 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1135 an observer’s role can reduce their own intention to use social media evaluated the woman in the three scenarios, participants indicated in public. what activities they would do while waiting in line at a café. Par- Studies 1-2 tested negative downstream consequences using an ticipants in the after condition (56.1%) were less likely to choose incentive-compatible choice. In study 1 (n=333), participants viewed to browse social media than those in the before condition (71.7%; the bios of two personal trainers at a local gym. One trainer indicated p = .002). that he liked to browse social media in his free time while the other This research contributes the literature on social media. Practi- liked to watch movies. 35.44% of participants chose the trainer who cally, our findings can better guide consumers’ impression manage- liked to browse social media (p < .001) because they perceived the ment strategies. trainer as having lower self-control (p < .001). In study 2 (n=118), participants selected a partner with whom I Care About Why You Share: Inferences about Sharing they would like to work on a word search task. We showed partic- Motives Influence Observers’ Engagement in Similar ipants the profiles of the two partner candidates. The two profiles Experiences were similar except that [one/the other] student spent [55%/10%] of her free time using social media. 33% of participants chose a heavy EXTENDED ABSTRACT social media user (p < .001) because they perceived the heavy user While some consumers post about their experiences on social as having less self-control (p < .001). media because they genuinely enjoy giving others insight into their Are the participants’ perceptions accurate? In a separate test, lives, others share to demonstrate their uniqueness or consumption we asked participants to engage in the same word search task by of conspicuous/luxury experiences (Bronner and de Hoog 2018; themselves. Social media use was not correlated with participants’ Kim 2018). Given that consumers’ sharing motives differ, do others performance (p = .98), suggesting that people’s beliefs about social who view shared content make different inferences about the sharers’ media users are not accurate. posting motives? If so, might these inferred sharing motives influ- In studies 3A-3B, we tested our multi-step mechanism (social me- ence observers’ desire to have a similar experience? dia use→perceived addiction→perceived self-control→downstream Prior research on perceptions of others’ sharing motives has consequences). In study 3A (n=247), participants in the [social me- only explored perceived motives behind sharing about experiences dia/online magazine] condition viewed a picture of a woman scroll- relative to material goods. This work demonstrates that perceived ing through [a social media account/an online magazine] on her sharing motives influence evaluations of the sharer—observers infer phone at the airport. In both conditions, we showed the same images the sharer to have relatively more intrinsic motives when viewing on the phone to control for the content the woman was browsing. The shared experiences (vs. material goods), leading to more positive woman in the social media (vs. magazine) condition was evaluated evaluations of the sharer (Van Boven et al. 2010). However, it is less positively overall (p = .012), as having less self-control (p = unlikely that all experiential sharing is perceived to be equally in- .006), and as being more addicted (p < .001). The hypothesized serial trinsically-motivated. In addition, prior research has yet to consider mediation path was significant. Neither impression management nor how perceived motives behind sharing about an experience might materialism served as a significant mediator. In study 3B (n=488), influence observers’ preference to purchase a similar experience. we replicated the findings of study 3A by comparing browsing so- We first propose that consumers will infer a broad range of cial media with other control conditions: playing an online game and sharing motives when viewing others’ shared experiences. As ini- browsing news. The negative effect of social media occurred when tial support, pilot study participants (n=92) viewed a post featuring the target person browsed social media on a smartphones or laptop. someone’s camping experience and described the poster’s sharing Study 4 (n=418) tested the moderating effect by the percentage motives. Two independent coders rated these qualitative descriptions of free time used for social media. Participants in the [100%/4%] (-3=purely extrinsic motives; 0=equally intrinsic/extrinsic motives; condition read a scenario that the flight was boarding in [five min- 3=purely intrinsic motives; M= -.223; SD=1.91; interrater reliabil- utes/two hours] and the woman was looking at her phone for five ity=.812). This distribution suggests that consumers infer sharers to minutes by browsing [Instagram/the Wall Street Journal]. We mea- have different motives when sharing about their experiences. sured willingness to hire the woman for a job as a downstream mea- Second, we propose that when consumers infer a sharer’s post sure. The interaction effect was significant for the intention to hire to be more extrinsically-motivated, it will decrease the desirability of (p = .014). Participants in the social media (vs. news) condition re- the shared experience itself. This will occur because consumers’ in- ported lower intention to hire in the 100% condition (p = .001), not ferences about the sharer’s motives will influence the observers’ own in the 4% condition (p = .89). We found similar patterns of result for perceptions of how others might view them if they had a similar ex- perceived self-control. perience (social motivation contagion; Wild et al. 1992). If consum- Study 5 (n=431) tested the moderating role of attitudes to- ers anticipate others will see them as having more extrinsic motives wards social media. Participants read a scenario that a 30-year-old if they engage in an experience, it should decrease the experience’s man at the airport was [browsing his Instagram account/reading a desirability (Wild and Enzle 2002). Thus, while firms are assumed magazine]. The regression analysis for the overall favorability of the to benefit when customers share about their firm-related experiences man revealed a significant interaction (p = .023). Specifically, the (WOM, buzz, etc.), we suggest such sharing may sometimes deter negative effect of social media (vs. magazine) on overall favorability other customers if the sharer is perceived to have extrinsic sharing occurred for participants with attitudes towards social media lower motives. than 5.80. We found similar results for perceived self-control. In Study 1, participants (n=361) viewed a mock Instagram post Study 6 (n=378) tested the proposed interventions to reduce so- featuring either a jacket (product) or camping scene (experience). cial media use. All participants read the three scenarios describing a Participants reported their likelihood to purchase the featured ex- woman [browsing her social media account on her phone/browsing perience/product and their perceptions of the poster’s sharing mo- news on her phone/looking around at the café]. After reading each tives (scale from pilot study). They also evaluated the sharer on a scenario, participants rated the woman on self-control and overall fa- number of positive/negative traits (Van Boven et al. 2010). Repli- vorability. In the before (after) condition, before (after) participants cating prior research, those viewing the experience post perceived 1136 / The Social and Behavioral Consequences of Browsing Social Media the sharer to have more intrinsic motives (MExperience=.52; SD=1.89; The Time-Dependent Effects on Bragging

MMaterial= -.97; SD=1.82; F(1,359)=61.10, p<.001). Pertinent to our hypothesis, however, there was heterogeneity in perceived sharing EXTENDED ABSTRACT motives within the experience condition, evidenced by a mean near Much research has been dedicated to understand why people the scale midpoint and substantial variance. Furthermore, a regres- brag (Baumeister 1982; Speer 2012) and perceptions of those who sion revealed an interaction between the experience/product condi- do (Tal-Or 2010; Berman et al. 2015). Self-promotion is a fact of tion and perceived sharing motives on product/experience purchase life, but with the proliferation of social media usage, people encoun- likelihood (b= -.249; t(357)=2.26, p=.024). As expected, within ter braggarts more than ever before. More recently, the conversa- the experience condition, those who perceived the sharer to have tion has shifted to understand characteristics of social media posts more extrinsic motives reported lower purchase likelihood (b=.296; (Matley 2018a, 2018b). For example, Matley (2018a) demonstrates t(357)=3.76, p<.001). Interestingly, perceived sharing motives did that the explicit use of the hashtag “#brag” attenuates some of the not affect purchase likelihood in the product condition (b=.046; negativity that might be commuted through a self-praising message. t(357)=.60, p=.551). Perceived motives equally predicted evalua- In this research, we explore other message characteristics that might tions of the sharer in both conditions, suggesting our consumption- moderate the effect of a self-praising message. Our primary focus related consequences are driven by a distinct process than that in Van investigates the temporal framing of a message. Recent work has Boven et al. (2010). demonstrated that people are more likely to share information about In Study 2, undergraduates (n=338) identified a post in their rewards that will occur in the future vs. the past (Weingarten and own social media accounts featuring an experience shared by some- Berger 2017), but little is known about how the temporal framing of one they follow. After describing the experience, they reported how a message affects how consumers respond to it. much they would enjoy consuming that same experience (1=not at This research posits that past-oriented self-promotion is viewed all; 7=very much) and the sharer’s motives (-2=purely extrinsic; negatively relative to both present-oriented self-promotion as a func- 2=purely intrinsic)—these measures were counterbalanced and or- tion of the message’s perceived intentionality. Present-oriented mes- der had no effect. Perceived sharing motives were normally distrib- sages are thought to be spontaneous and emotion-laden, whereas uted (M=.097; SD=1.13). Furthermore, perceptions of the sharer’s past-oriented messages are perceived to be intentional signals sent by motives predicted anticipated enjoyment of the experience (b=.232; the messenger. Our first study employed a 2-cell design (time: past, t(334)=1.90, p=.059). Participants then also justified why they present), between-subjects. Participants were told that they would scored the sharer’s motives as they did. An independent research as- view a tweet then be asked to evaluate the individual. The tweet read sistant coded these explanations for themes about what aspects of either “I cannot believe I just won the Valley half marathon! Hard shared experiences led to inferences of intrinsic/extrinsic motives. work pays out! #proud #win #today” or “I cannot believe it is ex- Participants perceived sharers to have more extrinsic motives when actly 5 years since I won the Valley half marathon! Hard work paid the sharer did not seem truly interested in the experience itself (con- out! #proud #win #throwback”. We asked participants to what extent suming for the purpose of sharing) or when the sharer’s presence they perceived the tweet to be “bragging” and a one-way ANOVA in the photo seemed to be the focus (staged poses, heavy makeup, indicated that participants viewed tweeting about the past to be sig- provocative clothing, etc.). Captions featuring text unrelated to the nificantly more braggy (Mpast = 6.66, Mpresent = 6.05; F(1,163) = experience (poems, quotes, etc.) or that conveyed social comparison 4.82; p = .03). Moreover, we found impressions of the sender to be or tried to elicit sympathy also signaled extrinsic motives. significantly more positive when the person tweeted about the pres- In Study 3, undergraduates (n=114) viewed a mock Instagram ent (Mpresent = 5.30) compared to the past (Mpast = 4.68; (1,163) = post featuring a bowling experience, manipulated based on the 6.09; p = .01). In reflection, we realized that the message also had an themes discovered in the previous study. In the intrinsic post, the element of skill given the amount of effort required to win a mara- sharer was less prominent and was displayed in the act of bowling thon. Thus, in the next study we sought to test whether our effect held with a caption related to bowling. In the extrinsic post, the sharer when a brag lacked skill and happened due to sheer luck. was featured prominently in a staged pose with a caption unrelated In study 2, we employed a 2(time: past, present) x 2(deserving- to bowling. Participants then reported how likely they would be ness: luck, skill), between-subjects design. This time, an Instagram to purchase a bowling experience. They also reported how others post was used to further generalizability. A picturesque waterfall was might view their own motives for going bowling, including three accompanied with the caption “one year ago today, I was in Costa intrinsically-motivated explanations (e.g., for the enjoyment of the Rica…” or “right now, I am in Costa Rica…” to manipulate the tem- experience) and three extrinsically-motivated explanations (e.g., to poral factor. To manipulate the deservingness factor, the caption also impress others), averaged to create our focal mediator. As expected, mentioned that the trip was won from a “random raffle for new em- mediation analysis revealed that participants who viewed the ex- ployees” or “receiving the highest customer satisfaction score for trinsic post thought others would view their own motives for go- my company”. A 2x2 ANOVA on participants’ overall evaluation of ing bowling as more extrinsic (MExtrinsic=4.27; MIntrinsic=5.01; b=.368; the poster revealed a significant interaction (F(1,111) = 4.46; p= t(111)=3.57, p<.001), which then predicted purchase intentions .04). When the brag was framed in terms of skills, and thus had a

(b=.285; t(111)=2.16, p=.032; index of mediation: CI95=[.005,.245]). deservingness component, we replicate our earlier effect, such that a We demonstrate that when consumers are perceived to have present brag led to more positive evaluations of the sender compared extrinsic motives for sharing experiences, observers see the same to past brags (Mpast = 3.86 vs. Mpresent = 4.63, F(1,111) = 3.54; p = experience as less desirable to consume themselves. This occurs be- .06). Nonetheless, this was not the case when the brag was framed cause observers assume that others will perceive their own motives in terms of luck (Mpast = 4.55 vs. Mpresent = 4.22, F(1,111) = 1.23; p as being more extrinsic if they are to consume the experience. These = NS). results suggest that firms do not always benefit when their customers Thus far, we have established that temporal framing matters, share about their firm-related experiences and identify characteristics but we’ve done little to determine the velocity of the effect. In study that lead consumers to infer different motives behind shared content. 3, we assign participants to past, present, and future conditions to explore whether disclosing upcoming events are bragging as well. Advances in Consumer Research (Volume 48) / 1137

However, to account for individual differences in time perceptions ture people from their local community, rather than from the brand we ask participants to report the extent to which they view these community (H3). Second, building on previous research on FOMO events as occurring between the distant past and distant future identifying the critical role of anxious attachment (Rifkin, Chan, and (5-point scale) and use this item as our primary independent variable. Kahn 2020), we predict that the effects will be stronger among those We do this to account for the subjective nature of time and deter- who feel more anxious about their belonging in the brand commu- mine the trend between distant and near time horizons. Participants nity (H4). Finally, consistent with work on fear appeals, we predict viewed a Facebook post discussing the purchase of a new BMW. We a boundary condition in which the effects will be attenuated when then asked them to evaluate the poster. In this study we also mea- consumers are not offered a means of coping with their FOMO (e.g., sured behavioral intentions to determine the managerial relevance if an event is private and closed to the public; H5). of our effect. A one-way ANOVA of temporal perceptions on poster We first conducted a field experiment in conjunction with Lu- evaluations yielded a significant effect (F(4, 403) = 2.807;p = .025). lulemon (N = 278). The study began with an initial survey where we A curvilinear function emerged such that a positive relationship ex- measured participants’ connection to the brand and their brand inten- ists from the distant past to near future (Mdistant past = 1.56, Mrecent past = tions. We then invited participants to an actual in-store event; pro-

2.97, Mpresent = 3.06, Mnear future = 3.44) before exhibiting a negative fessional photos of the event were subsequently posted on the local effect in the distant future (Mdistant future = 2.92). Similarly, a one-way Lululemon Facebook page. Afterwards, we conducted a follow-up ANOVA of temporal perceptions on brand intentions yielded the survey in which participants were directed to a Lululemon Facebook same pattern (Mdistant past = 2.67, Mrecent past = 3.57, Mpresent = 3.62, Mnear album containing either photos of the event or Lululemon merchan- future = 4.12, Mdistant future = 2.92; F(4,403) = 2.486; p = 043). Natu- dise. Among participants who did not attend the event (>90% of the rally, some of these cells suffer from small sample sizes, but none- sample), we found that, across both photo conditions, participants theless, further investigation into temporal perceptions is warranted. with higher brand connections felt more FOMO after seeing the pho- Through three studies, this research demonstrates that temporal tos (p < .001), and this increased their future brand intentions (95% framing of messaging yields significant effects of not only messen- CI = .04, .15), providing support for H1 and H2. Contrasting the two ger perceptions but can also have consequences for brands. Bragging photo conditions, we found that, relative to seeing merchandise pho- about the past vs. the present (or the future) seems to have negative tos, seeing event photos featuring people from the local community consequences for both but this is the temporal frame in which most (i.e., fellow university students) increased brand intentions among people brag. Thus, our next steps will be to investigate strategies by people who were initially less connected to the Lululemon brand (p which firms can encourage consumers to brag about present or future = .03; H3). interactions vs. past, and to determine whether consumers are aware In a subsequent lab study (N = 355) with more experimental of this bias when craft a message in the first place. control, we re-ran the same three-phases of the field study (using the same photos from the previous event). In the pre-event survey, we The Role of Social Media-Induced FOMO in also measured participants’ anxious attachment to the brand com- Strengthening Brand Communities munity (adapted from Wei et al. 2007) and brand intentions toward Lululemon. We then used the same two conditions from the field EXTENDED ABSTRACT study: participants were directed to Facebook pages featuring either We are witnessing a dramatic change in the world of retailing. event or merchandise photos. Finally, we conducted a post-event Rather than shopping in store, consumers can order anything they survey measuring FOMO and brand intentions. We again found that want online. However, while online shopping has changed the retail- seeing social media photos increased FOMO among those with high- ing scene, no one is predicting that physical stores are going away. er brand connection (p = .003; H1), and this increased their future To drive traffic to physical stores, specialty and brand retailers are in- brand intentions (95% CI = .01, .05; H2). Supporting H4, we also creasingly hosting in-store events and then posting photos on social found that seeing event (vs. merchandise) photos increased brand media. For example, Footlocker in Harlem, NY hosted a challenge intentions among those who were more anxiously attached to the where fans competed to see who had the highest vertical jump. The brand community (interaction p = .02). event featured celebrity appearances and elite sock customization, As a further test of H3, we conducted a lab study (N = 702) in and photos were posted on Footlocker social media pages. Similarly, which we manipulated whether the people in the event photos were Lululemon frequently hosts in-store yoga classes and posts event from the local community or the brand community. In this and the photos to social media. subsequent study, we also allowed participants to choose from sev- While these in-store events likely generate positive outcomes eral brands to generalize our effects. We replicated the finding that for customers who participate, how do these strategies affect cus- highly brand-connected consumers felt more FOMO from seeing the tomers who do not attend? We predict that seeing photos of missed photos (p < .001; H1), which in turn increased brand intentions (95% brand-community events will elicit a negative emotional feeling of CI = .03, .14; H2). We also observed a photo condition by brand FOMO (Fear of Missing Out) among those who feel more connected connection interaction in which seeing local (vs. brand) community to the brand (H1). Although this negative response may seem un- photos increased brand intentions among less-connected consumers desirable, prior research has found that fear appeals can positively (p < .001; H3). influence intentions and behaviors (Tannenbaum et al. 2015). We Our final study (N = 704) tests a boundary condition (H5) by therefore predict that consumers will cope with their negative feel- asking participants to imagine seeing social media photos from a ings of FOMO by increasing their intentions to engage with the brand event that was either open or closed to the public. We pre- brand in the future (e.g., by attending a future event; H2). dicted that only open events would offer consumers a means for cop- We also identify three moderators. One benefit of these events ing with their feelings of FOMO and would therefore increase brand is that they foster local brand communities, attracting people who intentions, whereas closed events would not. We observed a signifi- are members of the local community, but may not be brand loyal- cant mediated moderation whereby seeing event photos increased ists. Therefore, we predict that less brand-connected consumers FOMO among highly brand-connected consumers (95% CI = -.26, will respond more favorably when seeing event photos that fea- -.12). However, this enhanced FOMO only increased brand inten- 1138 / The Social and Behavioral Consequences of Browsing Social Media tions when the event was open (95% CI = .11, .19), but not when it Kim, Youngseon, (2018), “Power Moderates the Impact of Desire was closed to the public (95% CI = -.05, .06). for Exclusivity on Luxury Experiential Consumption,” In summary, we find that seeing social media photos of local Psychology & Marketing, 35 (4), 283-93. brand events triggers negative feelings of FOMO, which increases Van Boven, Leaf, Margaret C. Campbell, and Thomas Gilovich intentions to engage with the brand in the future. These effects are (2010), “Stigmatizing Materialism: On Stereotypes and stronger among those who feel more connected to the brand and Impressions of Materialistic and Experiential Pursuits,” more anxious about their belonging. This research broadens our con- Personality and Social Psychology Bulletin, 36 (4), 551-63. ceptualization of brand community and has implications for social Wild, T. Cameron and Michael E. Enzle (2002), “Social media strategies. Contagion of Motivational Orientations,” Handbook of Self- Determination, eds. Edward L. Deci and Richard M. Ryan, REFERENCES University of Rochester Press: Rochester, New York. Wild, T. Cameron, Michael E. Enzle, Glen Nix, and Edward SESSION OVERVIEW L. Deci (1997), “Perceiving Others as Intrinsically or Barasch, Alixandra, Gal Zauberman, and Kristin Diehl (2018), Extrinsically Motivated: Effects on Expectancy Formation “How the Intention to Share Can Undermine Enjoyment: and Task Engagement,” Personality and Social Psychology Photo-Taking Goals and Evaluation of Experiences,” Journal Bulletin, 23 (8), 837-848. of Consumer Research, 44 (6), 1220–37. Buechel, Eva C. and Jonah Berger (2018), “Microblogging and the The Time-Dependent Effects on Bragging Value of Undirected Communication,” Journal of Consumer Baumeister, R. F. (1982). A self-presentational view of social Psychology, 28 (1), 40–55. phenomena. Psychological Bulletin, 91(1), 3. Statista (2019), “Daily Time Spent on Social Networking by Berman, J. Z., Levine, E. E., Barasch, A., & Small, D. A. (2015). Internet Users Worldwide from 2012 to 2018,” https://www. The Braggart’s dilemma: On the social rewards and penalties statista.com/statistics/433871/daily-social-media-usage- of advertising prosocial behavior. Journal of Marketing worldwide/ (Retrieved at January 2, 2019). Research, 52(1), 90-104. Toubia, Olivier and Andrew T. Stephen (2013), “Intrinsic vs. Matley, D. (2018a). “This is NOT a# humblebrag, this is just a# Image-Related Utility in Social Media: Why Do People brag”: The pragmatics of self-praise, hashtags and politeness Contribute Content to Twitter?,” Marketing Science, 32 (3), in Instagram posts. Discourse, context & media, 22, 30-38. 368–92. Matley, D. (2018b). “Let’s see how many of you mother fuckers unfollow me for this”: The pragmatic function of the hashtag# Social Media Users are Penalized for Lacking Self- sorrynotsorry in non-apologetic Instagram posts. Journal of Control Pragmatics, 133, 66-78. Barasch, Alixandra, Gal Zauberman, and Kristin Diehl (2018), Speer, S. A. (2012). The interactional organization of self-praise: “How the Intention to Share Can Undermine Enjoyment: Epistemics, preference organization, and implications for Photo-Taking Goals and Evaluation of Experiences,” Journal identity research. Social Psychology Quarterly, 75(1), 52-79. of Consumer Research, 44 (6), 1220–37. Tal-Or, N. (2010). Bragging in the right context: Impressions Buechel, Eva C. and Jonah Berger (2018), “Microblogging and the formed of self-promoters who create a context for their boasts. Value of Undirected Communication,” Journal of Consumer Social Influence, 5(1), 23-39. Psychology, 28 (1), 40–55. Weingarten, E., & Berger, J. (2017). Fired up for the future: How Herhold, Kristen (2018), “How People Use Social Media in 2018,” time shapes sharing. Journal of Consumer Research, 44(2), The Manifest, https://themanifest.com/social-media/how- 432-447. people-use-social-media-2018 (Retrieved at January 2, 2019). Kuss, Daria J. and Mark D. Griffiths (2011), “Online Social The Role of Social Media-Induced FOMO in Networking and Addiction—A Review of the Psychological Strengthening Brand Communities Literature,” International Journal of Environmental Research Rifkin, J. R., Chan, C., & Kahn, B. E. (2020). FOMO: How the and Public Health, 8 (9), 3528–52. Fear of Missing Out Leads Missing Out. Working paper. Toubia, Olivier and Andrew T. Stephen (2013), “Intrinsic vs. Tannenbaum, M. B., Hepler, J., Zimmerman, R. S., Saul, L., Jacobs, Image-Related Utility in Social Media: Why Do People S., Wilson, K., & Albarracín, D. (2015). Appealing to fear: Contribute Content to Twitter?,” Marketing Science, 32 (3), A meta-analysis of fear appeal effectiveness and theories. 368–92. Psychological Bulletin, 141(6), 1178–1204. Wilcox, Keith and Andrew T. Stephen (2013), “Are Close Friends Wei, M., Russell, D. W., Mallinckrodt, B., & Vogel, D. L. (2007). the Enemy? Online Social Networks, Self-Esteem, and Self- The experiences in close relationship scale (ECR)-short Control,” Journal of Consumer Research, 40 (1), 90–103. form: Reliability, validity, and factor structure. Journal of Personality Assessment, 88(2), 187-204. I Care About Why You Share: Inferences about Sharing Motives Influence Observers’ Engagement in Similar Experiences Bronner, Fred and Robert de Hoog (2018), “Conspicuous Consumption and the Rising Importance of Experiential Purchases,” International Journal of Market Research, 60 (1), 88-103. New Insights from Computational Models of Cognition in Consumer Research Chair: Rory M. Waisman, University of Alberta, Canada

Paper #1: Insights From a Process Model of Retrospective tal representation-based and are considered subjective decisions vs. Evaluation objective judgments. Finally, the last paper offers a cautionary note Rory M. Waisman, University of Alberta, Canada about how best to apply computational models of cognition in consumer research. The authors show that estimation and interpretation Paper #2: What I Like Is What I Remember: Memory of free model parameters is hampered under certain modeling condi- Modulation in Preferential Choice tions, and they offer solutions to help researchers overcome these Ada Aka, University of Pennsylvania, USA challenges and avoid deriving spurious insights from computational Sudeep Bhatia, University of Pennsylvania, USA models of cognition. Paper #3: Packaging vs . Product: Distinguishing Between The papers in this session offer concrete examples of how mod- Choice Domains with Computational Modeling els and methods developed in cognitive science can be fruitfully ap- Stephanie M. Smith, University of California Los Angeles, plied in consumer research to deepen our understanding of consumer USA behavior phenomena. This session will appeal to scholars interested Ian Krajbich, The Ohio State University, USA in preference construction, decision making, cognitive processes (at- Paper #4: Structural Dependencies in Computational Models of tention and memory in particular), as well as those with an interest in Cognition in Consumer Research the application of modeling methods in consumer research. Antonia Krefeld-Schwalb, Columbia University, USA Thorsten Pachur, Max Planck Institute for Human Insights From a Process Model of Retrospective Development, Germany Evaluation Benjamin Scheibehenne, University of Geneva, Switzerland EXTENDED ABSTRACT SESSION OVERVIEW Consumers’ evaluations of consumption experiences are im- This session showcases how computational models of cogni- portant to researchers, firms, and consumers themselves. Scholars tion can be leveraged to offer new insights for consumer research. depend on retrospective self-reports (e.g. Heitmann, Lehmann, and Marketing scholars have been calling for greater cross-pollination Herrmann 2007), firms treat customer ratings as metrics ofsuc- between theory-rich cognitive science and the more pragmatically cess (Keiningham, Cooil, Andreassen, and Aksoy 2007; Reichhel, oriented field of consumer research (e.g., Bartels and Johnson 2015; 2003), and consumers rely upon their own (Moore 2011) and oth- Weber and Johnson 2006). This call is being increasingly answered ers’ (Nielse, 2015) evaluations as a basis for consumption decisions in the cognitive and psychology literature (e.g., Bhatia and Mullett (Babić Rosario, Sotgiu, De Valck, and Bijmolt 2016). Research has 2016; Bhatia and Pleskac 2019; Trueblood, Brown, and Heathcote revealed innumerable variables that influence retrospective evalua- 2014; Trueblood, Brown, Heathcote, and Busemeyer 2013), allow- tion, but more limited progress is evident in understanding the basic ing cognitive science to profit from validating models and theories cognitive mechanisms responsible. Weber and Johnson (2006) called in more realistic consumer settings (e.g., Vinson, Dale, and Jones for explicit exposition of the memory processes involved in prefer- 2019). But there has been less advancement along these lines in the ences construction. I respond with a process model of retrospective field of consumer research. In fact, this call was echoed again last evaluation: Minerva Evaluative Memory (Minerva-EM). year during a Knowledge Forum on Models of Behavioral Decision Comparison of model predictions to empirical results suggests Making held at the 50th ACR Conference. Participants appealed for the model’s instance encoding, resonance-based retrieval, and evalu- greater openness to incorporating modeling approaches from differ- ative mechanisms explain consumers’ evaluative behavior. Deriva- ent disciplines, and some argued that consumer research can benefit tion from the model of novel, empirically testable hypotheses dem- by the application of theory and methods from cognitive psychol- onstrates the value of applying computational models of cognition in ogy. This session aims to answer the call by presenting four papers consumer research. that apply computational models of cognition to better understand Inspired by Hintzman’s (1984) Minerva-2 memory model, Mi- the mechanism that underlie consumers’ evaluations of products and nerva-EM is a precise mathematical account of memory and evalu- experiences and their construction of preferences. ative mechanisms. According to Minerva-EM, memory is single The first paper introduces a process model of basic encoding, system for encoding and storage of episodic memory traces, and retrieval, and evaluative mechanisms that account for the behav- retrieval operates by principles of resonance. In the computational ior of consumers when they rely on memory to evaluate their con- instantiation of the theory, memory is a matrix with each trace consumption experiences. The author uses model simulations to derive stituting one vector. An experience is encoded as a series of traces novel hypotheses about variables that alter consumers’ retrospective for constituent events. Retrospective evaluation involves probing evaluations. In the second paper, a computational model of language memory to retrieve a representation matching the specified context. representation is utilized to help identify commonalities and isolate This produces an effect analogous to singing into a piano. The strings key differences between the cognitive processes employed in pure- of the piano (the traces) resonate (are activated) in response to the memory tasks and those employed in making preferential choices singer’s voice (the probe). In effect, the piano (memory) produces an that rely on memory. The key insight is that choice modulates mem- echo, with each string (trace) contributing in proportion to its match ory, altering the prioritization of retrieval to favor preference- and to the incoming sound (probe). The echo produced at retrieval is an behavior-relevant items. The authors of the third paper use machine imperfect representation of the target experience because traces from learning and a computational model of attention in decision mak- other experiences also contribute to the echo. Evaluation proceeds by ing to show that preferences are influenced by attention to different comparison of the target evaluative dimension to the corresponding degrees depending on whether choices are stimulus-based vs. men- features of the retrieved representation.

I tested the model against empirical data in simulations of the Most choices that we make on a day-to-day basis are memory- rating effect—the effect of explicitly evaluating an experience on based, with consideration sets and choice items being retrieved from retrospective evaluation of that experience (Ibrahim, Häubl, and memory at the time of decision. A large body of work has estab- Waisman, Working Paper). Ibrahim et al. had participants watch an lished the importance of memory for many different decisions (e.g., enjoyable time-lapse video and then either rate their enjoyment of Dick, Chakravarti, and Biehal 1990; Hertwig, Barron, Weber, and it (rating condition), briefly describe the first and last scenes (recall Erev 2004; Lynch, Marmorstein, and Weigold 1988; Murty, Feld- condition), or go on with their day (control condition). One, three, or manHall, Hunter, Phelps, and Davachi 2016). While previous work seven days later participants reported their enjoyment of the video. in the field of cognitive psychology has established regularities in After one day no difference in retrospective evaluation was apparent memory processes and highlighted the importance of value-based across conditions. After three days, participants in the rating condi- recall using free recall list-learning paradigms, memory processes in tion reported having enjoyed the video marginally more than those choice behavior have not been examined in a similar manner (Klein, in the control condition, with the recall condition in between. The Addis, and Kahana 2005; Castel 2007). Thus, we do not know the same pattern emerged as significant after seven days. To simulate degree to which memory processes at play in pure-memory tasks in- the rating effect in Minerva-EM the probes and echoes produced fluence memory-based choice, and if memory-based choice involves via initial rating or recall tasks were encoded as events. Three time a unique set of decision-specific memory mechanisms. intervals were simulated by adding 1000, 5000, or 10000 random The order in which people recall items in pure-memory tasks traces to memory, and 10000 participants were simulated per condi- is sensitive to primacy and recency effects (higher likelihood of re- tion. Consistent with the empirical results, the evaluative response membering words that appear at the beginning and end of lists) and (echo intensity) was greatest in the rating condition, with the recall temporal and semantic clustering (tendency to make recall transi- condition in the middle, and the difference increased as more events tions among temporally-close and semantically-related items) (Ka- intervened. In another experiment, Ibrahim et al. tested if previously hana and Miller 2013). Some of these effects have also been shown rating an experience increases enjoyment regardless of experience in memory-based choice. For example, brand memories display se- valence. Participants in rating and control conditions watched an un- rial position effects (e.g., Li 2010; Sherrick et al. 2016), and items enjoyable history video and seven days later reported how enjoyable recalled in everyday choice are semantically clustered (Hutchinson it was. Again, simulation results correspond well with the empirical 1994; Bhatia 2019). That said, thus far, researchers have examined data. Simulated and experimental participants evaluated the experi- pure-memory and memory-based choice in isolation, and have not ence as less enjoyable if they previously rated the video. systematically compared memory-based choice with pure-memory I derived novel hypotheses for future investigation from a para- in list-learning settings. We provide such a systematic comparison. metric analysis across 81 simulation conditions. For instance, the We first investigated the domain of food choice (Study 1). In model predicts that a request to rate a dining experience will influ- the subsequent study (Study 2), we extended our tests into the so- ence retrospective evaluation to a greater extent when the experience cial decision domain, namely gift choices. In both of our pre-regis- involves grabbing a snack in a familiar fast-food outlet compared to a tered studies, we adopted a list learning paradigm from the episodic more involved and lengthy experience such as eating dinner at a new memory literature. More specifically, we asked our participants to fine dining establishment. Correspondence of simulation and empiri- study a list of 25 items (food items or gift items, in Study 1 and 2, cal results indicates that the mechanisms described by Minerva-EM respectively), and manipulated the recall task across our participants. account for the rating effect and suggest the model captures basic We randomly asked half of our participants to recall as many items processes responsible for retrospective evaluation more generally. as they could from the just-presented list (pure-memory condition). This work makes three important contributions. First, the model The other half of the participants deliberated through the items and extends a classic theory of memory (Hintzman 1984) to offer a unify- recalled them during a decision making task with the ultimate goal ing framework for our understanding of the behavior of consumers of making a choice (choice condition). when they rely on memory to evaluate consumption experiences. By analyzing the particular sequence of presented and recalled Second, simulation in Minerva-EM can benefit marketers seeking items of each participant, we were able to identify memory processes to understand when soliciting consumer feedback is advantageous. that underlie pure-memory and choice behavior tasks and compare Third, I demonstrate that cognitive computational modeling offers a them based on the particular task participants were engaged in. Our useful and fruitful strategy for investigating and understanding con- results showed that traditional memory regularities such as the pri- sumer behavior. macy effect, as well as semantic and temporal clustering, appear in both pure-memory and decision making situations. That said, we What I Like Is What I Remember: Memory Modulation found a stronger influence of temporal clustering in pure-memory in Preferential Choice than in choice. Additionally, preferences and frequency of consumption have a bigger effect in choice, with participants being more like- EXTENDED ABSTRACT ly to recall items they like or items they have consumed previously. Imagine meeting your new next-door neighbor for an afternoon Computational models fit to participant data provided converging tea. First, try to remember all of the restaurants in your neighborhood support for our results. to give her a list of available options. Now try to remember these Our results suggest a substantial overlap in the memory pro- restaurants with the explicit goal of choosing where to go for dinner. cesses used in pure-memory settings and in choice. However, these It is clear that your choice depends fundamentally on memory: A res- processes are not identical. Choice involves the modulation of taurant cannot be selected unless it is successfully recalled. But what memory to prioritize the retrieval of decision-relevant items, such as is less clear is how the choice task (the goal of selecting somewhere items that are highly desirable. One result of this is that other mem- to eat) modulates memory. How do memory processes during pref- ory effects, such as temporal clustering, weaken. Ultimately, our erential decision making compare with memory processes that guide experiments showcase a novel experimental paradigm for studying recall when individuals do not have to make a choice? memory in decision making. This paradigm is able to connect decision making with established memory-theories in cognitive psychol- Advances in Consumer Research (Volume 48) / 1141 ogy. Future work should build on these connections, so as to better things they rated more highly, they did spend longer, on average, characterize the domain-general and domain-specific memory mech- looking at heavier foods and looking at food images that take up anisms at play in choice, and to better understand the ways in which more of the screen. memory can be intervened on to influence and improve choice. In order to identify the strongest connections (and thus, separa- tions) between the tasks, we use a combination of subject-level mod- Packaging vs . Product: Distinguishing Between Choice eling and machine learning. First, we draw on the attentional drift Domains With Computational Modeling diffusion model (Krajbich et al. 2010)—a computational model of decision making—to characterize the degree to which attention mod- EXTENDED ABSTRACT erates the choice process. We estimate subject-level effects in each When people shop, they are undoubtedly influenced by both the task and then examine inter-task correlations (or lack thereof). We products’ packaging and their mental representations of the products find the strongest positive correlations (1) between the two stimulus- inside the packaging. Therefore, it is important to understand how based tasks (i.e. image size and package preference) and (2) between people make decisions based on both types of information. Although the two representation-based tasks (i.e. taste and weight), while the choosing based on the packaging and choosing based on the prod- rest of the inter-task correlations were much smaller in magnitude (or uct inside the packaging are clearly different decisions, they may even negative). This provides some evidence that subjects’ stimulus- have similar underlying mechanisms. For instance, past research has based decision processes and mental representation-based decision uncovered a consistent effect of visual attention in many different processes are similar (regardless of whether the decision is objective types of choices; specifically, more attention to an option leads to or subjective in nature), specifically with regard to the link between an increased likelihood of choosing that option. This link applies to attention and choice. choices in many domains, including consumer goods (Krajbich et al. In a separate set of analyses, we combine ~30 machine learning 2012), monetary gambles (Stewart, Hermens, and Matthews 2015), classification methods to estimate inter-task (dis)similarity. These and even complex decision environments, like moral dilemmas (Pär- machine learning models utilize a wide variety of decision process namets et al. 2015). This attention-choice link also extends to per- measures (spanning choice, response time, and attention-based met- ceptual judgments, such as choosing the option that more closely rics) from the odd-numbered trials to learn patterns in the data. Then, resembles a target stimulus (Tavares, Perona, and Rangel 2017). the models use these learned patterns to try to accurately classify the However, past research (Polanía et al. 2014) has also demon- even-numbered trials into the appropriate category. We use pairwise strated, using neural (EEG) data, that there may be differences in classification accuracy rates as an index of dissimilarity, since the the decision-making process between preference-based choices (e.g. most dissimilar tasks should be the easiest to classify correctly. (On choosing which food to eat) and perceptual judgments (e.g. deciding the other hand, the most similar tasks should have the lowest classifi- which image is larger). However, in our initial example, choosing cation accuracy.) The machine learning models are most accurate at an item based on its packaging is both a perceptual- and preference- distinguishing between the taste and size tasks (in line with Polanía based task. Accordingly, we argue that the separation of perceptual et al. 2014), and the models are least accurate at distinguishing be- judgments and preference-based choices in the literature thus far has tween the representation-based tasks (i.e. taste and weight). Overall, been conflated with the separation of stimulus-based decisions (e.g. the machine learning results (in combination with the drift diffusion choosing which image is larger, based on the image itself) vs. de- approach described above) imply greater similarity in the decision cisions about the object represented by the stimulus (e.g. choosing process within the stimulus/representation categories than within the which food to eat, based on pictures representing the foods). There- subjective/objective categories. fore, it has not been possible to determine if the differences observed In this study, we investigate the separation of perceptual- and between perceptual- and preference-based choices are truly due to preference-based decisions suggested in the literature and demon- the perceptual/preference categorization or instead due to the stimu- strate that a more important distinction is the presentation of the lus/representation split. choice (i.e. whether the decision is about the stimulus itself or about Here, we present an alternative way to categorize decisions, a mental representation of the stimulus), rather than in the nature of based on two independent dimensions: (1) subjective vs. objective the choice (i.e. whether it is based on objective or subjective crite- and (2) stimulus vs. representation. In this project, we cross these ria). Therefore, there seems to be greater similarity in the decision dimensions and investigate the decision-making process in the four process among choices about which food is tastier or heavier (i.e. resulting categories, using the same stimulus set: food images. Spe- the representation-based tasks) than there is among choices based cifically, subjects (N = 42) first gave incentivized ratings for each of on taste or packaging (i.e. the preference-based tasks). Ultimately, the 100 food images in each of the four categories: taste (subjective; this study provides us with evidence that (1) although attention plays representation), weight (objective; representation), attractiveness of a large role across several different choice domains, (2) the choice the package (subjective; stimulus), and size of the image (objective; processes for products and packages are decidedly distinct. stimulus). Then, subjects made 100 incentivized, eye-tracked, binary choices in each of the four categories. Structural Dependencies in Computational Models of Overall, we find remarkable similarity between the tasks. In line Cognition in Consumer Research with past research, subjects chose in line with their subjective ratings (i.e. their preferences or perceptions, depending on the task). They EXTENDED ABSTRACT also tended to choose the option they looked at more; looking at One of the major purposes of cognitive modeling in consumer one option for a half-second longer translates to choosing the more research is measuring latent traits. This procedure is sometimes re- looked-at option ~70% of the time. Subjects also tended to choose ferred to as “cognitive psychometrics” (Batchelder and Riefer 1999; the option they looked at last; when subjects were indifferent be- Riefer, Knapp, Batchelder, Bamber, and Manifold 2002). Despite a tween the options, they chose the last looked-at option ~70% of the model’s fit and even despite a model’s predictive accuracy out-of- time. However, there were also some divisions between the tasks. sample, however, the model’s parameters might still not be correct For instance, although subjects did not spend more time looking at operationalizations of the underlying latent traits. Instead, the model 1142 / New Insights from Computational Models of Cognition in Consumer Research as-a-whole can be a very good proxy for the entire cognitive pro- reduced the parameter correlations, it greatly increased the model’s cess, while the individual parameter estimates do not represent the fit to the empirical data. suggested latent traits. An overconfidence in interpreting the param- Adding stochasticity to parameters has been proposed else- eters of cognitive models is often addressed with the famous quota- where as a better approach for risk (Blavatskyy and Pogrebna 2010) tion: “All models are wrong, but some are useful” (Box 1979). This as well as temporal discounting data (Apesteguia and Ballester idea emphasizes that cognitive models remain nonetheless statistical 2018). Our results dovetail with these findings. Following this line models, being proxies of cognitive processes. Researchers should be of reasoning, adding stochasticity to model parameters may be tested cautious when interpreting the model parameters, and even more so as a better-suited approach to cover the type of noise observed in if the necessary pre-condition to measuring latent traits is not ful- choice data, in general. filled—that is, the independent and reliable estimation of model pa- Two major suggestions that are relevant for the application of rameters. cognitive models in consumer research can be inferred from these Unfortunately, in several cognitive models that are commonly results. First, we suggest that the different model selection criteria used in consumer research, parameters are indeed structurally depen- should be weighted differently considering the purpose of the cogni- dent. One particular set of models that is commonly applied in con- tive model. Evaluating a model and comparing it with competitors sumer research is subject to very strong parameter intercorrelations. can lead to different results, depending on which model selection These models assume that behavior can be regarded as a probabilis- criterion is the focus. For instance, a model’s fit and out-of-sample tic choice among alternatives. The alternatives are assigned subjec- prediction are most important if the main purpose lies in predict- tive values. Free parameters in the model govern non-linear func- ing behavior across similar tasks. Independent parameter estimation, tions assigning the subjective values to the alternatives, and other parameter recovery and parameter correlations with neuronal or pro- parameters govern probabilistic choice functions, such as a logistic cess data are important if the researcher aims to measure the under- choice function to map the subjective values to the choices. This lying cognitive process or latent traits, and if the researcher aims to structure can be found in risky choice models, temporal discount- generalize predictions across tasks and domains. If the goal of the ing models or categorization models, to name but a few models that cognitive model lies in measuring latent traits that can subsequently are relevant to consumer research. In these models, parameters that be used for follow-up analyses, it must be guaranteed that the cor- determine the subjective value function also scale the input values, responding model parameters are estimated independently, in order while the response noise parameters in the choice function scale the to make valid inferences on the parameters. Second, we suggest subjective values to map them to the probability scale. As a conse- consumer researchers to conduct a parameter-recovery study. Every quence, the response noise parameter is dependent not only on the introduction of a new cognitive model to measure latent traits or in- scale introduced by the other parameter, it depends on the parameter dividual differences should be accompanied by a parameter-recovery itself that is governing the function. The dependency between the study to prove that parameter estimation is reliable. If the subsequent parameters leads to strongly correlated parameter estimates as for application of the cognitive model also involves the calculation of instance in Cumulative Prospect Theory, CPT (Tversky and Kahne- correlations of the model parameters with external variables or the man 1992), Generalized Context Model, GCM (Nosofsky 1986) and calculation of group difference on the parameters, the parameter- the SIMPLE model of free recall (Brown, Neath, and Chater 2007). recovery study should prove that group differences and correlations In order to illustrate the negative consequences of the correla- with external variables can be reliably recovered. tions, we implemented a generic choice model consisting of a power subjective value function with curvature parameter, and a logistic REFERENCES choice function with response noise parameter. By means of simula- Apesteguia, Jose and Miguel A. Ballester (2018), “Monotone tion and parameter-recovery studies, we show that the correlation stochastic choice models: The case of risk and time between parameters not only decreases the accuracy of parameter preferences,” Journal of Political Economy, 126 (1), 74–106. estimation, it systematically distorts parameter estimation. We illus- Babić Rosario, Ana, Francesca Sotgiu, Kristine De Valck, and trate that inference on the correlation of the parameters with external Tammo HA Bijmolt (2016), “The effect of electronic word of variables as well as inferences on group differences on the individual mouth on sales: A meta-analytic review of platform, product, parameters are distorted by the correlated parameters. and metric factors,” Journal of Marketing Research, 53 (3), We then make several proposals for implementation in the con- 297–318. cerned models to reduce the dependency of the parameters and thus Bartels, Daniel M. and Eric J. Johnson (2015), “Connecting the negative consequences resulting from the correlation. Those pro- cognition and consumer choice,” Cognition, 135, 47-51. posals either focus on applying different ways of rescaling the input Batchelder, William H. and David M. Riefer (1999), “Theoretical of the probabilistic choice function, to cut the dependency between and empirical review of multinomial process tree modeling,” parameters, or they replace the individual response noise parameter Psychonomic Bulletin & Review, 6 (1), 57–86. with a different type of individual error parameter. The latter ap- Bhatia, Sudeep (2019), “Semantic processes in preferential decision proach seemed to be most effective. We tested two alternative er- making,” Journal of Experimental Psychology: Learning, ror models: adding a trembling hand error while fixing the response Memory, and Cognition, 45 (4), 627. noise parameter to one and adding stochasticity to the correlated Bhatia, Sudeep and Timothy J. Pleskac (2019), “Preference model parameter while likewise fixing the response noise parameter accumulation as a process model of desirability ratings,” to one. Here, instead of assuming that every individual had one in- Cognitive Psychology, 109, 47-67. dividual parameter, the model assumes that every individual has a Bhatia, Sudeep and Timothy L. Mullett (2016), “The dynamics of distribution of parameters with an individual mean and an individual deferred decision,” Cognitive Psychology, 86, 112-51. variance. The variance of the parameter distribution determines the Blavatskyy, Pavlo R. and Ganna Pogrebna (2010), “Models individual stochasticity. In the simulation and parameter-recovery of stochastic choice and decision theories: why both are studies, we found that this proposal reduced the correlation and the important for analyzing decisions,” Journal of Applied negative consequences enormously. Implementing it in CPT not only Econometrics, 25 (6), 963–86. Advances in Consumer Research (Volume 48) / 1143

Box, George E. P. (1979), “Robustness in the strategy of scientific Lynch Jr, John G., Howard Marmorstein, and Michael F. Weigold model building,” Robustness in Statistics. ACADEMIC (1988), “Choices From Sets Including Remembered Brands: PRESS, INC. Use of Recalled Attributes and Prior Overall Evaluations,” Brown, Gordon DA, Ian Neath, and Nick Chater (2007) , “A Journal of Consumer Research, 15 (2), 169-84. temporal ratio model of memory,” Psychological Review, 114 Moore, Sarah G. (2011), “Some things are better left unsaid: (3), 539–76. How word of mouth influences the storyteller,” Journal of Castel, Alan D. (2007), “The adaptive and strategic use of memory Consumer Research, 38(6), 1140-54. by older adults: Evaluative processing and value-directed Morwitz, Vicki G., and Gavan J. Fitzsimons (2004), “The mere- remembering,” Psychology of Learning and Motivation, 48, measurement effect: Why does measuring intentions change 225–70. actual behavior?” Journal of Consumer Psychology, 14 (1–2), Dardis, Frank E., Mike Schmierbach, Brett Sherrick, Frank 64-74. Waddell, Jose Aviles, Sushma Kumble, and Erica Bailey Murty, Vishnu P., Oriel FeldmanHall, Lindsay E. Hunter, Elizabeth (2016), “Adver-Where? Comparing the Effectiveness of A. Phelps, and Lila Davachi (2016), “Episodic memories Banner Ads and Video Ads in Online Video Games,” Journal predict adaptive value-based decision-making,” Journal of of Interactive Advertising, 16 (2), 87–100. Experimental Psychology: General, 145 (5), 548. Dick, Alan, Dipankar Chakravarti, and Gabriel Biehal (1990), Nielson, A. C. (2015), “Global trust in advertising,” NY: USA, “Memory-based Inferences During Consumer Choice,” Nielsen Media Research, ACNielsen. Retrieved from http:// Journal of Consumer Research, 17 (1), 82-93. www.nielsen.com/us/en/insights/reports/2015/global-trust-in- Fredrickson, Barbara L. (2000), “Extracting meaning from past advertising-2015.html. affective experiences: The importance of peaks, ends, and Nosofsky, Robert M. (1986) , “Attention, Similarity, and the specific emotions,”Cognition and Emotion, 14 (4), 577–606. Identification-Categorization Relationship,” Journal of Heitmann, Mark, Donald R. Lehmann, and Andreas Herrmann Experimental Psychology: General, 115 (1), 39–57. (2007), “Choice goal attainment and decision and consumption Polanía, Rafael, Ian Krajbich, Marcus Grueschow, and Christian satisfaction,” Journal of Marketing Research, 44 (2), 234–50. C. Ruff (2014), “Neural oscillations and synchronization Hertwig, Ralph, Greg Barron, Elke U. Weber, and Ido Erev (2004), differentially support evidence accumulation in perceptual and “Decisions from experience and the effect of rare events in value-based decision making,” Neuron, 82 (3), 709–720. risky choice,” Psychological Science, 15, 534–39. Reichheld, Frederick F. (2003), “The one number you need to Hintzman, Douglas L. (1984), “MINERVA 2: A simulation model grow,” Harvard business review, 81 (12), 46–55. of human memory,” Behavior Research Methods, Instruments Riefer, David M., Bethany R. Knapp, William H. Batchelder, and Computers, 16 (2), 96–101. Donald Bamber, and Victor Manifold (2002), “Cognitive Hutchinson, J. Wesley, Kalyan Raman, and Murali K. Mantrala psychometrics: Assessing storage and retrieval deficits in (1994), “Finding choice alternatives in memory: Probability special populations with multinomial processing tree models,” models of brand name recall,” Journal of Marketing Research, Psychological Assessment, 14 (2), 184–201. 31 (4), 441–61. Stewart, Neil, Frouke Hermens, and William J. Matthews (2015), Ibrahim, Nahid, Gerald Häubl, and Rory M. Waisman (Working “Eye movements in risky choice,” Journal of Behavioral Paper), “Good gets better, bad gets worse: The polarizing Decision Making, 29(2-3), 116-136. effect of rating a consumption experience.” Tavares, Gabriela, Pietro Perona, and Antonio Rangel (2017), “The Kahana, Michael J., and Jonathan F. Miller (2013), “Memory, recall attentional drift diffusion model of simple perceptual decision- dynamics,” Encyclopedia of the Mind. Thousand Oaks, CA making,” Frontiers in Neuroscience, 11, 468. SAGE Publications Inc. Trueblood, Jennifer S., Scott D. Brown, and Andrew Heathcote Keiningham, Timothy L., Bruce Cooil, Tor Wallin Andreassen, (2014), “The multiattribute linear ballistic accumulator model and Lerzan Aksoy (2007), “A longitudinal examination of net of context effects in multialternative choice,”Psychological promoter and firm revenue growth,” Journal of Marketing, 71 Review, 121 (2), 179. (3), 39–51. Trueblood, Jennifer S., Scott D. Brown, Andrew Heathcote, and Klein, Krystal A., Kelly M. Addis, and Michael J. Kahana (2005), Jerome R. Busemeyer (2013), “Not just for consumers: “A comparative analysis of serial and free recall,” Memory and Context effects are fundamental to decision making,” Cognition, 33 (5), 833–39. Psychological Science, 24 (6), 901-908. Krajbich, Ian, Carrie Armel, and Antonio Rangel (2010), “Visual Tversky, Amos and Daniel Kahneman (1992) , “Advances in fixations and the computation and comparison of value in Prospect-Theory - Cumulative Representation of Uncertainty,” simple choice,” Nature Neuroscience, 13 (10), 1292–98. Journal of Risk and Uncertainty, 5(4), 297–323. Krajbich, Ian, Dingchao Lu, Colin Camerer, and Antonio Rangel Vinson, David W., Rick Dale, and Michael N. Jones (2019), (2012), “The attentional drift-diffusion model extends to “Decision contamination in the wild: Sequential dependencies simple purchasing decisions,” Frontiers in Psychology, 3. in online review ratings,” Behavior research Methods, 51 (4), Pärnamets, P., Johansson, P., Hall, L., Balkenius, C., Spivey, 1477-1484. M. J., and Richardson, D. C. (2015). Biasing moral decisions Weber, Elke U., and Eric J. Johnson (2006), “Constructing by exploiting the dynamics of eye gaze. Proceedings of the preferences from memory,” In: Lichtenstein, S., and Slovic, P. National Academy of Sciences, 112 (13), 4170–75. (Eds.). The Construction of Preference. Cambridge University Li, Cong (2010), “Primacy effect or recency effect? A long‐term Press. memory test of Super Bowl commercials,” Journal of Consumer Behaviour: An International Research Review, 9 (1), 32–44. The Social, Psychological and Behavioral Consequences of Being Poor Chair: Yuechen Wu, Johns Hopkins University, USA

Paper #1: Resource Sharing in the Sharing Economy: Low should do both an overnight work-shift and sleep), and, when the Childhood Socioeconomic Status as a Barrier poor fail to meet these impossible expectations, they judge them as Yuechen Wu, Johns Hopkins University, USA lazier and less competent than the rich. Meng Zhu, Johns Hopkins University, USA Taken together, this session integrates new perspectives regarding the psychological, behavioral and social consequences of being Paper #2: Effects of Childhood Socioeconomic Status on Product poor. We believe that this session will appeal to a broad audience, Retention and Disposal Behavior in Adulthood including researchers interested in resource scarcity, socioeconomic Lisa Park, Georgetown University, USA status, poverty, the sharing economy, prosocial and ethical behaviors, Rebecca Hamilton, Georgetown University, USA and person perception. Paper #3: Are Rich/Educated Consumers less Ethical and Prosocial? Two Direct, Preregistered Replications of Piff et al.’s Resource Sharing in the Sharing Economy: Low (2012) Field Studies Childhood Socioeconomic Status as a Barrier Minah Jung, New York University, USA Joachim Vosgerau, Bocconi University, Italy EXTENDED ABSTRACT Paul Smeets, Maastricht University, The Netherlands In the past decade, the sharing economy has blossomed across Jan Stoop, Erasmus University, The Netherlands a range of domains such as lodging (e.g., Airbnb, HomeAway) and Paper #4: Impossible Expectations for the Poor transportation (e.g., Lyft, Uber, Turo). While extant literature has Anuj K. Shah, University of Chicago, USA systematically examined consumers’ motivation to join the sharing economy on the demand side as users (Belk 2013; Lamberton and SESSION OVERIVEW Rose 2012), it remains understudied what factors might impact con- Although abundance has emerged as the norm in modern in- sumers’ willingness to share their own assets on the supply side. dustrialized societies, resource distribution inequality and scarcity In this paper, we examine one fundamental factor that could remains prevalent across the globe (Roux, Goldsmith and Bonezzi impact consumers’ tendency of providing their own assets in the 2015). According to the World Bank, in 2015 about 10 percent of the sharing economy: socioeconomic status (SES). Somewhat intuitive- world population lived on less than $1.90 per day. Even within devel- ly, current SES might positively impact a person’s resource sharing oped economies like the United States, approximately 38.1 million behavior in the sharing-economy, as higher current SES consumers people lived in poverty in 2018, and more than 25% of children (20.1 have more assets (e.g., rooms, cars, spare tools) that can be shared million) live in households where no parent has secure employment with others in the first place. However, it is unclear whether and how (Kids Count Data Book 2019). Despite the accumulating interest in the socioeconomic environment one grew up in might exert an inde- topics regarding poverty and scarcity, many questions still need to be pendent impact on consumers’ willingness to share their resources answered regarding the consequences of being poor. For example, in the sharing economy. Although current SES and childhood SES how are consumption behaviors shaped by the experience of being are correlated, a rich body of research in psychology and sociology poor? Does low socioeconomic status make a consumer more self- has suggested that SES exposures during childhood, independent of ish? Are there social penalties to being poor? adulthood, are predictive of individuals’ preferences, decisions and This session sheds light on these questions; four papers exam- behaviors in the adulthood, including reproductive timing, eating ine the behavioral, psychological, and social consequences of being habits, and health outcomes (Amir et al. 2016; Cohen et al. 2010; poor. Wu and Zhu investigate whether socioeconomic status influ- Hill et al. 2016). In the present paper, we seek to isolate the effects ences consumers’ participation in the sharing economy as service of childhood SES on consumers’ resource sharing decisions from the providers. They find that consumers growing up in a lower socio- possible impact of current SES and resource availability. economic environment are less likely to share their own resources We propose (and find) that lower childhood SES decreases con- the sharing economy, controlling for current SES. Further, they show sumers’ willingness to share their own assets in the sharing economy. that greater territorial feelings over one’s own possessions serve as a In particular, drawing from prior research characterizing different so- central mechanism driving this effect. Consistent with these territori- cioeconomic environments (Amir et al. 2016; Evans and Kim 2012; al feelings, Park and Hamilton show that childhood SES influences Evans and Cassells 2013; Kraus et al. 2011; Stamos, Altsitsiadis, and consumer decision making related to product retention and dispos- Dewitte 2019) and past work on territoriality (Brown, Lawrence, and al. Controlling for current SES, low childhood SES consumers are Robinson 2005; Kirk, Peck, and Swain 2017), we argue that the psy- found to experience more negative emotions during disposal, report chological costs of sharing ownership can be higher for consumers greater desire to retain products, and dispose of products less fre- growing up in a lower socioeconomic environment, as these consum- quently than consumers with more abundant resources growing up. ers might have developed greater territorial feelings towards their The third and fourth papers examine social interactions and in- own property. As a result of the greater territoriality, lower childhood ferences. Jung, Vosgerau, Smeets, and Stoop investigate whether SES consumers might be less likely to share their own assets in the socioeconomic status predicts ethical behaviors. Contrary to prior sharing economy, an effect that goes above and beyond the effect of findings suggesting that the poor behave more ethically compared individuals’ current socioeconomic status. to the rich (e.g., are less likely to cut-off others at intersections and We conducted three preregistered studies as well as used a cross-walks), they find no relationship between socioeconomic status national-level field dataset to test the proposed effect and the under- and consumers’ ethical behaviors. The last paper examines the social lying mechanism. Study 1 (N = 204) tested the effect in a peer-to- consequences of being poor. Shah finds that consumers can have peer car-sharing domain using a real-behavioral dependent variable. inconsistent and even impossible expectations of the poor (e.g., they Participants first described a car that they currently have, and read

Advances in Consumer Research 1144 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1145 about a short description about Turo, a car-sharing platform. We then versus resource scarce environment (Griskevicius et al. 2011). In our asked participants to choose between reading two topics about Turo: research, we explore the lasting effects of childhood SES on consum- an article about how to list cars on Turo, and a recent news article er retention and disposal habits. More specifically, when controlling about Turo. Participants then indicated their childhood SES (“When for current SES, we predict that low childhood SES consumers will you were growing up, what was your assessment of your family’s experience more negative emotion during disposal, report greater de- social and economic status?”; Amir, Jordan and Rand 2018) and cur- sire to retain products, and dispose of products less frequently than rent SES (e.g., “I have enough money to buy things I want”; Hill et high childhood SES consumers. al. 2016) As predicted, controlling for current SES, individuals with Because childhood SES adjusts individuals’ expectations and lower childhood SES were significantly less likely to read the article patterns of responses, it may be more predictive of adult behavior about listing cars on Turo, implying decreased interests of resources than current SES (Chen and Miller 2012; Griskevicius et al. 2011; sharing. Griskevicius and Kenrick 2013), especially combined with the im- Study 2 (N = 201) explored the underlying mechanism. Partici- pact of parental influence. In particular, early life conditions may pants read a scenario where they imagined that they had a spare room program individuals to respond to future adversity in different ways. in their house, indicated their likelihood of listing their spare room Although adults with different SES backgrounds may behave simi- on Airbnb, and explained in an open-ended question, why they made larly in benign situations, they may behave very differently when the decision. We coded the open-ended responses (Khan et al., 2011, facing threatening conditions. For example, in response to a threat, Zhu et al. 2018) and found that people with lower SES growing up individuals of low childhood SES were found to be more impulsive, had stronger territorial feelings over their possessions, which medi- risk-taking, and suggestible to temptation compared to their high ated the effect of childhood SES on participating in sharing economy. childhood SES counterparts (Griskevicius et al. 2011). Study 3 (N = 594) provided further evidence for the territorial- We propose that low childhood SES consumers may experi- feeling mechanism by directly manipulating participants’ level of ence more negative emotion during disposal and greater reluctance ownership toward the object. Participants described an object in to dispose of products than high childhood SES consumers due to their house that they purchased (control condition) or other people early life experiences that encourage them to focus on self-preser- purchased (low ownership condition) and indicated their likelihood vation and resource preservation. In a sense, disposal of products of listing this object on a community resource-sharing platform. We contradicts the life history strategy they adopted as resource-scarce found that in the control condition, lower childhood SES participants children. Notably, childhood SES may have distinct effects across were less likely to list their own possessions on the sharing-economy stages of the consumer decision journey (Hamilton et al. 2019). For platforms, replicating prior findings. However, in the low ownership example, while low childhood SES seems to predict greater impul- condition, the effect of childhood SES on likelihood to share own siveness during initial choice (Griskevicius et al. 2011), it predicts assets in the sharing economy was attenuated. greater patience after consumers make an initial choice and learn that Study 4 (N = 57,155) employed a publicly available nation-lev- their choice is not immediately available (Thompson, Hamilton and el field dataset: China Family Panel Studies (2010). Our primary DV Banerji 2020). Although there is some evidence that low socioeco- was whether participants had ever rented out a spare room in their nomic status consumers tend to repurpose consumption items (see house. We used maternal education attainment as a proxy for early- Hill 2001; Rosa et al. 2012), there has not been a systematic exami- life socioeconomic stability (Duncan and Brooks-Gunn 1997) and nation of the effects of childhood SES on consumers’ disposal and coded individuals as high (vs. low) childhood SES if their mother retention habits. Thus, we used a multi-method approach to examine had completed some formal education (vs. no formal education). We the effects of childhood SES on consumers’ disposal and retention used propensity score matching method to control for confounding habits and emotions experienced during disposal. variables including the mother’s year of birth, household current Our first step was to conduct 22 interviews with consumers to SES, house space, family size, participants’ number of siblings, age, understand their disposal and retention habits, potential relationships gender, marital status, education and job types. Replicating prior between these habits, their current and childhood socioeconomic sta- studies, individuals whose socioeconomic stability was high (vs. tus and other demographic variables. We leveraged peer networks low) in their early life were more likely to rent out their spare room. to recruit participants who varied in gender (10 male, 12 female), Further, as a robustness check, we regressed participants’ room- age (21-78), ethnicity (13 Caucasion, 7 Asian, 2 Hispanic), and both renting behaviors on childhood SES index (i.e., parents’ completed childhood and current SES. Overall, participants were more likely to schooling and occupational prestige), controlling for the covariate self-categorize as “savers” than as “tossers,” perhaps due to the social variables, and found similar results. desirability of avoiding wastefulness. However, their descriptions of To conclude, the current research reveals that lower childhood when and why they preferred to save items rather than dispose of SES can lead to decreased likelihood of sharing own resources the them diverged substantially. Eight prominent themes emerged after sharing economy, an effect for which greater territorial feelings over analyzing all of the interviews: concerns about price paid for items, one’s own possessions emerge as a central driver. emotional attachment to items, clutter/space concerns, desire to repurpose items, concern over having to replace products, childhood Effects of Childhood Socioeconomic Status on Product experiences, and the impact of parents on retention and disposal. Retention and Disposal Behavior in Adulthood These interviews revealed that childhood experiences have a strong impact on post-purchasing habits, primarily through parental EXTENDED ABSTRACT influence. Many interviewees admitted to emulating their parents’ Earlier research has demonstrated that socioeconomic status habits in adulthood. For example, those who observed frugal par- (SES) in childhood has enduring effects on consumers’ attitudes ents were more likely to be conscious of their disposal and reten- towards products (Chaplin et al. 2014), their attitudes towards risk tion habits than those who did not. Moreover, changes in childhood (Griskevicius et al. 2013), and their willingness to wait for products SES seemed to have a particularly strong effect. Participants who (Thompson, Hamilton and Banerji 2020). Childhood SES reflects experienced financial shocks (e.g., immigrating to a new country, the degree to which an individual grew up in a resource abundant bankruptcy) during their childhood often experienced more nega- 1146 / The Social, Psychological and Behavioral Consequences of Being Poor tive emotions when disposing of products in adulthood than those The Piff et al. (2012) paper, and particularly these two field who did not experience financial shocks. Changes in disposal and studies, have been discussed extensively in the popular media (e.g., retention habits also coincided with changes in SES. Ultimately, one New York Times 2010, 2013, 2015, 2017; The Economist 2010; of the most critical determinants of retention and disposal habits Business Insider 2012, 2015; BBC 2015; CNN 2018; NPR 2018; in adulthood seemed to be how a consumer remembered his or her The Washington Post 2018) and in the academic literature (at the childhood SES rather than more objective measures of childhood in- time of writing this proposal, the paper had over 825 Google Scholar come or parental education. citations). Next, we conducted confirmatory research to measure the re- Research in economics and sociology, in contrast, has hypoth- lationship between childhood SES and retention and disposal habits esized and found the opposite, a positive relationship between SES in adulthood. Specifically, we developed items to capture each of and prosocial/ethical behavior. From an economic point of view, so- the eight themes identified in our exploratory research, disposal and cial preferences such as caring about others’ wellbeing and valuing retention habits, and emotions experienced during disposal. We also ethical principles implicate substantial opportunity costs, costs that included measures of both objective and subjective childhood SES only those with sufficient resources can afford to bear (Trautmann and current SES. Two hundred twenty-eight consumers of varying et al. 2013). In line with this reasoning, individuals living in richer ages, ethnicities, education and income levels were recruited to par- areas have been shown to be more likely than those from poorer ticipate via Amazon MTurk. The results point to a significant rela- areas to send back wrongly addressed letters containing cash (An- tionship between childhood SES and retention and disposal habits dreoni, Nikiforakis, & Stoop 2017). Korndörfer et al. (2015), using in adulthood. Consumers with higher childhood SES reported more the German Socio-Economic Panel (SOEP), the GSS, the American frequent disposal and different reasons for disposing of products. Consumer Expenditure Survey (CEX), and the International So- Specifically, high childhood SES consumers were more likely to cial Survey Programme (ISSP), showed that―across 30 countries dispose of products based on how they looked and dispose of unfin- including the US―high SES individuals are more likely to make ished products. The data also revealed that positive emotions when charitable donations and to contribute a higher percentage of their disposing, such as pride, strength and enthusiasm, were more closely family income to charity, to be more helpful and likely to volunteer, correlated to high childhood SES than current SES. In contrast, low and to be more trusting and trustworthy in an economic game when childhood SES consumers were more likely to report retaining items interacting with a stranger than low SES individuals. because they had paid a lot for them or holding onto them based on We conducted two direct, highly-powered, and pre-registered concern about needing them in the future. replications of the two field studies in Piff et al. (2012). In summary, both our exploratory and confirmatory research Study 1 is a direct replication of field study 1 in Piff et al. (2012). suggest that childhood SES has a lasting effect on consumer post- We observed the behavior of drivers at the same four-way intersec- purchase behavior in adulthood. Consumers who reported lower sub- tion in Berkeley, California, as in the original study. We aimed at jective childhood SES are more reluctant to dispose of products than collecting 2.5 times the original’s study’s sample size (Noriginal = 274, those who reported higher subjective childhood SES. Notably, con- Nreplication= 685). With a correlation of r = .12 between vehicle status sumers’ memory of financial hardships during childhood seems to be and likelihood of cutting off as observed in the original study, the a more important predictor than the accuracy of recall. Building on replication had 88% power to detect the effect at the 5% level (two- prior research linking childhood SES to adult choice and consump- sided test). The training of the coders (research assistants), coding of tion behavior, our research suggests that experiences in the formative vehicle status (1 = cars worth $3,000 or less; 2 = $3,001-$10,000; 3 years of childhood have a particularly strong effect on adult habits = $10,001-$25,000; 4 = $25,001-$40,000; 5 = cars worth more than related to retention and disposal of products. $40,000), age and gender of drivers, and position of the coders were the same as in the original study. Unlike in Piff et al.’s (2012) study Are Rich/Educated Consumers less Ethical and 1, vehicle status and cutting-off as judged by the first team of coders Prosocial? Two Direct, Preregistered Replications of Piff were not correlated, r = .014, p = .71, nor were vehicle status and et al .’s (2012) Field Studies cutting-off as judged by the second team of coders, r = -.024, p = .53. Binary logistic regressions on the likelihood of cutting-off other EXTENDED ABSTRACT vehicles, controlling for traffic density, time of day, and driver’s age Who behaves more miserly and unethically, the poor or the and sex, neither showed a relationship between vehicle status and rich? Some psychologists maintain that it is the rich, or more specifi- likelihood of cutting-off ps( > 0.60). cally, citizens of high socioeconomic status (SES). It is argued that Study 2 is a direct replication of field study 2 in Piff et al. (2012) people of high SES are more independent and less socially attuned at a crosswalk in Berkeley, California. Research assistants blind to than the lower social classes, and thus feel less compassion towards the study hypothesis observed and coded the likelihood of drivers others (Stellar et al. 2012). As a consequence, they are hypothesized not yielding right of way to pedestrians at the crosswalk. The cross- to be more likely to engage in antisocial and unethical behaviors walk on Bancroft Way at the intersection with Ellsworth in Berkeley (Piff et al. 2010; Kraus et al. 2012). Consistent with this hypothesis, was chosen rather than the crosswalk at the intersection of Bancroft it has been shown that high SES individuals are less likely to give to Way and Dana Street used in the original study, because Paul Piff charity and to help and trust others (Piff et al. 2010), and to be more informed us that the original crosswalk now has pedestrian lights likely to engage in selfish and unethical behaviors, such as taking installed and is hence no longer suited for the study. We aimed at candies from children, lying in job negotiations, cheating, pilfering, collecting 2.5 times the sample size of the original study (N = 152). and outright stealing (Piff et al. 2010; Piff et a. 2012; Côté et al. The final dataset included 439 observations. Due to a data encoding 2013; Guinote et al. 2015; Dubois et al. 2015). Piff et al. (2012) con- error we lost 75 observations, leaving us with 364 complete records, ducted two field studies, which show that ― compared to drivers of 16 short of the 380 that we had pre-registered. With a correlation of less luxurious cars ― drivers of luxurious vehicles are more likely to r = .17 between vehicle status and likelihood of not yielding right cut off others at intersections and are less likely to yield right of way of way to pedestrians as reported in Piff et al. (2012), study 2 had to pedestrians at crosswalks. 91% power to detect the effect at the 5% level. Unlike in Piff et al. Advances in Consumer Research (Volume 48) / 1147

(2012), vehicle status and likelihood of not yielding right of way to day across different tasks, participants’ answers sum to more than 24 pedestrians did not correlate, r = .004, p = .93. Also, binary logistic hours for the poor, not so for the rich. regressions controlling for traffic, time of day, and driver’s age and Studies 5A-D show that the poor are judged as lazier and less sex showed no relationship between vehicle status and likelihood of competent for failing to meet these (impossible) expectations; the not yielding to pedestrians (lowest p = .59). rich are not. For example, in Study 5D participants read a scenario describing how a (poor/rich) person could participate in one of two Impossible Expectations for the Poor gig-economy promotions that offered higher wages than usual. One of the promotions was a “driving” promotion (e.g., driving for Uber); EXTENDED ABSTRACT the other was a “delivery” promotion (e.g., deliveries for Uber Eats). People often believe poor individuals should do various things These two promotions were mutually exclusive. Half of the partici- to improve their lives (e.g., work longer hours, sell certain posses- pants were asked whether they thought the person should do the driv- sions). It is possible that these are just matters of opinion (about ing promotion. The other half were asked whether they thought the which people can agree or disagree). This paper asks whether these person should do the delivery promotion. Overall, significantly more beliefs are formed in a way that might lead to inconsistent or im- participants thought the poor should do the promotion they were possible expectations for the poor (but not the rich). That is, some asked about (86%) than should the rich (63%). Critically, expecta- of these beliefs might not just be things that people can agree or tions for the poor were essentially impossible—88% of participants disagree about, but are actually impossible and violate normative said the poor should do the delivery promotion, 85% said they should standards. do the driving promotion (both significantly greater than 50%). Why does this occur? When evaluating whether a person should The next phase of the study asked participants to make an attri- do something to make money, people focus narrowly on the oppor- bution about the person they just read about. Participants were either tunity without considering other things a person might do. There are assigned to the “control” condition or the “failure” condition. Par- many intuitive reasons why the poor should take the opportunity ticipants in the control condition were asked how lazy they think the (they need money). There are fewer intuitive reasons why the rich person is (on a scale from 1 to 7). Participants in the failure condition should (they have enough money). So when people consider whether were asked to first imagine that the person does not do the promo- the poor should do something to make money, the intuitive answer tion that the participants were just asked about. Then participants is “Yes.” were asked how lazy they think the person is. There was a signifi- Here, I document three effects that stem from this. First, this cant interaction such that participants in the failure condition rated intuitive belief means that people’s expectations for the poor depend the poor as significantly lazier (M=3.47) than in the control con- heavily on how a question is framed; their expectations are therefore dition (M=2.36). But participants did not rate the rich significantly often inconsistent. Moreover, people often argue that the poor should differently in the failure condition (M=2.86) and control condition do mutually exclusive things (i.e., do the impossible). Second, when (M=2.68). Taken together, people have impossible expectations for the poor fail to meet these inconsistent or impossible expectations the poor (more so than for the rich) and people judge the poor more (which is inevitable), they rate the poor are lazier or less competent. harshly (but not the rich) when they fail to meet these expectations. Third, this leads people to set inconsistent or impossible conditions Studies 6A-B show that this leads people to set inconsistent and for poverty alleviation policies. I find evidence for this across over impossible conditions for anti-poverty policies. For example, when 10 studies with ~100 participants per cell (e.g., 400 Ps for a 2x2 be- participants are asked how long the poor should have to work in tween-subjects study) in all studies. Below, I highlight a few studies. order to qualify for food stamps, the median response is age 50. But Study1 (2x2 between-subjects): Participants read about a rich when asked whether the poor should have to work past age 51 to or poor person, and participants were assigned to Phase1 or Phase2. qualify for food stamps, 64% of people say yes. This leads to escalat- Phase1 participants were asked how many hours the (rich or poor) ing requirements for the poor to receive benefits. person should work for a given wage. Median response for the poor: 10 hours; rich: 5 hours. Phase2 participants were asked a narrower REFERENCES version of this question—whether the person should work more than Amir, Dorsa, Matthew R. Jordan, and Richard G. Bribiescas the median response from Phase1. Normatively, Phase2 answers (2016), “A Longitudinal Assessment of Associations Between should be around 50% (because people are asked whether someone Adolescent Environment, Adversity Perception, and Economic should work more than the median number of hours). But, 67% of Status On Fertility and Age of Menarche,” PLoS One, 11, Phase2 participants said the poor should work more than the Phase1 e0155883. median; 46% said so for the rich. Depending on the question, people Andreoni, James, Nikos Nikiforakis, and Jan Stoop (2017), “Are have inconsistent expectations for how much the poor should work. The Rich More Selfish Than The Poor, Or Do They Just Have Study2 replicates this within-subjects. More Money? A Natural Field Experiment,” National Bureau Study3 (2x2 between-subjects) shows that this process leads to of Economic Research, No. w23229. impossible expectations. Participants read that a (poor/rich) student BBC. Are drivers of luxury cars less likely to stop for pedestrians? could either do an overnight work-shift or sleep before a scholar- [video file] (2015, March 13). Retrieved from http://www.bbc. ship exam; options were mutually exclusive. Participants were asked com/news/av/magazine-31854075/are-drivers-of-luxury-cars- one of two questions: Whether the student should work, or Whether less-likely-to-stop-for-pedestrians. the student should sleep. When asked whether the rich should sleep, Belk, Russell W. (2013). “Extended Self in A Digital World,” 81% said yes. Asked whether the rich should work, 35% said yes. Journal of Consumer Research, 40, 477–500. Asked whether the poor should sleep, 78% said yes. Asked wheth- Business Insider (2012), retrieved on March 03, 2018, from http:// er the poor should work, 66% said yes. A majority of people think www.businessinsider.com/how-rich-people-think-differently- the poor should both work and sleep, which is impossible. Study4 from-the-poor-2012-8?IR=T. shows that when asked how people should allocate their hours in a 1148 / The Social, Psychological and Behavioral Consequences of Being Poor

Business Insider (2015), retrieved on March 03, 2018 from http:// Hill, Sarah E., Marjorie L. Prokosch, Danielle J. DelPriore, Vladas www.businessinsider.com/rich-people-teach-their-kids-to-be- Griskevicius, and Andrew Kramer (2016), “Low Childhood rich-2015-9?IR=T. Socioeconomic Status Promotes Eating In The Absence Of Brown, G., Lawrence, T. B., & Robinson, S. L. (2005). Energy Need,” Psychological Science, 27, 354–364. Territoriality in organizations. Academy of Management Hill, R. P. (2001). Surviving in a Material World: The Lived Review, 30(3), 577-594. Experience of People in Poverty. Notre Dame: University of Chaplin, L. N., Hill, R. P., and John, D. R. (2014), “Poverty Notre Dame Press. and Materialism: A Look at Impoverished versus Affluent Khan, U., M. Zhu & A. Kalra (2011),“When Tradeoffs Matter: The Children,” Journal of Public Policy & Marketing, 33, 78–92. Effect of Choice Construal on Context Effects,”Journal of Chen, E., and Miller, G. E. (2012), “’Shift-and-persist’ Strategies: Marketing Research, 48(1), 62-71. Why Low Socioeconomic Status Isn’t Always Bad for Kirk, C. P., Peck, J., & Swain, S. D. (2018). Property lines in Health,” Perspectives on Psychological Science, 7(2), the mind: Consumers’ psychological ownership and their 135–158. territorial responses. Journal of Consumer Research, 45(1), CNN (2018), retrieved on March 03, 2018 from https://edition.cnn. 148-168. com/2017/12/18/health/poor-rich-happiness-study/index.html. Korndörfer, Martin, Boris Egloff, and Stefan C. Schmukle (2015). Cohen, S., Janicki‐Deverts, D., Chen, E., & Matthews, K. A. “A large scale test of the effect of social class on prosocial (2010). Childhood socioeconomic status and adult health. behavior.” PloS one, 10(7), e0133193. Annals of the New York Academy of Sciences, 1186(1), 37-55. Kraus, Michael W., Paul K. Piff, Rodolfo Mendoza-Denton, Côté, Stéphane, Julian House, and Robb Willer (2015), “High Michelle L. Rheinschmidt, and Dacher Keltner (2012). “Social Economic Inequality Leads Higher-Income Individuals To class, solipsism, and contextualism: how the rich are different Be Less Generous,” Proceedings of the National Academy of from the poor.” Psychological Review, 119(3), 546-572. Sciences, 112(52), 15838-15843. Kraus, Michael W., E. J. Horberg, Jennifer L. Goetz, and Dacher Cutti, S. (2018, February). More money can mean Scrooge-like Keltner (2011), “Social Class Rank, Threat Vigilance, and pride, study says. CNN. Retrieved from https://edition.cnn. Hostile Reactivity,” Personality and Social Psychology com/2017/12/18/health/poor-rich-happiness-study/index.html. Bulletin, 37, 1376–1388. Dubois, David, Derek D. Rucker, and Adam D. Galinsky (2015). Lamberton, Cait Poynor, and Randall L. Rose. (2012), “When Is “Social Class, Power, And Selfishness: When And Why Upper Ours Better Than Mine? A Framework For Understanding And Lower Class Individuals Behave Unethically,” Journal of And Altering Participation In Commercial Sharing Systems,” Personality and Social Psychology, 108(3), 436-449. Journal of Marketing, 76, 109–125. Duncan, Greg J., and Jeanne Brooks-Gunn (Eds.) (1997). National Public Radio (NPR 2018), retrieved on March 03, 2018 Consequences of growing up poor. New York, NY: Russell from https://www.npr.org/2018/01/23/579976635/the-power- Sage. hour. Economist (2010). Retrieved on March 03, 2018 from https://www. New York Times (2010), retrieved on March 03, 2018 from http:// economist.com/node/16690659. www.nytimes.com/2010/08/22/magazine/22FOB-wwln-t. Evans, Gary W., and Rochelle C. Cassells (2013), “Childhood html?_r=1. Poverty, Cumulative Risk Exposure, and Mental Health in New York Times (2013), retrieved on March 03, 2018 from Emerging Adults,” Clinical Psychological Science, 2, 287-296. https://wheels.blogs.nytimes.com/2013/08/12/the-rich-drive- Evans, Gary W. and Pilyoung Kim (2012), “Childhood Poverty differently-a-study-suggests/. and Young Adults’ Allostatic Load: The Mediating Role New York Times (2015), retrieved on March 03, 2018 from Of Childhood Cumulative Risk Exposure,” Psychological https://www.nytimes.com/2015/05/07/style/the-shaky-moral- Science, 23, 979-983. compass-of-silicon-valley.html. Griskevicius, V. and Kenrick, D. T. (2013), “Fundamental Motives: New York Times (2017), retrieved on March 03, 2018 from https:// How Evolutionary Needs Influence Consumer Behavior,” www.nytimes.com/2017/11/07/your-money/money-bubble- Journal of Consumer Psychology, 23(3), 372–386. economic-inequality.html Griskevicius, V., Tybur, J. M., Delton, A. W., & Robertson, T. Piff, Paul K., Michael W. Kraus, Stéphane Côté, Bonnie Hayden E. (2011), “The Influence of Mortality and Socioeconomic Cheng, and Dacher Keltner. (2010). “Having less, giving Status on Risk and Delayed Rewards: A Life History Theory more: the influence of social class on prosocial behavior.” Approach,” Journal of Personality and Social Psychology, Journal of Personality and Social Psychology, 99(5), 771-784. 100(6), 1015–1026. Piff, Paul K., Daniel M. Stancato, Stéphane Côté, Rodolfo Guinote, Ana, Ioanna Cotzia, Sanpreet Sandhu, and Pramila Siwa Mendoza-Denton, and Dacher Keltner (2012). “Higher social (2015). “Social status modulates prosocial behavior and class predicts increased unethical behavior.” Proceedings of egalitarianism in preschool children and adults.” Proceedings the National Academy of Sciences, 109(11), 4086-4091. of the National Academy of Sciences, 112(3), 731-736. Rosa, J. A., Geiger-Oneto, S., and Fajardo, A. B. (2012), “Hope Hamilton, Rebecca W., Debora Thompson, Sterling Bone, Lan N. and Innovativeness: Transformative Factors for Subsistence Chaplin, Vladas Griskevicius, Kelly Goldsmith, Ronald Hill, Consumer-merchants,” In Transformative Consumer Research Deborah Roedder John, Chiraag Mittal, Thomas O’Guinn, for Personal and Collective Well-Being, eds. David, Glen M., Paul Piff, Caroline Roux, Anuj Shah and Meng Zhu (2019), Simone Pettigrew, Cornelia Pechmann and Julie L. Ozanne, “The Effects of Scarcity on Consumer Decision Journeys,” New York, NY: Routledge, 151–70. Journal of the Academy of Marketing Science, 47 (3), 532-550. Roux, Caroline, Kelly Goldsmith, and Andrea Bonezzi (2015), “On The Psychology Of Scarcity: When Reminders Of Resource Scarcity Promote Selfish (And Generous) Behavior,” Journal of Consumer Research, 42(4), 615-631. Advances in Consumer Research (Volume 48) / 1149

Stellar, Jennifer E., Vida M. Manzo, Michael W. Kraus, and Dacher Thompson, Debora V., Rebecca W. Hamilton and Ishani Banerji Keltner (2012). “Class and compassion: Socioeconomic (2020), “The Effect of Childhood Socioeconomic Status on factors predict responses to suffering.”Emotion, 12(3), 449- Patience,” Organizational Behavior and Human Decision 460. Processes, 157 (March), 85-102. Stamos, Angelos, Efthymios Altsitsiadis, and Siegfried Trautmann, Stefan T., Gijs van de Kuilen, and Richard J. Dewitte (2019), “Investigating The Effect Of Childhood Zeckhauser (2013). “Social class and (un) ethical behavior: A Socioeconomic Background On Interpersonal Trust: Lower framework, with evidence from a large population sample.” Childhood Socioeconomic Status Predicts Lower Levels Of Perspectives on Psychological Science, 8(5), 487-497. Trust,” Personality and Individual Differences, 145, 19-25. Vernon-Feagans, Lynne, Patricia Garrett-Peters, Allison De The Annie E. Casey Foundation. 2019 Kids Count Data Book: Marco, and Mary Bratsch-Hines (2012), “Children Living In State Trends in Child Well-Being. https://www.aecf.org/m/ Rural Poverty: The Role Of Chaos In Early Development,” resourcedoc/aecf-2019kidscountdatabook-2019.pdf. In V. Maholmes & R. King (Eds). The Oxford handbook of The Washington Post (2018). “Are rich people more likely to poverty and child development (pp. 448-466). Oxford: Oxford lie, cheat, steal? Science explains the world of Manafort University Press. and Gates.” retrieved on August 14th 2018 from https:// Zhu, M., Y. Yang & C. K. Hsee (2018), “The Mere Urgency Effect,” www.washingtonpost.com/news/speaking-of-science/ Journal of Consumer Research, 45(3), 673-90. wp/2018/08/13/are-rich-people-more-likely-to-lie-cheat-steal- science-explains-the-world-of%E2%80%A6/ The Troubles of Gifts: Exploring the Challenges Faced by Givers and Receivers Chair: Lingrui Zhou, Duke University, USA

Paper #1: Buying Gifts for Multiple Recipients: How Culture ceivers, and that this is caused by mispredictions in feelings of guilt. Affects Whose Desires Are Prioritized Indebtedness then leads to increased likelihood for receivers to get a Ruomeng Wu, Western Kentucky University, USA return gift, which givers also do not anticipate. Mary Steffel, Northeastern University, USA In paper 4, Ward, Lowe, Reich, and Fajardo examine circum- Sharon Shavitt, University of Illinois at Urbana-Champaign, stances where giving a spontaneous gift results in worse outcomes USA than not giving a gift. The authors find that for high-investment relationships, receiving a low-expenditure gift decreases affiliation with Paper #2: Buying Love: Gifts as a Form of Social Support the giver compared to not receiving a gift at all. The effect is attenu- Hillary Wiener, University at Albany, USA ated for low-investment relationships. Holly Howe, Duke University, USA These four papers highlight the various obstacles givers and re- Tanya Chartrand, Duke University, USA ceivers encounter and some of the unintended consequences of gifts. Paper #3: The Burden of Gifts: How Givers Mispredict Negative Taken together, we provide insight on givers’ thought processes Reactions of Receivers when choosing what to give, as well as the potentially negative reac- Lingrui Zhou, Duke University, USA tions experienced by receivers. Accordingly, this session should have Tanya Chartrand, Duke University, USA broad appeal to those who are interested in gift giving, cultural differ- Paper #4: Thanks for Nothing: When Giving a Gift is Worse ences, social support, and consumer relationships. than Giving Nothing at All Morgan Ward, Emory University, USA Individuating Gifts out of Liking and Respect: Michael L. Lowe, Georgia Institute of Technology, USA Expanding Gift Giving Theory with a Cross-Culture Taly Reich, Yale University, USA Perspective Tatiana M. Fajardo, Florida State University, USA EXTENDED ABSTRACT SESSION OVERVIEW Gift givers often tailor their gift selections to show consideration Gift exchange is a common societal practice that has been a to recipients. Consequently, givers who shop for multiple recipients tradition for thousands of years (Culture and Society 2015). The pur- at once tend to pass up gifts that they know would be preferred in chasing of gifts is a giant industry all across the world. The aver- favor of different gifts for each recipient (Steffel and LeBoeuf 2014). age American was expected to spend over $900 on gifts in 2019, Although prior research illuminates how the immediate context in- a number that has been steadily growing over the past few years fluences overindividuation, little is known about the influence of the (Haury 2019). Moreover, gifts have important psychological effects, broader cultural context. The present research explores how cultural being able to bring people closer together and perpetuate relation- values influence overindividuation. ships (Belk and Coon 1993; Chan and Mogilner 2016). On the other We predict that overindividuation is not a uniquely Western hand, gifts also have the potential to create conflict and negatively phenomenon—Westerners’ independent values may foster overin- impact relationships (Sherry, McGrath, and Levy 1993). Given such dividuation by emphasizing the value of treating people as unique immense financial and psychological consequences, understanding individuals, but Easterners’ interdependent values may also foster the potential impacts, both positive and negative, that underly gift overindividuation by emphasizing the value of moderating choices exchange can provide important insight for consumers. In particular, based on the social context (Fukuoka et al. 2005). Additionally, how this session aims to shed light on the challenges faced during gift giv- people from different cultures individuate gifts may differ depending ing, from both the givers’ and receivers’ perspectives. on norms regarding how to properly acknowledge others’ standing Papers 1 and 2 examine the burdens givers experience when within a given social context. Whereas Western values emphasize given the opportunity to give a gift. In paper 1, Wu, Steffel, and choosing mainly based on personal preferences, Eastern cultures em- Shavitt explore the overindividualization of gift giving for multiple phasize choosing based on social hierarchy and status (Riemer et al. recipients from a cross-cultural perspective. The authors find that 2014). As a result, the criteria that consumers use to decide whose while both Easterners and Westerners tend to overindividuate when preferences to prioritize should differ across cultures. We predict that choosing for multiple others, Easterners are more likely to overindi- Westerners will prioritize the preferences of recipients that they like viduate when there are status differences amongst recipients. better by choosing a better-liked gift for personally preferred recipi- Sometimes gift giving is the response to a difficult situation ents and a unique but less-liked gift for less preferred recipients. But rather than the cause of one. Wiener, Howe, and Chartrand look they will not differentiate recipients based on their social status. By at gift giving as a method of social support. They demonstrate that contrast, Easterners will prioritize the preferences of recipients who when encountering someone in need of support, consumers are more have higher social status and not necessarily the preferences of re- likely to give gifts instead of providing emotional support when the cipients they like better. other is perceived as difficult to support. Interestingly, gift giving Experiment 1 examined whether givers from both a Western actually reduces the receiver’s negative affect more so than providing culture (United States) and an Eastern culture (China) overindividu- emotional support. ate gifts for multiple recipients. Participants were recruited simulta- Papers 3 and 4 explore the potential negative reactions that re- neously from the U.S. (N = 114) and China (N = 149) and randomly ceivers experience when getting a gift. Zhou and Chartrand examine assigned to imagine choosing a birthday gift for either one or two the mismatch between givers and receivers in terms of predicting unacquainted recipients. In the two-recipient condition, the target receivers’ feelings of indebtedness caused by the gift. They find that recipient preferred silver but also liked gold, and the other recipi- givers underestimate the amount of indebtedness experienced by re- ent preferred only silver. In the one-recipient condition, participants

Advances in Consumer Research 1150 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1151 chose gifts only for the target recipient. Despite their cultural differ- Overindividuating gifts for multiple recipients occurs across ences, givers in both cultures were less likely to get the silver frame cultures. However, the dimensions along which overindividuation for the target recipient in the two-recipient condition than in the one- occurs vary in ways that reflect cultural values. Although Americans recipient condition (41% vs. 76%; p < .001), and overindividuation overindividuate based only on personal liking, Chinese overindivid- did not vary based on culture (p = .98). uate based on both liking and status. Experiment 2 sought to determine whether givers from West- ern and Eastern cultures overindividuate differently based on - per Buying Love: Gifts as a Form of Social Support sonal liking, relative status, or both, and whether these differences are driven by cultural values dictating who should be shown greater EXTENDED ABSTRACT consideration. Participants from the U.S. (N = 318) and China (N Social support is critically important to people’s health and = 341) imagined choosing iPad cases for two unacquainted recipi- happiness (Thoits 1995), and not providing or providing low-qual- ents who either differed in how much they were liked (favorite co- ity social support damages relationships (Dakof and Taylor 1990). worker vs. liked coworker) or in their status (higher-status coworker However, people often do not provide support to others. When con- vs. same-status coworker). Both recipients preferred a black case, fronted with a loved one’s tragedy, would-be support providers often but the higher-status/favorite coworker also liked brown, and the avoid the affected, refuse to talk about the difficult topic, or mini- same-status/liked coworker also liked beige. Participants were free mize the problem (Barbee and Cunningham 1995; Dakof and Taylor to choose the same or different iPad cases for the recipients. Finally, 1990). One reason people avoid providing support is because provid- participants reported to what extent they felt they should show con- ing high-quality emotional support can be difficult and unpleasant sideration to each recipient. (Wortman and Dunkel-Schetter 1987). Overall, participants overindividuated by choosing the pre- In this paper we propose an alternate way that consumers can ferred black iPad case for the higher status/favorite coworker more support others who are struggling: giving them a gift. Gifts have often than for the same status/liked coworker (p < .001). More im- been found to be important in relationship maintenance, but most gift portantly, there was a three-way interaction between culture, rela- research has focused on gifts given for celebratory events (Ward and tionship, and recipient on gift choices (p = .01). Both Chinese and Chan 2015). We focus on gifts given in response to others’ negative American participants chose the preferred gift more often for the events, such as a loved one’s death. Consumers regularly give gifts, favorite coworker than the liked coworker (both p = .001). By con- such as flowers or care packages, under these circumstances (Post trast, Chinese participants chose the preferred gift more often for and Senning 2017). However, little is known about when consumers the higher-status coworker than the same-status coworker (p < .001), are most likely to give these gifts or how supported recipients feel but American participants did not differentiate between recipients when receiving them. based on status (p = 1.00). Furthermore, both American and Chinese We propose that the more difficult it is to provide emotional participants felt their favorite (vs. liked) coworker should be treated support, the more likely the support provider is to give a gift. In with greater consideration (both p < .05). However, Chinese partici- particular, support providers report that providing support to pessi- pants felt their higher (vs. same)-status coworker should be treated mistic, negative people is stressful and difficult (Forest et al. 2014). with greater consideration more so than did American participants (p Giving a gift is an alternative way to fulfil the support obligation < .001 vs. p = .07, respectively). (Clark 1987), so we propose that givers will be more likely to pro- Experiment 3 examined whether power distance belief, the de- vide support by giving a gift when the recipient is pessimistic. gree to which people believe in or endorse hierarchy and status dif- We also investigate how effective recipients find emotional sup- ferences in society, predicts gift choices in the same manner as the port versus gifts to be. Recipients may view gifts negatively, as giv- national culture differences. The design of Experiment 3 was similar ers “buying their affection” but also report that emotional support is to that of Experiment 2, except that participants completed a measure often ineffective (Lehman, Ellard, and Wortman 1986). Therefore, of power distance belief (α = .55; Zhang et al. 2010) rather than a we propose that receiving a gift will lead to a larger improvement in measure of consideration norms. mood than receiving emotional support will. As expected, Westerners and Easterners followed different Study 1a examines when givers choose to support someone criteria to overindividuate gifts for multiple recipients. U.S. givers with a gift rather than by providing emotional support. Participants (N = 210) prioritized the preference of the better-liked recipient ac- (n = 48) came to the lab in groups and were told that they would cording to their relative attitudes towards the recipients, such that be paired with another participant. They were shown the personality U.S. givers (p < .001), but not Chinese givers (N = 241, p = .26), profile of their “partner.” They did not have a partner; instead these chose the preferred gift more often for the best friend than for the profiles contained the recipient personality manipulation: The part- acquittance. However, Chinese givers prioritized the preferences of ner was either pessimistic or optimistic. Then participants were told the higher-status recipient according to the relative social status of their partner had been through a break-up and were asked whether the recipients, as Chinese givers (p < .001), but not U.S. givers (p = they wanted to spend five minutes talking to them about it (i.e., pro- .11) chose the preferred gift more often for the manager than the co- vide emotional support) or pay $1 (from their $7 study payment) worker. Moreover, power distance belief predicted gift choices (the to give them a gift. Participants were significantly more likely to three-way interaction: p < .05) in the same manner as did national provide support by giving a gift, rather than by providing emotional culture differences. This pattern of results was consistent with prior support, when the support recipient seemed pessimistic (54%) rather research, such that individual differences in power distance belief than optimistic (25%, χ2 = 4.27, p = .039). predict attention to the status signals sent by one’s consumption (e.g. Study 1b (n = 225) is a conceptual replication of study 1a. Par- Gao et al. 2016). In line with this, givers higher in power distance ticipants read a scenario about a pessimistic or optimistic person in belief prioritized the preferences of the recipient with higher status, need of support. Participants chose whether to support the recipient whereas those lower in power distance belief prioritized the prefer- by sending them a care package or talking to them on the phone and ences of the recipient that they liked better, but not vice versa. indicated how difficult providing support would be. Participants who read about the negative recipient were significantly more likely to 1152 / The Troubles of Gifts: Exploring the Challenges Faced by Givers and Receivers send a care package (46%) than were those who read about the posi- feelings of indebtedness that givers are unable to anticipate. Further- tive recipient (27%, χ2 = 9.60 p = .002). A mediation analysis showed more, we hypothesize that this effect is driven by recipients feeling that difficulty of providing support mediated the effect of condition more guilty than givers expect them to feel after receiving the gift. on gift giving (95% CI: [.02, .48]). People experience guilt when they think they have gone against an Study 2 (n = 118) is a dyadic lab study (59 dyads) that inves- important value (Sznycer 2019), in this case feeling currently unbal- tigates how recipients feel about receiving social support through a anced on reciprocity norms. We propose that feelings of guilt lead gift. This study had a 2 (emotional supported provided v. no emo- to a cognitive recalibration of the situation (Tooby and Cosmides tional support provided) x 3 (time: initial, post-support manipula- 2008) in which receivers realize they are indebted. This indebtedness tion, post-gift opportunity) mixed design. Participants came to the increases a need for reciprocity (Belk and Coon 1993), leading re- lab with a close other. One participant from each pair was randomly ceivers to want to give a return gift in order to reinstate equilibrium. assigned to the role of giver or recipient. Recipients first wrote about Across four studies, we demonstrate the inaccuracy of givers in (un- a current life problem that did not involve their partner and answered der)anticipating the burdens placed on receivers. questions about their mood and feelings about the problem. Then In study 1, we examined the mismatch between feelings of in- they talked to their partner for five minutes about their problem debtedness that givers predict receivers will experience compared (emotional support condition) or completed a control task before to what receivers actually experience. Participants were recruited completing the attitude measures again. Then givers were given the right after Christmas and were randomly assigned to think back to option to spend some of their study payment to purchase 0-4 gifts the winter holidays to either recall a gift they gave to someone or for their partner. The research assistant gave recipients their gift(s) if to recall a gift they received from someone. Givers then rated how they were given one, and recipients completed the measures a third indebted they thought the recipient would feel while receivers rated time. how indebted they actually felt (100-point slider scale). Results re- We first examined whether receiving emotional support made vealed that those who recalled giving a gift (M = 17.21) significantly recipients feel better than did completing a control task. We conduct- underestimated the indebtedness felt by those who recalled receiving ed a repeated measures ANOVA with the time at which the recipient a gift (M = 26.8; t(198) = -2.53, p = .01). rated their feelings as the within-subjects factor (time 1 [immediately Study 2 conceptually replicated the results of study 1 using after writing about their problem] v. time 2 [post-manipulation]) and dyads in the lab. Participants came into the lab in pairs (63 pairs), whether they received emotional support as the between-subjects with one partner assigned as the giver and the other assigned as the factor. Receiving emotional support did not reduce participants’ neg- receiver. The giver chose a gift for the receiver on the premise that ative mood (F (1, 57) = 2.15, p = .15), and only marginally reduced one receiver would be picked from the lottery to obtain the gift the their negative feelings toward the problem (F (1, 57) = 3.03, p = giver chose for them. Givers and receivers then rated feelings of the .087) more than completing the control task did. receiver’s indebtedness. Using a multi-level modeling approach to In contrast, receiving a gift reduced negative mood and feelings account for violations of statistical independence for pairs (Kenny, about the problem significantly more than not receiving a gift did. Kashy, and Cook 2006), we again find that givers underestimated We conducted a repeated measures ANOVA with the time at which feelings of indebtedness (B = 7.77, t(64.22) = 2.03, p = .05). the recipient rated their feelings as the within-subjects factor (time In study 3, we conducted a 2 (role: giver vs. receiver) x 2 2 [post-manipulation] v. time 3 [post-gift]) and whether or not they (friend: close vs. distant) experiment to test the proposed mechanism received emotional support as the between-subjects factor. Partici- and examine whether closeness with the exchange partner would af- pants who received a gift experienced a larger decrease in negative fect the results. Participants first named either a close friend or an emotions (F (1, 57) = 7.61, p = .008) and negative feelings toward acquaintance. Then they read a scenario in which they met with the the problem (F (1, 57) = 8.56, p = .005) than those who did not. friend and either gave or received a gift. Givers predicted receivers’ There was no main effect of having received support on post-gift feelings of indebtedness and guilt while receivers rated own feelings feelings (ps > .15). of indebtedness and guilt. An ANOVA revealed only a main effect In conclusion, these three studies show that gifts are not con- of role (F(1, 395) = 15.34, p < .001), with givers again mispredict- sumers’ default way of providing social support, but they use them ing indebtedness (Mgiver = 20.25 vs. Mreceiver = 31.01). There was no when they think the recipient will be difficult to support, and that effect of closeness or interaction between role and closeness (ps > they are quite effective at improving recipient mood. .46). Receivers also felt more guilty after getting the gift than givers

anticipated (Mgiver = 7.40 vs. Mreceiver = 11.55; F(1, 395) = 15.34, p = The Burden of Gifts: How Givers Mispredict Negative .03). Ratings of guilt mediated the mispredictions on indebtedness Reactions of Receivers (ab = -3.05, SE = 1.36, 95% CI [-5.63, -.36]). Finally, in study 4, we examined the predictions for both the EXTENDED ABSTRACT negative experiences of indebtedness and guilt as well as the positive Givers make many mispredictions when it comes to getting emotion of happiness. We conducted a 2 (role: giver vs. receiver) gifts for receivers. For example, they mistakenly think that unsolicit- x 2 (gift: desirable vs. not) experiment in which the pretested gift ed gifts would be appreciated more (Gino and Flynn 2011; Ward and was either a desirable or undesirable gift card. Then givers predicted Broniarczyk 2016) and that desirability matters more than practical- while receivers rated own feelings of happiness, indebtedness and ity (Baskin et al. 2014). Additionally, givers often focus on maximiz- guilt. We also measured predicted likelihood to give a return gift ing the happiness of receivers (Baskin et al. 2014), but do not seem as a downstream consequence of indebtedness (1 = not likely, 7 = to consider the potential negative consequences of the gift. Thus, very likely). First, we ran an ANOVA on happiness, which revealed the present research focuses on givers’ mispredictions of receivers’ a main effect of role (F(1, 397) = 16.66, p < .001), a main effect of negative feelings during gift giving. desirability (F(1, 397) = 7.35, p = .007), and a marginal interaction Social norms dictate that we cannot always be on the receiving between the two (F(1, 397) = 3.10, p = .08). For desirable gifts, giv- end of exchanges (Ruth, Otnes, and Brunel 1999). Thus, we propose ers and receivers had similar estimates of happiness (Mgiver = 79.35 that from the receiver’s perspective, receiving a gift will lead to more vs. Mreceiver = 84.68; p = .10), whereas for undesirable gifts, givers un- Advances in Consumer Research (Volume 48) / 1153 derestimated the happiness that receivers actually felt (Mgiver = 69.10 Gift giving is a cultural tradition meant to perpetuate future vs. Mreceiver = 82.50; p < .001). relationships. Given the future-looking nature of gift exchange, we Next, we ran an ANOVA on indebtedness and guilt ratings. For consider recipients’ predicted future relationship ‘affiliation’ with indebtedness, we found a main effect of role (Mgiver = 18.81 vs. Mre- givers as the focal dependent measure. ceiver = 31.51; F(1, 397) = 21.80, p < .001), but no significant effect of In Study 1, an RA asked students who had completed various gift desirability or interaction (ps > .66). We found the same pattern research studies, if they would complete an ‘additional task’ to help for feelings of guilt. Finally, we examined likelihood of buying a the PR department at the university. Participants were told that no return gift. Although there were no effects regarding desirability, we compensation would be given but that their participation would be found a main effect of role such that receivers felt more obligated to of great assistance to the PR department. Participants were instructed get a return gift for givers than the givers anticipated (Mgiver = 5.05 vs. to search the university’s name on Google.com/news, then read and

Mreceiver = 6.06; F(1, 397) = 50.97, p < .001). This effect was serially summarize the top search results. To manipulate relationship invest- mediated by the mismatch between giver and receiver on feelings of ment, students were either assigned to summarize 10 (vs. 2) articles. guilt and indebtedness (index = .03, SE = .02, 95% CI [.002, .07]. Those in the 10 article condition felt they had worked harder and Taken together, we demonstrate that givers chronically under- thus were more highly invested in their relationship with the orga- estimate how indebted receivers actually feel after getting a gift be- nization. After completing the task, participants were given either a cause they cannot foresee the guilt experienced by the receiver. As low expenditure gift (i.e. three plain pencil erasers) or nothing at all. a downstream consequence, receivers become more likely to give a Subsequently, participants endorsed their feelings of affiliation with return gift for the giver, which givers also do not anticipate. the PR department. An ANOVA conducted on the affiliation measure revealed a significant interaction between the two factors (F(1, 296) Thanks for Nothing: When Giving a Gift is Worse than = 4.04, p = .04). When HIR participants received a low-expenditure

Giving Nothing at All gift, they expressed lower feelings of affiliation (MLE = 4.09) than

when they received no gift (MNo Gift = 4.67; t = -3.04, p < .01). How-

EXTENDED ABSTRACT ever, in LIRs, this effect was attenuated (MLE = 4.62 vs. MNo Gift = Receiving a gift is generally regarded by both the giver and re- 4.69; t = -.32, p = .75). cipient as a relationship-validating experience. Prior research shows In Study 2, we examined whether HIR recipients were more that even when a giver chooses a less preferred item, the relation- sensitive to the relational implications of low-expenditure gifts re- ship between giver and recipient is typically enhanced (Dunn et al. ceived under spontaneous (vs. normative) conditions. In a 2 (Gift: 2008) as the recipient is likely to view the gift as a signal of the None vs. Low-Expenditure) by 2 (Gifting Context: Normative vs. giver’s positive intentions toward the recipient. Given strong social Spontaneous) participants read a scenario. To manipulate gift con- endorsement to avoid ‘looking gift horses in the mouth,’ coupled text, participants read a scenario in which they imagined completing with consumers’ motivation to interpret gifts as positive relational a project at work, after which their employer usually presented them symbols, it may seem unlikely that receiving a gift would harm the a gift (normative) or did not usually present one (spontaneous). Next giver-recipient relationship. However, we predict and find that there they imagined receiving a low-expenditure gift (a $5.00 gift card to are instances in which certain gifts may damage relationships. a restaurant) or no gift. Subsequently, participants endorsed their af- The context in which gift exchange takes place influences the filiation with the firm. recipient’s interpretation of the offering. Much of the prior gifting re- An ANOVA on affiliation revealed a significant interaction be- search focuses on ‘normative’ gift contexts wherein recipients have tween gift context and gifting scenario (F(1,178) = 5.24, p = .02). In high expectations of receiving a gift (e.g., birthday) and the items normative conditions, there was no significant difference in affilia- chosen are pre-determined by social norms (Joy 2001). However, tion when participants received no (vs. low-expenditure) gift (MNo Gift many gifts are offered spontaneously -- outside of prescribed giv- = 4.03 vs. MLE= 3.96; F(1, 178) = .082, p = .76). However, in spon- ing occasions. According to Carrier (1990), spontaneous gifts rep- taneous conditions, receiving no (vs. low-expenditure) gift resulted resent the truest form of giving because givers select these items for in significantly higher feelings of affiliation with the giverNo (M Gift recipients without the pressure of social obligation. Consequently, = 4.83 vs. MLE = 3.83; F(1, 178) = 10.93, p = .001). Thus, low- we suggest that recipients are likely to be more attuned to the mes- expenditure gifts given under normative circumstances ameliorated sages imbued in spontaneous gifts, because they are perceived to recipients’ feelings of disaffiliation. demonstrate the giver’s ‘true’ perceptions of the relationship (Belk In Study 3, in the context of interpersonal relationships, we and Coon 1993). examine how recipients’ perceptions of givers’ individual resources We predict that the context of spontaneous gift giving recipi- influence their predicted affiliation. In a 2 (Relationship investment: ents will interpret givers’ expenditures of resources on gifts as prox- High vs. Low) x 3 (Giver resources: Control vs. Restricted vs. Unre- ies for givers’ value for the relationships (Ruth et al. 1999). Thus, stricted) x 2 (Gift Expenditure: Low vs High) between-subjects ex- low-expenditure gifts may devalue the relationship relative to re- periment, participants imagined receiving either a low-expenditure ceiving no gift; in response recipients may distance themselves from gift (a $3, easily found, generic keychain) or no gift, after a book the giver. club meeting (spontaneous context). Notably, recipients were told We suggest that when gifts are offered spontaneously, low ex- that their HIR (LIR) friend either had “a really tight (light) schedule penditure gifts have especially detrimental effects on high-invest- this semester.” Under control conditions, participants were given no ment relationships (HIR) – relationships in which individuals have information about the giver’s time resources. committed substantial time or effort to one another. Conversely, in An ANOVA revealed a significant three-way interaction the context of low-investment relationships (LIR) – relationships in (F(2,378)=5.07, p<.01). When no mention of the givers’ resourc- which the individuals are less involved with or committed to one es was made in the unrestricted resource condition, results reveal another – recipients are likely to have lower expectations for the gifts significant two-way interactions between relationship investment they receive. Thus, receiving a low expenditure gift should be less and gift type (Control: (F(1,378)=3.88, p<.05); Unrestricted: (F(1, damaging for LIRs. 378)=4.31, p<.05)). However, when participants knew the giver had 1154 / The Troubles of Gifts: Exploring the Challenges Faced by Givers and Receivers restricted resources, the focal interaction among high investment re- Haury, Amanda (2019), “Average Cost of an American Christmas,” lationships was attenuated (MLEt=4.69, MNoGift=4.66; F(1,378)=0.25, Retrieved from: https://www.investopedia.com/financial- p=.87). edge/1112/average-cost-of-an-american-christmas.aspx In conclusion, we demonstrate that in spontaneous gift contexts, Joy, Annamma (2001), “Gift Giving in Hong Kong and the giving nothing to an HIR recipient may be more likely to preserve Continuum of Social Ties,” Journal of Consumer Research, the relationship than giving a low-expenditure gift. 28(2), 239-56. Kenny, David A., Deborah A. Kashy, and William L. Cook (2006), REFERENCES Dyadic Data Analysis. Guilford Press. Barbee, Anita P. and Michael R. Cunningham (1995), “An Lehman, Darrin R., John H. Ellard, and Camille B. Wortman Experimental Approach to Social Support Communications: (1986), “Social Support for the Bereaved: Recipients’ and Interactive Coping in Close Relationships,” Communication Providers’ Perspectives on What is Helpful,” Journal of Yearbook, 18(1), 381-413. Consulting and Clinical Psychology, 54 (4), 438-446. Baskin, Ernest, Cheryl J. Wakslak, Yaacov Trope, and Nathan Post, Lizzie and Daniel Post Senning, Emily Post’s Etiquette, Novemsky (2014), “Why Feasibility Matters More to Gift 19th Edition: Manners for Today, New York, New York: Receivers than to Givers: A Construal-level Approach to Gift HarperCollins Publishers. Giving,” Journal of Consumer Research, 41(1), 169-182. Riemer, Hila, Sharon Shavitt, Minkyung Koo, and Hazel Rose Belk, Russell W., and Gregory S. Coon (1993), “Gift giving as Markus (2014), “Preferences Don’t Have to be Personal: Agapic Love: An Alternative to the Exchange Paradigm Based Expanding Attitude Theorizing with a Cross-cultural on Dating Experiences,” Journal of Consumer Research, Perspective,” Psychological Review, 121(4), 619-648. 20(3), 393-417. Ruth, Julie A., Cele C. Otnes, and Frederic F. Brunel (1999), Carrier, James G. (1990), “Gifts in a World of Commodities: The “Gift Receipt and The Reformulation of Interpersonal Ideology of the Perfect Gift in American Society,” Social Relationships,” Journal of Consumer Research, 25(4), 385- Analysis, 29, 19-37. 402. Chan, Cindy and Cassie Mogilner (2017), “Experiential Gifts Sherry Jr, John F., Mary Ann McGrath, and Sidney J. Levy (1993), Foster Stronger Social Relationships Than Material Gifts,” “The Dark Side of the Gift,” Journal of Business Research, Journal of Consumer Research, 43(6), 913-931. 28(3), 225-244. Clark, Candace (1987), “Sympathy Biography and Sympathy Steffel, Mary and Robyn A. Le Boeuf (2014), “Overindividuation Margin,” American Journal of Sociology, 93 (2), 290-321. in Gift Giving: Shopping for Multiple Recipients Leads Culture and Society (2015), Curious History, Retried from: http:// Givers to Choose Unique but Less Preferred Gifts,” Journal of www.curioushistory.com/the-origins-and-history-of-gift- Consumer Research, 40(6), 1167-1180. giving/ Sznycer, Daniel (2019), “Forms and Functions of the Self- Dakof, Gayle A. and Shelley E. Taylor, “Victims’ Perceptions of conscious Emotions,” Trends in Cognitive Sciences, 23(2), Social Support: What is Helpful From Whom?,” Journal of 143-157. Personality and Social Psychology, 58(1), 80-89. Thoits, Peggy A. (1995), “Stress, Coping and Social Support Dunn, Elizabeth W., Jeff Huntsinger, Janetta Lun, and Stacey Processes: Where are we? What’s Next?”,” Journal of Health Sinclair (2008), “The Gift of Similarity: How Good and Bad and Social Behavior, (Extra Issue), 53-79. Gifts Influence Relationships,”Social Cognition, 26(4), 469- Tooby, John and Leda Cosmides (2008), “The Evolutionary 81. Psychology of the Emotions and Their Relationship to Internal Forest, Amanda, David R. Kille, Joanne V. Wood, and John G. Regulatory Variables,” in Handbook of Emotions, ed. M. Holmes (2014), “Discount and Disengage: How Chronic Lewis, J. M. Haviland-Jones, and L. F. Barrett, The Guilford Negative Expressivity Undermines Partner Responsiveness Press, 114-137. to Negative Disclosures,” Journal of Personality and Social Ward, Morgan K. and Cindy Chan (2015), “Gift Giving” in The Psychology, 107(6), 1013-1032. Cambridge Handbook of Consumer Psychology, Michael Fukuoka, Yoshimi, Kathleen Dracup, Sally H. Rankin, Erika I. Norton, Derek D. Rucker, and Cait Lamberton, eds. Sivarajan Froelicher, Fumio Kobayashi, Haro Hirayama, Cambridge, UK: Cambridge University Press, 398-418. Miyoshi Ohno, and David Matsumoto (2005), “Prehospital Ward, Morgan K., and Susan M. Broniarczyk (2016). “Ask and Delay and Independent/Interdependent Construal of Self You Shall (Not) Receive: Close Friends Prioritize Relational Among Japanese Patients with Acute Myocardial Infarction,” Signaling over Recipient Preferences in Their Gift Choices.” Social Science & Medicine, 60(9), 2025-2034. Journal of Marketing Research, 53(6), 1001-1018. Gao, Huachao, Karen Page Winterich, and Yinlong Zhang (2016), Wortman, Camille B. and Christine Dunkel-Schetter (1987), “All that Glitters is Not Gold: How Others’ Status Influences “Conceptual and Methodological Issues in the Study of Social the Effect of Power Distance Belief on Status Consumption,” Support,” in Handbook of Psychology and Health: Stress, Vol. Journal of Consumer Research, 43(2), 265-281. 5, Andrew Baum and Jerome E. Singer, eds. Hillsdale, NJ: Gino, Francesca and Francis J. Flynn (2011), “Give them what they Lawrence Erlbaum Associates, Inc., 63-108. want: The Benefits of Explicitness in Gift Exchange,”Journal Zhang, Yinlong, Karen Page Winterich, and Vikas Mittal (2010), of Experimental Social Psychology, 47(5), 915-922. “Power Distance Belief and Impulsive Buying,” Journal of Marketing Research, 47(5), 945-954. Subjective to Objective Value of Humans and Algorithms Chairs: Gizem Yalcin, Erasmus University, The Netherlands Evan Weingarten, Arizona State University, USA

Paper #1: Thumbs Up or Down: Consumer Reactions to stress and compulsive behaviors. In four studies, the authors find Decisions by Algorithms Versus Humans that consumers perceive AI (vs. human) mental health providers to Gizem Yalcin, Erasmus University, The Netherlands be less warm, which harms their perceived competence, resulting in Sarah Lim, Cornell University, USA higher aversion of AI providers. Finally, the paper by Weingarten Stefano Puntoni, Erasmus University, The Netherlands et al . tests whether AI outperform human experts in creative mar- Stijn van Osselaer, Cornell University, USA kets, specifically in logo design. The authors consistently find that the quality of the logos from human experts outperform those from Paper #2: People Prefer Forecasting Methods Similar to the artificial intelligence. Event Being Predicted This session provides timely insight into objective and subjec- Lin Fei, University of Chicago, USA tive valuation of algorithms and humans, an increasingly important Berkeley J. Dietvorst, University of Chicago, USA topic given the proliferation of this technology in consumer contexts. Paper #3: Stress, Addiction, and Artificial Intelligence Our proposed session directly addresses the ACR 2020 conference Mohamed Ayman Hussein, Stanford University, USA theme as it examines the value of deep thinking, rational thought, and Szu-chi Huang, Stanford University, USA reason for humans relative to algorithms in a new era of machine in- Paper #4: Experts Outperform Technology in Creative Markets telligence. We believe that this proposed session should be of interest Evan Weingarten, Arizona State University, USA to a broad audience of scholars and practitioners working on judg- Michael Meyer, University of California – San Diego, USA ment and decision-making, advice/recommendations, aesthetics, and On Amir, University of California – San Diego, USA new technologies. All projects have at least four completed studies. Amit Ashkenazi, Fiverr, USA Thumbs Up or Down: Consumer Reactions to Decisions SESSION OVERVIEW by Algorithms Versus Humans With the accelerated demand for new technologies, algorithms / artificial intelligence (AI) have integrated diverse and refined skills. EXTENDED ABSTRACT These advancements make them a viable alternative to using humans Companies are increasingly adopting algorithms to make deci- to complete the same tasks. Algorithms can outperform humans on sions that affect existing and potential customers, such as accepting medical diagnoses (Hutson 2017), taste-based recommendations and rejecting applications. Today, algorithms are commonly used to (Yeomans et al. 2019), predictions (Grove et al. 2000), and games decide who a company should hire (e.g., JetBlue) or provide services (Hosanagar 2019). to (e.g., rayatheapp.com). This growing trend calls for marketing Despite advancements in these new technologies in businesses, researchers to gain a better understanding of customers’ reactions however, the literature diverges on the extent consumers embrace or to decisions made by algorithms and humans. Previous research has dislike algorithms. It is still unclear from the literature when consum- predominantly focused on how individuals choose between an algo- ers are open to using algorithms (e.g., Dietvorst et al. 2015; Logg rithmic and human service provider (Castelo et al. 2019; Logg et al. et al. 2019). This disagreement raises important questions about the 2019; Longoni et al. 2019). Unlike this line of research, we investi- dynamics between humans and their valuation of technology: What gate the responses of individuals as a recipient of decisions made by factors affect consumers’ reactions towards decisions by algorithms either an algorithm or a human. (vs. humans)? Do consumers value decisions differently when gen- In this research, we propose that customers react to decision- erated by algorithms or humans? Does the framing of algorithms makers (algorithms vs. humans) differently depending on the valence change consumers’ willingness to use it? Under which situations are of decision outcomes, namely whether they are accepted or rejected consumers more likely to (de)value algorithms? Can algorithms ob- by a firm. Specifically, we hypothesize that customers react less posi- jectively outperform humans in domains involving aesthetics? tively (e.g., less perceived self-worth, less positive attitudes towards This session aims to identify the boundary conditions of when the firm), to an algorithm than a human decision-maker in the case of consumers appreciate algorithms more than humans (subjective val- favorable decisions (e.g., acceptances), whereas such a negative re- uation), and when algorithms do or do not objectively outperform action to an algorithm (vs. a human) would be attenuated in the case humans (objective). In the first paper,Yalcin et al . study consumers’ of unfavorable decisions (e.g., rejections). Our theorizing is based reactions to decisions that are made by algorithms versus humans. on attribution theory demonstrating that individuals tend to attribute Across seven studies, they reveal less positive reactions to favorable their successes and failures in a self-serving way (Halperin et al. outcomes (e.g., acceptances) by algorithms (vs. humans), whereas 1976). Namely, people tend to take credit for their success but blame they find no increased negative responses to algorithms when the others for failures so as to defend their self-esteem. Accepted cus- outcome is unfavourable (e.g., rejections). They also show that these tomers would be motivated to view the positive outcome as a result differences are explained by a shift in perceptions depending on the of their individual characteristics, thereby reacting more positively valence of the decision outcome. Lin and Dietvorst explore how to a human (vs. an algorithmic) decision-maker, who is perceived as the way algorithms are described changes consumers’ likelihood of more capable of incorporating individuals’ uniqueness into the deci- using them. Across seven studies and different prediction methods sion. Rejected customers, however, would attribute the unfavorable (e.g., navigation), they demonstrate that consumers prefer methods outcome to decision-makers (regardless of who the decision-maker of prediction that replicate the event in question even when doing is), viewing a human decision-maker as less objective and an algo- so is counterproductive. Next, Hussein and Huang investigate what rithm as more ignorant of their individual uniqueness. type of professional help consumers value more when dealing with

We tested our predictions across seven experiments (N = 3,535) In this paper, we find that the more similar a prediction method and showed that customers react differently to favorable (vs. unfa- is to the event being predicted (e.g. in its outcome distribution, pro- vorable) decisions that are told to be made by algorithms and humans cess, etc.), the more people like it, even when it does not perform as across various contexts (e.g., dating websites, bank loans), In studies well as alternatives. Humans often learn by mimicking and mirror- 1 and 2, we demonstrated that participants felt lower feelings of self- ing others’ actions (Meltzoff and Moore 1977), and people tend to worthiness when favorable decisions were made by an algorithm (vs. “over-imitate” - copying actions that are unnecessary to accomplish a human). This relatively negative reaction to algorithms is, howev- the given goal (McGuigan et al. 2011; Hoehl et al. 2019). We pro- er, mitigated when the decision outcome is unfavorable. Extending pose that this innate behavior transfers to consumers’ preferences these findings, study 3 revealed the same pattern of customers’ reac- for prediction methods: consumers prefer prediction methods that tions towards firms: participants demonstrated less positive attitudes best resemble the event being predicted. However, this preference towards the firm when an algorithm (vs. a human) accepts them, but can lead consumers to prefer prediction methods that offer subopti- such a relatively negative reaction to algorithmic decision-making mal performance. For example, when the outcomes of an event are was mitigated in the case of unfavorable decisions. at least partially determined by random chance (Fox and Ülkümen In studies 4a, 4b and 5, we aimed to understand possible driv- 2011), mimicking the event in question to predict its outcome will ers of such an interaction effect. In studies 4a-b, we revealed the result in overfitting of random error. In a set of 7 studies, we inves- negative impact of disclosing algorithmic decision-makers. Specifi- tigate consumers’ preference for similar prediction methods using cally, when the decision-maker was not explicitly mentioned by the both incentivized studies and real-world consumer scenarios. company, participants reacted similarly as they did to a human de- In Study 1, we present evidence that people like a prediction cision-maker. Next, study 5 tested whether participants would react method to be similar to the event in question by having participants more positively to humans even when they do not actively engage in predict the outcome of a die roll. We chose this task because rolling decision-making but monitoring the process. In line with our theori- a die has a clear process to be replicated, and its outcome is deter- zation, our results demonstrated that the relatively positive effect of mined by random chance. Participants were tasked with predicting human decision-making is driven by knowing that a human actively the outcome of a 7-sided die roll (sides 1,2,3,4,5,6,7). Participants made a decision to accept the applicants instead of passively moni- chose between two prediction methods: one that always predicts “4” toring the evaluation process. Finally, in study 6, we directly exam- (constant), and another that chooses a random number between 1 and ined the psychological mechanism underlying different reactions 7 (similar). Participants learned that they would receive a linearly to algorithmic versus human decision-makers. In these studies, we increasing bonus depending on the difference between their predic- showed that the differences in customers’ reactions stem from how tion and the actual outcome ($0.21 for a perfect prediction, -$.03 for they perceive these two decision-makers and that their perceptions each unit of error). With this payment scheme, choosing “4” (i.e. the of algorithms and humans (i.e., perceived objectivity, consideration constant method) offers the highest possible return in expectation. of applicants’ uniqueness). Thus, choosing the similar method instead of the constant method We believe that our research makes several contributions. Ex- is costly. However, we found that a substantial proportion of partici- tending the previous work that has predominantly studied situations pants (44.5%) chose the similar method. Even among participants where people choose whether to rely on algorithms or humans, we who passed a comprehension check by reporting that choosing “4” study situations where customers are recipients of decisions made by produced the highest expected earnings, 36.8% chose the similar algorithms versus humans. Furthermore, the current research dem- method, which suggests that consumers’ preference for a similar onstrates how motivated attribution plays a role in people’s percep- method is not due to ignorance. tion of algorithmic versus human decision-making. From a practical In Studies 2 and 3, we found that the more similar a prediction perspective, our work offers important managerial guidance. Manag- method is to the event in question, the more people like it. In study ers are often worried about deploying algorithms in customer-facing 2, we offered participants a choice between two randomly selected functions as they fear algorithms to amplify customers’ negative re- methods out of a set of three: the similar and constant options from actions in the case of unfavorable decisions. Our findings, however, Study 1, and a “mixed” option that was a combination of the two. suggest that managers should be more concerned about deploying The mixed option used the similar method with a probability of 50% algorithms in the case of favorable experiences as it can result in and the constant method with a probability of 50%. Participants were less positive customer reactions. Together, our work provides valu- more likely to select the similar method both when the alternative able insights on how firms can effectively communicate decision was the constant method (58%) and the mixed method (56.2%). In outcomes to customers. Study 3, we show that method similarity can be increased along multiple dimensions. Specifically, we hypothesized that participants People Prefer Forecasting Methods Similar to the Event would prefer a method that uses the same process as the event (roll- Being Predicted ing a die instead of drawing a marble) in addition to matching the outcome distribution. Participants rated the constant method and 4 EXTENDED ABSTRACT methods that varied on their proximity to the event being predicted Consumers face many scenarios where they have to make pre- (a fair 7-sided die roll) on 5-point scales: rolling a die that is skewed dictions by choosing between prediction methods. People can check towards the optimal answer (process resemblance), drawing a mar- multiple navigation tools for traffic predictions, analysts choose ble with values 1-7 from a jar (outcome resemblance), rolling a die between models to make financial market predictions, and patients with the same numbers as the focal die (perfect resemblance), and can choose between doctors and AI for medical diagnosis. It is clear drawing a marble with values skewed towards the optimal answer from past research that people don’t always pick the best perform- (no resemblance). Participants also rated the similarity of each pre- ing method (Arkes et al. 1986; Dietvorst et al. 2015; Yeomans et al. diction method to the event being predicted on two 5-point scales. 2019). However, it is still unclear how consumers choose between Consistent with our hypothesis, participants rated the prefect re- prediction methods. What features do people look for in prediction semblance method higher than all other methods (t’s(404)≥ 7.16, methods? p’s<.001). Further, participants ratings of the similarity between Advances in Consumer Research (Volume 48) / 1157 each method and the focal die roll mediate the differences in method We replicate this finding in a variety of different stress domains ratings [10.29,35.99]. These results suggest that the more similar a (Studies 1b-e). We consistently find that the AI adviser is dominated prediction method is to the event in question, the more people like it. by the two other advisers. The percentage of participants choosing In the remaining studies, we extend these findings to real-world an AI adviser ranges from 5.41% (academic/relationship stress) to consumer scenarios. In study 4, participants made an incentivized 20.99% (financial stress). choice between two prediction methods to predict a consumer’s rat- In Study 2, we investigate consumers’ choice of mental health ing of a movie. Participants chose between a similar method that adviser in a high-stigma context: porn addiction, which has harmful finds the person who gave the most similar responses to the focal consequences such as cognitive decline and isolation (Alarcón et al. consumer in a survey of movie preferences and uses their rating, 2019). We recruited men who watch at least 5 hours of porn per week and an alternative that uses the average rating of the movie among (2 SD higher than average). We provided these participants with an all consumers. 71% of participants chose the similar method even article that described the negative consequences of porn overuse and though it performs 1.5 times worse than the alternative. In study 5, asked them to choose which of the three advisers they would con- participants preferred a navigation tool that based a travel time pre- sider discussing their porn consumption habits with. Similar to the diction on the most similar trip in its data over another that averaged results above, we find that the AI adviser is dominated by the two among many somewhat similar trips. In follow up studies (Studies other choices (χ2 = 12.6, p-value<0.001). Interestingly, the percent- 6 and 7), we investigate how companies can frame the prediction age of participants willing to choose the AI adviser was higher (23%) methods that they offer consumers to be more similar to the event than in the stress-related domains we tested. in question. In Study 3, participants were told about a new service in which People’s systematic preference for similar prediction methods customers share symptoms with a (virtual vs. human) doctor who provides novel insight into consumer decision making, and suggests diagnose if they have common mental health disorders, such as de- interventions to boost consumers’ use of a prediction method. For expression and anxiety. Participants then reported their willingness to ample, it suggests that framing a prediction method as similar to the pay for the service. We found that participants in the AI condition event in question will make it more palatable to consumers, which reported a lower willingness to pay compared to the human condi- can help companies boost the use of prediction methods that they of- tion ( = -2.8, t(402) = -2.8, p = .005). Importantly, we found support fer consumers (e.g. recommendation systems, calculators, etc.). for our proposed serial mediation model, such that an AI (vs. human) provider was associated with lower perceived warmth, which Stress, Addiction, and Artificial Intelligence lowered the perceived competence of the provider, thereby leading to lower willingness to pay for the service (indirect effect = −.91, SE EXTENDED ABSTRACT = .27, 95% CI [−1.44, −.39]). As a robustness check, we included About 90.2% of people suffering from mental illness worldwide uniqueness neglect as a parallel mediator in the model and found do not receive adequate treatment (Alonso et al. 2018). A solution to that, while the indirect effect through warmth and competence re- this problem is therapy powered by artificial intelligence (AI), which mained significant (indirect effect = −.89, SE = .26, 95% CI [−1.40, is scalable at a low cost, convenient, and private. Are consumers −.38]), the indirect effect through uniqueness neglect was not signifi- open to adopting AI to manage their mental health? We first inves- cant (indirect effect = −.012, SE = .048, 95% CI [−.11, .082]), sug- tigate if consumers would choose an AI adviser over other types of gesting that uniqueness neglect did not affect consumers’ willingness human advisers present in the marketplace to treat anxiety and stress to pay for mental health providers in this context. (Studies 1a-e). We find that consumers are averse to choosing an AI In Study 4, participants were first asked to reflect on a recent provider to manage their mental health. We replicate our finding in time they experienced financial stress. Participants were then pre- a highly stigmatized context—porn addiction (Study 2)—and using sented with a [virtual] mental health coach [powered by artificial additional dependent variables, such as willingness to pay and shar- intelligence] that could help them deal with stressful situations. To ing intentions (Studies 3 & 4). We argue that perceptions of warmth ensure that consumers’ AI aversion generalized across other depen- shape consumers’ aversion to AI mental health providers: AI (vs. hu- dent variables of importance to marketers, we included a measure man) providers were perceived as less warm, which harmed their of participants’ intent to share information about the provider with perceived competence, resulting in higher aversion. others. Consistent with Study 3, we found that participants in the In Study 1a, we asked college students to fill out a survey about AI (vs. human) condition reported a lower willingness to pay (β = stress-related resources. After reflecting on an instance in which they −2.56, t(498) = −3.23, p = .001) and a lower interest in sharing inexperienced academic stress, participants were offered three types formation about this provider with others (β = −.39, t(498) = −2.05, of advisers to choose from to help them improve how they deal with p = .041). Importantly, perceptions of warmth again mediated the stress. Descriptions for all three advisers were based on real-world effect of provider on perceived competence, which in turn influenced descriptions. We scrapped the top 500 Google search results for participants’ willingness to pay (indirect effect = −.83, SE = .30, 95% mental health care providers, and two independent coders catalogued CI [−1.42, −.23]) and sharing intentions (indirect effect = −.41, SE = the providers into 5 categories. Three types of advisors—clinical, .07, 95% CI [−.55, −.28]). integrative, and virtual/AI—together form more than 80% of the Overall, we found that consumers are reluctant to use AI mental market. We thus focus on these three types of mental health provid- health advisers, and that perceptions of warmth played an important ers. The first was a clinical adviser with an MD from the university, role in shaping consumers’ preference for a mental health provider. who has published peer-reviewed articles on stress. The second was AI (vs. human) providers were judged as less warm, which harmed an integrative adviser, a life coach who has triumphed over dealing their perceived competence, and resulted in higher aversion. We rep- with stress. The third was a virtual adviser powered by artificial in- licated this finding across different types of mental health domains, telligence; the order of these advisers was counterbalanced. We find such as stress and addictive behaviors (e.g. porn addiction), across that only 5.4% of the participants chose the AI adviser (χ2 = 59.14, experimental paradigms (within-subject and between-subject), and p-value<0.001), compared to 64.9% choosing the integrative adviser across dependent variables (choice, willingness to pay, and sharing and 29.7% choosing the clinical adviser. intentions). 1158 / Subjective to Objective Value of Humans and Algorithms Experts Outperform Technology in Creative Markets In Study 1c, we replicate Studies 1a and 1b but with additional ratings designed to determine what dimensions the human expert lo- EXTENDED ABSTRACT gos exceeded artificial intelligence logos on. Undergraduates ina Recent decades have seen an increase in machines and other design course completed the procedure from Study 1a and also rated technological advancements coopting jobs from humans (Autor to what extent the logos conveyed the company’s industry, were aes- 2015; Brynjolfsson and McAfee 2014; Hosanagar 2019). Notably, in thetically pleasing, and were unique. In addition to replicating Study many cases algorithms or technology based on training from initial 1a on overall evaluations (F(1, 144) = 131.92, p < .001), human ex- human inputs can outperform humans (see Dawes et al. 1989; Diet- pert logos were also evaluated, relative to artificial intelligence, to vorst et al. 2015, 2016). Once thought unlikely, the machines have convey the company’s industry better (F(1, 144) = 290.20, p < .001), become dominant in games of human skill such as Jeopardy, chess, to be more aesthetically pleasing (F(1, 144) = 51.85, p < .001), and and Go (Hosanagar 2019). to be more unique (F(1, 144) = 57.40, p < .001). Recently, there has been discussion of technology based on ar- In Study 1d, we conceptually replicated the previous studies tificial intelligence supplanting workers (Lohr 2018; Peiser 2019). with incentivized choice with a set of west coast laboratory partici- While this replacement is argued to occur more for lower-skilled pants. That is, participants were informed that those participants who jobs (Autor 2015; Mokyr et al. 2015) and is expected more for chose the logo that was selected most often would be eligible for a jobs involving thinking (Waytz and Norton 2014), there have also $50 Amazon Gift card as a prize. Consistent with the previous stud- been market entrants from artificial intelligence in domains that are ies, participants chose logos produced by human experts more often typically considered to require the human spark of creativity (Rand than would be expected by chance (z = 8.49, p < .001). 1968). For example, for logos design, sites such as logomaster.ai Study 2a replicates Study 1a on the 62 logos produced for and Wix offer affordable artificial intelligence builders for logos. ONAK using Amazon Mechanical Turkers. Again, compared with Like with human experts, these artificial intelligence makers present the artificial intelligence logos, the human expert logos were evalu- people with multiple initial design concepts from which they can ated more favorably (F(1, 394) = 114.47, p < .001). Further, Study 2b choose, and they allow for iteration and revision on the initial con- replicated Study 1c on overall evaluations (i.e., human expert logos cept they like most (Goodwin 2009). were judged to be better than artificial intelligence logos) with a set Does artificial intelligence outperform human experts in cre- of west coast laboratory participants (F(1, 423) = 170.40, p < .001). ative markets, such as in logo design? To explore this question, we Finally, Study 2c attempts to replicate Study 1d on choice with a run two logo elicitation studies (Studies 1 and 2) in which subjects sample of west coast laboratory participants. When choosing among worked with artificial intelligence and/or human expert designers, logos for ONAK, participants were only marginally more likely to and then run seven follow-up studies (Studies 1a-1d and 2a-2c), us- select logos produced by human experts (z = 1.95, p = .051). ing both ratings and incentivized choice, in which an external set of Overall, using two types of logo elicitation methods, we find participants evaluate the quality of the logos. that logos from human experts outperform those from artificial intel- Importantly, we align the logo design process in the two logo ligence. This advantage may result from human expert logos being elicitation studies. In each, participants start with a creative brief, superior on aesthetics and clarity. However, subsequent coding and have a designer selected for them or by them, see at least two initial analyses revealed that those managers in Study 2 who invested more concepts for a logo, and then go through a revision process before effort into their creative briefs (based on an expert designer coding arriving at a final logo. In Study 1, we impose more experimental the strength of each brief) had more favorably evaluated artificial control by providing a fixed set of creative briefs for three (ficti- intelligence logos, which might mean that the advantage to human tious) companies (the data science company Empirical, the fashion experts may disappear with more experience or clarity of what is company Forward, and a restaurant Hyperion) and assigning par- desired from the design process. ticipants to work with either human experts or artificial intelligence logo makers, both of which provide initial concepts and a revision REFERENCES process. In Study 2, with a separate set of participants we allow for Alarcón, Rubén de, Javier de la Iglesia, Nerea Casado, and Angel more natural design by having participants write creative briefs for Montejo (2019), “Online Porn Addiction: What We Know and a company (ONAK, an origami canoe company), and then let them What We Don’t—A Systematic Review.” Journal of Clinical work both with one human expert designer of their choice and one Medicine 8 (1), 91. artificial intelligence logo maker. Study 1 yielded 17 human expert Alonso, Jordi, Zhaorui Liu, Sara Evans-Lacko, Ekaterina Sadikova, logos and 30 artificial intelligence logos, while Study 2 yielded 62 Nancy Sampson, Somnath Chatterji, Jibril Abdulmalik, et logos (31 human expert, 31 artificial intelligence). In Study 1, the al. (2018), “Treatment Gap for Anxiety Disorders Is Global: participants were MBA students enrolled in a new product develop- Results of the World Mental Health Surveys in 21 Countries.” ment course; in Study 2, the participants were managers enrolled in Depression and Anxiety 35 (3), 195–208. an evening MBA program. Arkes, Hal R., Robyn M. Dawes, and Caryn Christensen In Study 1a, we first seek overall evaluations on the logos from (1986), “Factors influencing the use of a decision rule in Study 1 using external raters. Amazon Mechanical Turkers evaluated a probabilistic task,” Organizational behavior and human all of the logos from one of the three companies (Empirical, Hyper- decision processes, 37 (1): 93-110. ion, or Forward) from Study 1 on a seven-point scale (1 = Very Bad, Autor, David H. (2015), “Why Are There Still So Many Jobs? The 7 = Very Good). Importantly, this evaluation was blind to whether History and Future of Workplace Automation,” Journal of the logos were produced by artificial intelligence or human experts Economic Perspectives, 29 (3), 3-30. (see Dietvorst et al. 2015). Across all three companies, relative to artificial intelligence logos, human expert logos were evaluated more favorably (F(1, 287) = 222.78, p < .001). In Study 1b, human expert designers with at least fourteen years industry experience evaluated all logos and showed directionally consistent results. Advances in Consumer Research (Volume 48) / 1159

Blanco, Carlos, Mayumi Okuda, Crystal Wright, Deborah S. Hasin, Hunt, Justin, and Daniel Eisenberg (2010), “Mental Health Bridget F. Grant, Shang-Min Liu, and Mark Olfson (2008), Problems and Help-Seeking Behavior Among College “Mental Health of College Students and Their Non–College- Students,” Journal of Adolescent Health, 46 (1), 3–10. Attending Peers: Results from the National Epidemiologic Leung, Eugina, Gabriele Paolacci, and Stefano Puntoni (2018), Study on Alcohol and Related Conditions,” Archives of “Man versus machine: Resisting automation in identity-based General Psychiatry 65 (12), 1429. consumer behavior,” Journal of Marketing Research, 55 (6), Brynjolfsson, Erik, and Andrew McAfee (2014), The Second 818-31. Machine Age: Work Progress, and Prosperity in a Time of Logg, Jennifer M., Julia A. Minson, and Don A. Moore (2019), Brilliant Technologies. “Algorithm Appreciation: People Prefer Algorithmic to Human Castelo, Noah, Maarten W. Bos, and Don R. Lehmann (2019), Judgment,” Organizational Behavior and Human Decision “Task-Dependent Algorithm Aversion,” Journal of Marketing Processes, 151, 90−103. Research, 56 (5), 809−25. Longoni, Chiara, Andrea Bonezzi, and Carey Morewedge (2019), Dawes, Robyn M., David Faust, and Paul E. Meehl (1989), “Resistance to Medical Artificial Intelligence,”Journal of “Clinical Versus Actuarial Judgment,” Science, 243 (4899), Consumer Research, 46 (4), 629–50. 1668-74. McGuigan, Nicola, Jenny Makinson, and Andrew Whiten (2011), Dietvorst, Berkeley J., Joseph P. Simmons, and Cade Massey “From over‐imitation to super‐copying: Adults imitate causally (2015), “Algorithm Aversion: People Erroneously Avoid irrelevant aspects of tool use with higher fidelity than young Algorithms After Seeing Them Err,” Journal of Experimental children,” British Journal of Psychology, 102 (1), 1-18. Psychology: General, 144 (1), 114−26. McQueen, Amy, and William M. P. Klein (2006), “Experimental Dietvorst, Berkeley J., Joseph P. Simmons, and Cade Massey Manipulations of Self-Affirmation: A Systematic Review,” Self (2016), “Overcoming Algorithm Aversion: People Will Use and Identity 5 (4), 289–354. Imperfect Algorithms If They Can (Even Slightly) Modify Meltzoff, Andrew N., and M. Keith Moore (1977), “Imitation of Them,” Management Science, 64 (3), 1155-70. facial and manual gestures by human neonates,” Science, 198 Ferrari, Francesco, Maria Paola Paladino, and Jolanda (4312), 75-78. Jetten (2016), “Blurring Human–Machine Distinctions: Mokyr, Joel, Chris Vickers, and Nicolas L. Ziebarth (2015), “The Anthropomorphic Appearance in Social Robots as a Threat History of Technological Anxiety and the Future of Economic to Human Distinctiveness,” International Journal of Social Growth: Is This Time Different?,”Journal of Economic Robotics 8 (2), 287–302. Perspectives, 29 (3), 31-50. Fox, Craig R., and Gülden Ülkümen (2011), “Distinguishing two Rand, Paul (1968), The Designer’s Problem. A Designer’s Art. Yale dimensions of uncertainty,” Perspectives on thinking, judging, University Press. and decision making, 21-35. Sontag-Padilla, Lisa, Michael S. Dunbar, Feifei Ye, Courtney Kase, Goodwin, Kim (2009), Designing for the digital age: How to create Rebecca Fein, Sara Abelson, Rachana Seelam, and Bradley D. human-centered products and services. John Wiley & Sons. Stein (2018), “Strengthening College Students’ Mental Health Grove, William M., David H. Zald, Boyd S. Lebow, Beth E. Knowledge, Awareness, and Helping Behaviors: The Impact of Snitz, and Chad Nelson (2000), “Clinical Versus Mechanical Active Minds, a Peer Mental Health Organization,” Journal of Prediction: A Meta-analysis,” Psychological Assessment, 12 the American Academy of Child & Adolescent Psychiatry 57 (1), 19-30. (7), 7500–7. Halperin, Keith, C. R. Snyder, Randee J. Shenkel, and B. Kent Steele, Claude M. (1988), “The Psychology of Self-Affirmation: Houston (1976), “Effects of Source Status and Message Sustaining the Integrity of the Self.” In Advances in Favorability on Acceptance of Personality Feedback,” Journal Experimental Social Psychology, edited by Leonard of Applied Psychology, 61 (1), 85−88. Berkowitz, 21, 261–302. Hoehl, Stefanie, Stefanie Keupp, Hanna Schleihauf, Nicola Waytz, Adam, and Michael I. Norton (2014), “Botsourcing and McGuigan, David Buttelmann, and Andrew Whiten (2019), Outsourcing: Robot, British, Chinese, and German Workers “‘Over-imitation’: A review and appraisal of a decade of Are for Thinking—Not Feeling—Jobs,” Emotion, 14 (2), research,” Developmental Review, 51, 90-108. 434−44. Hosanagar, Kartik (2019), A Human’s Guide to Machine Yeomans, Michael, Anuj Shah, Sendhil Mullainathan, and Jon Intelligence: How Algorithms Are Shaping Our Lives and How Kleinberg (2019), “Making Sense of Recommendations,” We Can Stay in Control. Journal of Behavioral Decision Making, 32, 403-414. Rejecting over Choosing: The Unknown Strengths of the Reject Frame Chair: Adelle X. Yang, National University of Singapore, Singapore

Paper #1: Decision Diagnosticity: Rejecting Induces a Larger options “stickier” in subsequent decisions. The rejecting procedure Post-Decision Evaluation Gap Than Choosing can thus be adopted to mitigate variety-seeking tendencies that im- Adelle X. Yang, National University of Singapore, Singapore pede consumer welfare. Jasper Teow, National University of Singapore, Singapore Besharat, Romero, and Haws find that the choose versus reject frames can influence the perception of customized products in Paper #2: “Swipe Left for No”: Why Options That Survive the context of food ordering. Their studies demonstrate that the re- Rejection Lead to Greater Purchase jecting procedure (i.e., eliminating ingredients) leads consumers to Jen H. Park, Stanford University, USA estimate the customized food as containing fewer calories and being Itamar Simonson, Stanford University, USA healthier than the choosing procedure (i.e., selecting ingredients), Paper #3: Rejections Are Stickier than Choices which also induces positive changes in intended future consumption Minzhe Xu, University of Florida, USA and the type of subsequent consumption. Yang Yang, University of Florida, USA Together, this special session sheds new light on when and how Paper #4: When More Is Less: How Rejecting (vs . Selecting) decision framing can be adopted to influence important post-decision Food Ingredients Leads to Lower Estimates of Calories consequences. Ali Besharat, University of Denver, USA Marisabel Romero, Colorado State University, USA Decision Diagnosticity: Rejecting Induces a Larger Post- Kelly L. Haws, Vanderbilt University, USA Decision Evaluation Gap Than Choosing

SESSION OVERVIEW EXTENDED ABSTRACT Many consumer decisions can be reached by either choosing A decision can be framed as either choosing wanted options or preferred options, or rejecting less preferred options. While prior re- rejecting unwanted options in a consideration set. For example, when search has established that framing a decision as choosing versus managers make a hiring decision between two job candidates, the rejecting influences the decision-making process, much of the previ- decision may be posed either as “Which candidate would you rec- ous studies focuses on how decision framing produces inconsistent ommend to accept?” or “Which candidate would you recommend to decision outcomes that violate the principle of procedural invari- reject?” While the difference between the two decision frames may ance (e.g., Dhar and Wertenbroch 2000; Shafir 1993; Sokolova and appear nominal, research shows that framing a decision as choose Krishna 2016). By contrast, the downstream consequences of choose versus reject produces systematically different decision outcomes versus reject framing, beyond the decision outcomes, have been rela- and consequences (Dhar and Wertenbroch 2000; Shafir 1993; Nag- tively underexplored. pal and Krishnamurthy 2008; Perfecto, Galak, Simmons, and Nelson This special session presents four papers that examine in- 2017; Sokolova and Krishna 2016). teresting consequences of decision framing, and highlight previ- In this research, we examine how decision framing affects ously unexamined strengths of the reject frame over the choose choice-induced preference changes. Six pre-registered experiments frame . Each paper uncovers an important consumer-relevant con- (N = 3,412) show that when a decision is framed as rejecting un- sequence of decision framing, from post-decision preference updat- wanted options, it leads to a larger post-decision evaluation gap being (paper 1), satisfaction and purchase likelihood (paper 2), variety tween the wanted and unwanted options, than when the decision is seeking in repeated consumption (paper 3), and perceptions of con- framed as choosing wanted options. Building on self-perception the- sumption outcomes (paper 4). ory (Bem 1972), and previous findings that rejecting tends to invoke Yang and Teow examine how decision framing influences post- greater processing depth (e.g., Nagpal and Krishnamurthy 2008; So- decision preference updating. They find the reject frame leads to a kolova and Krishna 2016) and greater attitude strength (Bizer and bigger post-decision evaluation gap between the wanted and unwant- Petty 2005; Bizer, Larsen, and Petty 2011), we propose that people ed options, compared with the choose frame, a finding attributable to perceive rejecting as more diagnostic of one’s true underlying pref- the act of rejecting being perceived as more diagnostic than the act of erences. Across the studies, we find process evidence consistently choosing. This effect was mitigated when the focal decision induced supporting the role of decision diagnosticity underlying the effect. a high level of decision difficulty, and reversed when all options were The studies followed a similar basic paradigm, with participants negatively-valenced. randomly assigned into 2 (choose vs. reject) between-subjects con- Park and Simonson test how the availability of an explicit re- ditions. Each participant was simultaneously presented with two or jecting procedure in sequential evaluation influences psychological more options, and asked to either choose the wanted option(s) or closure and subsequent consequences. They demonstrate that, be- reject the unwanted option(s). After their decision, participants were ing able to explicitly reject options, such as “swipe left to reject,” asked to evaluate each option. We calculated the post-decision evalu- heightens evaluation closure over the forgone options compared with ation gap by taking the difference between their evaluation of the when the decision procedure involves implicit rejection procedures wanted option(s) and their evaluation of the unwanted option(s). only. Consequently, when explicit rejection is allowed, it increases Study 1 (N = 256) established the effect with a pair of visually the likelihood of subsequent purchase. different consumer products (smartphone cases). Participants’ post- Xu and Yang investigate the effect of decision framing on decision evaluation gap was larger after the rejecting decision than variety-seeking behaviors in repeated decisions. Their experiments after the choosing decision (Mchoose = 2.08, SD = 1.87 vs Mreject = 3.02, 2 reveal that, consumers seek less variety when the repeated decisions SD = 2.78, F(1, 254) = 10.14, p = .002, ηp = .04). Moreover, we mea- are framed as rejecting than choosing, because rejecting reduces the sured the perceived decision diagnosticity, decision difficulty, and liking of the less preferred options, and hence renders the preferred decision finality associated with each option, and found that decision

Advances in Consumer Research 1160 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1161 diagnosticity best accounted for the effect: rejecting was perceived “Swipe Left for No”: Why Options That Survive as more diagnostic of one’s true underlying preferences than choos- Rejection Lead to Greater Purchase ing (Mchoose = 5.56, SD = 2.09 vs Mreject = 6.12, SD = 2.31, F(1, 254) = 4.20, p = .042), which induced more post-decision preference up- EXTENDED ABSTRACT dating (indirect effect in mediation analysis = -.282, CI95% = [-.57, Most alternatives considered by decision makers are rejected or -.01]). The other process measures only partially contributed to the not chosen, yet the manner in which individuals pass on alternatives effect, and choice share of the options did not differ between condi- has not received much attention. In social apps like Tinder, users tions. Similar results on the process measures were observed in later “swipe left” to make an explicit rejection before proceeding to the studies. next match, while interviewers often pass on a job candidate rather Study 2a (N = 308) replicated this effect with consumer prod- implicitly by looking at the next candidate. Despite the common use ucts that were presented with descriptions of multiple attributes, to of rejection strategies in everyday life, little research has examined be more in line with related prior research (e.g., Meloy and Russo how individuals make rejections and the psychological consequenc- 2004). Replicating our initial findings, participants’ post-decision es of rejection. evaluation gap was larger after the rejecting decision than after the We suggest that having the option to explicitly reject each al- choosing decision (Mchoose = 2.00, SD = 2.53 vs Mreject = 2.94, SD = ternative during sequential evaluation increases subsequent purchase 2 2.81, F(1,306) = 9.54, p = .002, ηp = .03), and decision diagnosticity from the narrowed set, largely due to the feeling of “evaluation clo- mediated the effect (indirect effect = -.275, CI95% = [-.56, -.01]). In sure.” Building on prior literature on psychological closure (Gu, addition, decision difficulty moderated this effect (2-way ANOVA Botti, and Faro 2013), we propose “evaluation closure”— the state F(1,304) = 6.47, p = .011): the framing effect was attenuated when experienced by consumers when their evaluation is perceived to be the participants perceived a high level of decision difficulty (> 4.4, complete—as a key driver of determining whether consumers pro- the Johnson-Neyman value), which impeded the self-inferential proceed to the next phase of decision-making: making a purchase. We cess that lead to post-decision preference updating. The effect was test in four experiments the predicted difference between options that also replicated in study 2b (N = 875) with visually different photo- survive explicit rejection (e.g., add-or-reject) and those identified graphs. through implicit rejection (e.g., add-only). The implications of this Next, study 3 (N = 461) expanded the generalizability of this research are discussed with respect to our understanding of evalua- effect by crossing the framing manipulation (choose vs reject) with tion modes in the current information environment. the size of the choice set: (picking 1/2 vs 2/4 paintings). The effect Study 1 (N = 296) tested the effect of explicit rejection during generalized to both the 1/2 and 2/4 conditions: participants’ post- sequential evaluation on purchase. Participants reviewed 10 paint- decision evaluation gap was marginally larger after the rejecting de- ings and indicated their liking of each painting. In the like-only concision than after the choosing decision (Mchoose = 2.50, SD = 1.62 vs dition, participants were asked to “swipe right” (drag the slider to the

Mreject = 2.80, SD = 2.33, F(1,457) = 2.78, p = .096), regardless of the right) for the paintings that they liked before moving onto the next size of the choice set, which had a separate main effect (M ½ = 2.30, painting. In the like-or-dislike condition, participants had to not only

SD = 2.2, M2/4 = 2.99, SD = 1.73; F(1,457) = 2.08, p < .001), and did “swipe right” on the paintings that they liked but also “swipe left” not significantly interact with framing F( (1,457) = 2.07, p = .151). on the paintings that they did not like. Then, participants indicated Study 4 (N = 1,051) revealed that the effect reversed when the their purchase likelihood (“Assuming that the price is reasonable, options were negatively valenced instead of positively valenced (2- would you buy any of the 10 paintings that you have just evalu- 2 way ANOVA: F(1,1047) = 6.5, p = .011, ηp = .01), consistent with ated?”). Results revealed that participants who made explicit rejec- prior findings supporting a compatibility effect between framing and tions were more likely to purchase a painting than those who did not 2 attribute valence (e.g., Nagpal and Krishnamurthy 2008; Perfecto, make rejections (Mlike-or-dislike = 91.7% vs. Mlike-only = 84.4%), χ (1) = Galak, Simmons, and Nelson 2017; also see Tversky et al. 1988). 3.80, p = .051. Study 5 (N = 461) replicated the effect when participants ex- In Study 2 (N = 605), we extended the rejection effect to non- plicitly expressed their preference before the decision (Mchoose = 2.81, binary purchase intentions and added another condition (i.e., dislike- 2 SD = 2.14 vs Mreject = 3.49, SD = 2.43; F(1,457) = 5.31, p = .022, ηp only) to further delineate the effect of rejection. Participants were 2 = .01; overall main effectF (1,457) = 4.52, p = .034, ηp = .01). Thus, randomly assigned to one of three conditions (like-only, like-or- the effect was not contingent on existing preference uncertainty. dislike, and dislike-only) and engaged in the sequential evaluation Last, we explored the potential influence of the weak axiom task as in Study 1; participants in the dislike-only condition were of revealed preference (WARP) across these studies. When treat- instructed to “swipe left” to reject the paintings and proceed to the ing violation of WARP (i.e., choosing A over B, yet rating A lower next painting if otherwise. Then, participants reported their overall than B) as an attention failure, we screened out participants who had satisfaction with the paintings that survived the evaluation (1 = Not internally inconsistent choice and preferences, and compared their at all, 7 = Very much). We found that compared to participants who results with those previously reported. This comparison allowed us did not make explicit rejections (Mlike-only = 5.94), those who had the to examine if the findings were attributable to an alternative explana- option to reject each painting reported greater satisfaction (Mlike-or- tion: if participants were more impelled to appear consistent after dislike= 6.22, Mdislike-only = 6.15), p = .008. This effect persisted even the rejecting procedure than the choosing procedure. The ratio of after controlling for the number of surviving paintings, p = .007. inconsistent responses was low (6% on average) and similar between Study 3 (N = 320) extended the effect to the shopping cart choose and reject conditions. Moreover, this screening did not sig- context and captured evidence of “evaluation closure.” Participants nificantly affect the main results. In sum, the reject frame enlarges evaluated nine microwaves one at a time and decided whether to the post-decision evaluation gap primarily because of greater prefer- add each microwave to their cart. While the button to add micro- ence diagnosticity perceived in reject decisions. waves (“add this item to shopping cart”) remained the same, the button to reject before looking at the next microwave differed across three conditions: “reject this item” (explicit rejection condition), “look at other items” (implicit rejection), no button (add-only). After 1162 / Rejecting over Choosing: The Unknown Strengths of the Reject Frame reviewing all microwaves, participants reported how much closure seeking. Imagine, for instance, you need to decide which restaurant they had experienced and decided whether they would purchase any to go to for lunch every day. Would you be more likely to stick to microwave from their cart. Consistently, participants who made ex- your favorite restaurant if you decide by choosing or rejecting? plicit rejections were more likely to make a purchase from their cart Across five studies and a wide range of product decisions, we

(Mreject this item = 92.8%) compare to those who did not make such re- find that rejections are stickier than choices—consumers tend to seek jections (Mlook at others = 88%; Madd-only = 86.7%), p = .039. Importantly, less variety across repeated decisions when the decisions are framed participants also reported feeling greater closure over their evalua- as rejections (vs. as choices). This effect occurs because the act of tions after making explicit rejections (Mreject this item = 4.75) vs. implicit rejecting has an asymmetric effect on one’s preference for different rejections (Mlook at others = 4.02; Madd-only = 3.91), t(317) = 3.40, p = .001. options: while it does not influence the liking of the more-preferred Study 4 (N = 394) examined the effect of rejection on purchase option(s), it significantly reduces the liking of the less-preferred likelihood in a free search task whereby participants viewed as many option(s). As a result, the act of rejecting broadens the liking gap or as few alternatives as they wished. Participants reviewed an un- between options, increasing one’s likelihood of selecting the more- known number of $25 gift cards that differed in their brands, and preferred option(s) in future decisions. we used the stopping point (i.e., the number of cards participants re- Study 1 (N=111) tested our hypothesis in a natural setting in- viewed before terminating their search) as an indicator of the evalu- volving real decisions. Specifically, participants made two snack ation completion; at any point during this task, participants could decisions—they first selected a snack to eat at the end of the lab end their evaluation by clicking the checkout button. To ensure that session, and then selected a snack to take home. Participants made participants treated their evaluation as consequential, they were en- these two decisions by either choosing or rejecting between snacks. tered into a lottery to win the gift card of their final choice. Similar We found that participants who rejected between snacks sought to Study 3, the button to reject each card was labelled as either an less variety than those who chose between snacks (5% vs. 36%, explicit rejection (“reject this item”) or an implicit rejection (“look at χ2(1)=15.47, p<.001) because the act of rejecting reduced the liking other items”). Finally, when participants reached their checkout, they of the less-preferred snack (3.07 vs. 3.82, F(1, 109)=7.37, p=.008; could choose a gift card option in their cart or defer (i.e., receive a 95% CI=[.036, .501]). bonus payment of $10 instead). Results showed that making explicit Study 2 (N=400) tested whether our framing effect is attribut- rejections led to earlier stopping points after evaluating fewer cards able to the increased level of deliberation associated with rejections.

(Mreject this item = 28.64) compared to making implicit rejections (Mlook Specifically, participants imagined that they were using a meal kit at other items = 31.07), t(392) = 2.02, p = .044. As expected, participants service and needed to select three out of six recipes each week for the who passed on gift cards through explicit (vs. implicit) rejections upcoming two weeks. We orthogonally manipulated framing and the were more likely to commit to an option from their “rejection sur- amount of deliberation by asking participants to either deliberate or vivors” rather than to defer (Mreject this item = 89.9% vs. Mlook at other items = not before they made choice or rejection decisions. We found that 1) 84.2%), χ2(1) = 2.86, p = .091. Importantly, this effect remained after deliberation did not moderate our effect (F(1, 396) = .79, p = .374); controlling for the number of evaluated options (B = .62, Wald’s χ2(1, 2) participants who removed three less-preferred recipes (i.e., reject- N = 394) = 3.86, p = .050). ing) sought less variety, compared with those who added three pre- In four studies, we demonstrate that explicit rejection during a ferred recipes (i.e., choosing; switches: 1 vs. 1.42, F(1, 398)=18.43, sequential evaluation of alternatives can increase consumers’ choice p<.001); 3) the liking of the less-preferred recipes mediated the commitment (e.g., purchase) to the “rejection survivors.” Overall, framing effect on variety-seeking (95% CI=[.058, .167]). this research has not only theoretical implications for understanding Study 3 (N=301) replicated the framing effect in real decisions rejection motivations but also practical implications for marketers to involving real consumption experiences, and demonstrated an inter- recapture online consumers’ commitment. esting boundary condition. Participants were told that they would view one of two photos across five iterations. During each -itera Rejections Are Stickier than Choices tion, they decided which photo to view and then viewed the selected photo for 10 seconds. Critically, this study included three condi- EXTENDED ABSTRACT tions: choosing, rejecting, and a new choosing-and-rejecting condi- Consumers often make the same decision repeatedly, ranging tion where participants indicated which option they would choose from which product to buy, which restaurant to visit, to which song and which option they would reject simultaneously. Replicating our to listen to. A prominent pattern in repeated decisions is variety- previous findings, participants in the rejecting condition sought less seeking—instead of sticking to their favorite option, consumers tend variety than those in the choosing condition (switches: 1.84 vs. 2.28, to alternate between different options due to satiation (Inman 2001; F(1, 298)=6.19, p=.013). Interestingly, the newly added choosing- McAlister 1982), preference uncertainty (Kahneman and Snell 1992; and-rejecting condition yielded the same pattern as the choosing Simonson 1990), information-seeking (McAlister and Pessemier condition (switches: 2.19 vs. 2.28, F(1, 298)=.26, p=.610). 1982), need for stimulation (Huang et al. 2019; Menon and Kahn Study 4 (N=404) further tested the generalizability of our effect 1995), or merely for the sake of variety (Ratner, Kahn, and Kahne- by manipulating option valence. Participants imagined that they had man 1999). to consume a serving of sparkling water on each of three consecutive Although existing literature on variety-seeking has enriched days. Half of the participants were asked to either choose or reject our understanding of when and why people seek variety, it predomi- among three pleasant flavors (i.e., watermelon, lime, and blueberry), nantly focuses on choices when eliciting consumer preferences, ask- while the other half were asked to either choose or reject among ing participants to make decisions by choosing what they like more. three unpleasant flavors (i.e., buffalo wing, ranch dressing, and clam This focus on choices, however, may not fully capture what happens chowder). Rejecting decreased variety-seeking among both positive in the real world where it is equally likely that consumers make de- (switches: 1.28 vs. 1.84, F(1, 400)=24.42, p<.001) and negative op- cisions by rejecting what they like less (Perfecto et al. 2017; Shafir tions (switches: 1 vs. 1.38, F(1, 400)=10.88, p=.001). 1993; Sokolova and Krishna 2016). The current research investigates Consumers have been found to seek more variety than nor- whether choosing versus rejecting influences the degree of variety- matively warranted (Ratner, Kahn, and Kahneman 1999; Read and Advances in Consumer Research (Volume 48) / 1163

Loewenstein 1995; Steffel and LeBoeuf 2014). Study 5 (N=400) can have important implications on consumers’ desire to visit the aimed to demonstrate that in situations where consumers tend to seek restaurant and subsequent food orders. too much variety, framing the decisions as rejections (vs. as choices) We test this premise in a series of four studies using various will lead to better decisions. This study is incentive compatible. Each approaches. In all of our studies, we measured calorie estimation participant listed two friends and then selected gifts for either one or by calculating the difference between actual and estimated calories both friends. When the gifting decisions were framed as choices, we and then dividing by the actual number of calories. Therefore, higher replicated the overindividuation effect (Steffel and LeBoeuf 2014): numbers indicate lower estimation of calories. while more than 79.8% of participants correctly selected the better- In Study 1, we utilized a controlled approach that involved a liked gift for either friend individually, only 55.0% of participants step-by-step customization task designed to be as realistic as pos- selected the better-liked gift for both friends (χ2(1)=14.34, p<.001). sible, holding constant the contents of the final product. Undergradu- When the gifting decisions were framed as rejections, however, ate students (n = 133) were shown either a selection or rejection path 79.0% of participants selected the better-liked gift for both friends to a chicken burrito with shredded cheese, lettuce, and sour cream (χ2(1)=.02, p=.892). that another customer was ordering and asked to estimate the final In sum, this research identifies a novel framing effect on vari- calories. They were provided with calorie counts for the base product ety-seeking and significantly extends the literatures on variety-seek- and the product with all six possible ingredients included, such that ing and decision framing. Practically, it offers recommendations for the possible range for the calorie estimation was restricted. Consum- firms on how to increase market shares and enhance consumer loy- ers estimated a lower amount of calories in the final product when alty without incurring additional cost. It also yields implications for they were performing a rejection compared to a selection task (Mrejec- consumer welfare: for domains in which consumers typically seek tion = 3.19% vs. Mselection = -11.70%; F(1,129) = 9.21, p < .01). more variety than normatively warranted, framing decisions as rejec- In the next studies, we allow participants to customize their tions can improve decision quality and increase consumer welfare, own food product using either a selection or elimination task. Even whereas for domains in which consumers typically seek too little though the selection of ingredients could vary across conditions, the variety, framing decisions as rejections may exacerbate the problem index of calorie estimation accounts for this difference. In Study 2, and reduce consumer welfare. students (n = 190) built their own nachos that they received at the end of the session through a selection or rejection approach and then When More Is Less: How Rejecting (vs . Selecting) Food estimated calories in the final product. The results of an ANOVA Ingredients Leads to Lower Estimates of Calories revealed that the rejection (vs. selection) condition led to a lower

estimation of calories (Mrejection = 24.44%, Mselection = - 11.27%; F(1, EXTENDED ABSTRACT 172) = 5.54, p < .001; η2= .18). Consumers are increasingly interested in customizing their Studies 3-4, asked participants to customize their own sand- food orders. As such, restaurants often allow customers to tailor wich and then estimated calories. After customization, participants their products, but this can be achieved in different ways. For in- in Study 3 (n = 111) indicated how likely they would be to order stance, consumers can order a pizza using a “pare-down” approach food from this restaurant. An ANOVA on the individualized calorie in which they eliminate undesired ingredients from a specialty pizza estimation measure revealed that participants in the rejection (vs. or a “build up” approach in which they add desired ingredients to a selection) task estimated lower calories (Mrejection = 13.18%; Mselection base product (Coker and Nagpal 2013; Nagpal, Lei, and Khare 2015; = - 8.13%; F(1, 102) = 19.22, p < .001; η2= .16). Furthermore, we Park et al. 2000). An important benefit of these strategies is that they found that participants that rejected (vs. selected) ingredients had a allow consumers to tailor food consumption based not only on taste higher desire to visit the restaurant (Mrejection = 4.79; Mselection = 4.18; preferences but also on health considerations. F(1, 102) = 4.02, p < .05; η2= .04). We also found a significant indi- Prior literature on product customization and decision frames rect effect of calorie estimation on patronage (Hayes 2013; 95% CI documents how a pare-down, versus a build-up, strategy shifts focus = .02, .53). to negative product attributes (Laran and Wilcox 2011; Shafir 1993; Study 4 explores the influence of customization on subsequent Huber, Neale, and Northcraft 1987). Different from the previous dis- food choices. Given that a rejection task leads to lower calorie esti- cussion, our work focuses on the influence of rejection versus selec- mates, we expect that they will feel more licensed to eat an unhealth- tion customization routes on calorie estimation. Our work proposes ier snack after this process takes place. MTurk panelists (n = 116) that the nature of the customization task can serve as a signal that bi- rejected (vs. selected) ingredients for a sandwich and then estimated ases calorie estimation of the final product. For instance, the process calories in the final product. Next, participants indicated how likely of rejecting is considered to be in line with a consumption-limiting they would be to add a cookie to their order (1 = not at all likely, 7 = mindset (Krishnamurthy and Prokopec 2009). Given that the rejec- very likely). An ANOVA on this measure revealed that participants tion (vs. selection) approach highlights the food items consumers are in the rejection (vs. selection) task estimate lower calories (Mrejection avoiding, the process of rejecting (vs. selecting) is more likely to be = 6.54%; Mselection = - 7.50%; F(1, 106) = 3.95, p < .05; η2= .04). associated with avoiding immediate pleasure to achieve long-term Participants in the rejection (vs. selection) task were also more likely benefits (Hung and Labroo 2010; Vohs and Baumeister 2016). Thus, to add a cookie to their order after the initial customization occurred the literature suggests that the nature of the rejection (vs. selection) is 2 (Mrejection = 3.85; Mselection = 2.98; F(1, 106) = 4.38, p < .05; η = .04). related with a higher degree of restraint. Moreover, because consum- We also found a significant indirect effect of calorie estimation on ers lack expertise to accurately estimate calories (Burton et al. 2006; likelihood to order dessert (Hayes 2013; 95% CI = .02, .45). Sharpe, Staelin, and Huber 2008), they generally use impressions to Across multiple scenarios, we find that a customization task that infer calorie content (Chandon and Wansink 2007). Given that in the involves rejecting unwanted ingredients, versus selecting desired in- rejection task, consumers focus on the food options that they denied gredients, leads to lower estimation of the caloric intake, even when themselves, they may rely on this association to estimate fewer calo- the final product is exactly the same. As the customization trend ries in their final product. Furthermore, this calorie underestimation continues to grow, consumers are continuously modifying their food orders. Our research provides a deeper understanding of the detri- 1164 / Rejecting over Choosing: The Unknown Strengths of the Reject Frame mental consequences that a specific food customization process can Laran, Juliano and Keith Wilcox (2011), “Choice, Rejection, and have on caloric estimation, future food choices, and retail evaluation. Elaboration on Preference Inconsistent Alternatives,” Journal of Consumer Research, 38 (2), 229-41. REFERENCES McAlister, Leigh (1982), “A Dynamic Attribute Satiation Model of Bem, Daryl J. (1972). “Self-perception theory.” Advances in Variety-Seeking Behavior,” Journal of Consumer Research, 9 Experimental Social Psychology 6, no. 1 1-62. (2), 141–50. Bizer, George Y., and Richard E. Petty. (2005). “How we McAlister, Leigh and Edgar Pessemier (1982), “Variety Seeking conceptualize our attitudes matters: The effects of valence Behavior: An Interdisciplinary Review,” Journal of Consumer framing on the resistance of political attitudes.” Political Research, 9 (3), 311–22. Psychology 26, no. 4, 553-568. Menon, Satya and Barbara E. Kahn (1995), “The Impact of Context Bizer, George Y., Jeff T. Larsen, and Richard E. Petty. (2011). on Variety Seeking in Product Choices,” Journal of Consumer “Exploring the valence‐framing effect: Negative framing Research, 22 (3), 285–95. enhances attitude strength.” Political Psychology 32, no. 1 Nagpal, Anish, Jing Lei, and Adwait Khare (2015), “To Choose 59-80. or to Reject: The Effect of Decision Frame on Food Burton, S., Creyer, E. H., Kees, J., and Huggins, K. (2006). Customization Decisions,” Journal of Retailing, 91 (3), 422. “Attacking the Obesity Epidemic: The Potential Health Park, C Whan, Sung Youl Jun, and Deborah J MacInnis (2000), Benefits of Providing Nutrition Information in Restaurants.” “Choosing What I Want Versus Rejecting What I Do Not American Journal of Public Health, 96(9), 1669-1675. Want: An Application of Decision Framing to Product Option Chandon, Pierre and Brian Wansink (2007), “The Biasing Health Choice Decisions,” Journal of Marketing Research, 37 (2), Halos of Fast-Food Restaurant Health Claims: Lower Calorie 187-202. Estimates and Higher Side-Dish Consumption Intentions,” Perfecto, Hannah, Jeff Galak, Joseph P. Simmons, and Leif D. Journal of Consumer Research, 34 (3), 301-14. Nelson. (2017).”Rejecting a bad option feels like choosing a Coker, Brent and Anish Nagpal (2013), “Building-up Versus good one.” Journal of personality and social psychology 113, Paring-Down: Consumer Responses to Recommendations no. 5, 659. When Customizing,” Journal of Retailing, 89 (2), 190-206. Ratner, Rebecca K., Barbara E. Kahn, and Daniel Kahneman Dhar, Ravi, and Klaus Wertenbroch. (2000). “Consumer choice (1999), “Choosing Less‐Preferred Experiences For the Sake of between hedonic and utilitarian goods.” Journal of Marketing Variety,” Journal of Consumer Research, 26 (1), 1–15. Research 37, no. 1, 60-71. Read, Daniel and George Loewenstein (1995), “Diversification Gu, Y., Botti, S., & Faro, D. (2013). Turning the page: The impact Bias: Explaining the Discrepancy in Variety Seeking Between of choice closure on satisfaction. Journal of Consumer Combined and Separated Choices,” Journal of Experimental Research, 40(2), 268-283. Psychology: Applied, 1 (1), 34–49. Hayes, Andrew F. (2013), “Methodology in the Social Sciences,” Shafir, Eldar (1993), “Choosing Versus Rejecting: Why Some Introduction to mediation, moderation, and conditional process Options Are Both Better and Worse Than Others,” Memory analysis: A regression-based approach. New York: Guilford and Cognition, 21 (4), 546-56. Press. Sharpe, K. M., Staelin, R., and Huber, J. (2008). Using Extremeness Huber, Vandra L, Margaret A Neale, and Gregory B Northcraft Aversion to Fight Obesity: Policy Implications of Context (1987), “Decision Bias and Personnel Selection Strategies,” Dependent Demand.” Journal of Consumer Research, 35(3), Organizational Behavior and Human Decision Processes, 40 406-422. (1), 136-47. Simonson, Itamar (1990), “The Effect of Purchase Quantity and Huang, Zhongqiang (Tak), Yitian (Sky) Liang, Charles B. Timing on Variety-Seeking Behavior,” Journal of Marketing Weinberg, and Gerald J. Gorn (2019), “The Sleepy Consumer Research, 27 (2), 150–62. and Variety Seeking,” Journal of Marketing Research, 56 (2), Sokolova, Tatiana, and Aradhna Krishna. (2016). “Take it or 179–96. leave it: How choosing versus rejecting alternatives affects Hung, I. W., and Labroo, A. A. (2010). “From Firm Muscles information processing.” Journal of Consumer Research 43, to Firm Willpower: Understanding the Role of Embodied no. 4, 614-635. Cognition in Self-regulation.” Journal of Consumer Research, Steffel, Mary and Robyn A. Le Boeuf (2014), “Overindividuation 37(6), 1046-1064. in Gift Giving: Shopping for Multiple Recipients Leads Inman, J. Jeffrey (2001), “The Role of Sensory-Specific Satiety Givers to Choose Unique but Less Preferred Gifts,” Journal of in Attribute-Level Variety Seeking,” Journal of Consumer Consumer Research, 40 (6), 1167–80. Research, 28 (1), 105–20. Vohs, K. D., and Baumeister, R. F. (2016). Handbook of Self- Kahneman, Daniel and Jackie Snell (1992), “Predicting a Changing regulation: Research, Theory, and Applications: Guilford Taste: Do People Know What They Will Like?,” Journal of Publications. Behavioral Decision Making, 5 (3), 187–200. Krishnamurthy, P., and Prokopec, S. (2009). “Resisting that Triple- chocolate Cake: Mental Budgets and Self-control.” Journal of Consumer Research, 37(1), 68-79. Being a Human in the Age of Artificial Intelligence Chairs: Gizem Yalcin, Erasmus University, The Netherlands Nofar Duani, New York University, USA

Paper #1: Defending the Human Need to Be Seen: Recipient highly subjective or consisting of unstandardized tasks. In the final Identifiability Aggravates Algorithm Aversion in Resource paper, Yalcin, Mehta and Dahl present a conceptual framework for Allocation Decisions integrating AI in creative cognitive processes. Given that creativity Jasper Teow, National University of Singapore, Singapore is no longer a process that is single-handedly achieved by humans, Adelle X. Yang, National University of Singapore, Singapore the authors develop Cyber-Creativity Framework, that captures the dynamics of the relationship between humans and AI in creative Paper #2: The Human Black Box: Illusionary Understanding processes. Specifically, they theorize that collaborating with AI en- Drives Preference for Human over Algorithmic Decision-Making hances humans’ performances in earlier stages whereas humans have Andrea Bonezzi, New York University, USA an edge over AI in the later stages. Massimiliano Ostinelli, Winthrop University, USA This session provides timely insight into how-and when-con- Johann Melzner, New York University, USA sumers are able to cooperate with algorithms, an increasingly im- Paper #3: How Will AI Affect Job Choice? portant topic given the proliferation of this technology in consumer- Noah Castelo, University of Alberta, Canada contexts. This session directly addresses ACR’s call for papers that Gerald Häubl, University of Alberta, Canada generate new knowledge on how consumers make decisions in the Paper #4: Cyber-Creativity: Unraveling the Dynamics Between age of new technology. This session should be of interest to scholars Humans and Algorithms in Creative Processes working on choice and decision making, inference-making, recom- Gizem Yalcin, Erasmus University, The Netherlands mendations, creativity, innovation, and new technologies. Ravi Mehta, University of Illinois at Urbana-Champaign, USA Darren W. Dahl, University of British Columbia, Canada Defending the Human Need to Be Seen: Algorithm Aversion and the Identifiable Human Recipient in SESSION OVERVIEW Resource Allocation Decisions With rapid technological progress and data proliferation, decisions previously made by humans are increasingly being delegated to EXTENDED ABSTRACT algorithms and Artificial Intelligence (AI) technologies, which have Many important allocation decisions, which used to rely on hu- been shown to outperform humans in a wide array of domains (Don- man judgment, can now be made by algorithms with advanced data nelly 2017; Lohr 2016; Silver 2012). Prior research has shown that processing and analytic capabilities. While algorithms are widely consumers are often averse to algorithms (Castelo et al. 2019; Önkal adopted in market predictions and product recommendations, the et al. 2009; Longoni et al. 2019) and are concerned that these tech- involvement of algorithms have provoked much resistance in the nologies could one day replace them (Granulo et al. 2019; Rainie and decisions that distribute scarce resources among individuals. For in- Anderson 2017). Though they engender substantial resistance, the stance, a recent survey indicated that 67% of Americans find it unac- widespread adoption of these technologies appears inevitable. The ceptable to use algorithms in hiring decisions (Pew Research 2018). papers in this session utilize different methodological approaches to Recent research shed important light on people’s general atti- explore how consumers can adapt to this impending transformation. tudes towards algorithms in decision making: while people appreci- Together, they explore what it means to be human in the age of AI, ate algorithmic advice for its superior efficiency and accuracy, they and how technology may change how consumers think and act. are less tolerant of algorithms’ mistakes (Dietvorst, Simmons, and The session’s first two papers discuss impediments to human- Massey 2015; Logg, Minson, and Moore 2019). In certain domains, AI collaborations. Teow and Yang show that consumers’ algorithm such as medical decision making, people tend to reject the use of aversion in resource allocation decisions depends on their empathetic algorithms due to a concern that algorithms may overlook the unique reactions towards the recipients. Seven studies reveal stronger algo- characteristics of each individual (Longoni, Bonezzi, and More- rithm aversion when an identifiable person is highlighted among the wedge 2019). decision recipients, than when no identifiable person is highlighted. Joining this growing literature, we examine people’s prefer- This is because the presence of an identifiable recipient heightens the ence between human experts and algorithms in resource allocation sentiment that individual human circumstances should be attended decisions. We find in seven pre-registered experiments (N = 2,732) to by fellow human beings instead of emotionless algorithms. In the that information about an identifiable decision recipient exacerbates second paper, Bonezzi, Ostinelli and Melzner show that people fos- algorithm aversion by evoking greater empathic reactions than inter an illusion of understanding human better than algorithmic deci- formation about statistical recipients. In particular, the empathic sion-making, when in fact both are black boxes. This occurs because, reactions evoked by the identifiable decision recipient highlight the for human decision-makers, people confuse a superficial, high-level concern that algorithms cannot produce appropriate emotions in re- understanding of what the decision-maker does, for a deeper, low- sponse to each recipient’s circumstance. These findings may reflect level understanding of how the decision-maker operates. Such con- a shared sentiment for the “compassionate gaze,” namely, that indi- fusion is less likely to occur for algorithmic decision-makers. vidual human beings’ circumstances should be attended to by fellow The remaining papers explore potential pathways for success- human beings. ful human-AI collaborations. Castelo and Häubl report five studies In each study, we introduced participants to a resource alloca- that show that students and workers are discouraged from pursuing tion decision, and asked them to indicate their preference between jobs that AI can also perform. Importantly, they find that the effect two decision methods: a committee of human experts versus an algo- hinges on whether consumers perceive the technology as helping hu- rithmic judge. The context of the focal decision varied across studies. mans or competing with them and is reduced when the job is seen as

Across these studies, we also find process evidence in support of our ent category. Thus, the stronger algorithm aversion evoked by an theorization and rule out important alternative explanations. identifiable recipient reflects the concern that algorithms are unable Studies 1 (N = 241), 2a (N = 232), 2b (N = 427), 2c (N = 417), to understand emotions. and 2d (N = 231) all followed a mixed design, with 2 repeated measures (recipient description: statistical versus identifiable), and coun- The Human Black Box: Illusionary Understanding terbalancing of the presentation order of the recipient description Drives Preference for Human over Algorithmic Decision- between-subjects. The description of the statistical recipients em- Making phasized the number of total recipients and the differences in their circumstances, whereas the description of the identifiable recipient EXTENDED ABSTRACT featured an image of one individual recipient with descriptions of Algorithms are increasingly used to make judgments and deci- the person’s circumstance. Participants were first presented with one sions that profoundly impact people’s lives, from whom gets incar- version of recipient description (either statistical or identifiable), cerated or admitted to universities, to whom gets hired or granted followed by an 8-point scale measuring participants’ preference be- credit. Prior research shows that people are often hesitant to trust al- tween the two decision methods for the focal allocation decision (1 = gorithmic over human decision-makers (Castelo, Bos, and Lehmann definitely prefer human committee; 8 = definitely prefer algorithmic 2019; Dietvorst, Simmons, and Massey 2014; Longoni, Bonezzi, judge), as well as a binary choice between the two decision methods. and Morewedge 2019). We propose that people’s reluctance to trust Then, participants were asked to indicate their empathic reactions on algorithmic over human decision-makers stems, at least in part, from a 6-point scale adopted from Erlandsson, Björklund, and Bäckström an illusion of explanatory depth (IOED; Rozenblitz and Keil 2002). (2015), followed by a 9-point scale measuring the compassionate Specifically, people foster the illusion that they understand how hu- gaze belief (with 4 items, e.g., “I believe that their [her] situation man decision-makers arrive at judgments and decisions better than should be seen”). Next, the same participant was presented with the how algorithmic decision-makers do so, when in fact both are black other version of recipient description, following by the same depen- boxes. This occurs because, for human decision-makers, people are dent variables and process measures. Therefore, each participant was more likely to confuse a superficial, high-level understanding of presented with both versions of recipient descriptions, and explicitly what the decision-maker does, for a deeper, low-level understanding informed that both descriptions were about the same decision sce- of how the decision-maker does it. This is less likely to occur for nario. algorithmic decision making. These studies revealed highly consistent results: featuring an Study 1. The goal was to provide initial evidence that people identifiable decision recipient evoked stronger algorithm aversion foster the illusion that they understand human decision-makers bet- than presenting statistical recipients (despite participants being exter than algorithmic decision-makers. We test this hypothesis using plicitly reminded that the focal decision applies to each individual as a typical IOED paradigm, which unveils illusionary understanding well as the entire group); regardless of presentation order. More spe- by asking people to explain in detail how something works (Alter, cifically, these inconsistent preferences were observed both between- Oppenheimer, and Zemla 2010). subjects and within-subjects. We thus collapsed the data to compare 401 MTurkers were randomly assigned to a 2(decision-maker: the average preference under each version of recipient description. algorithm, judge) by 2(explanation: yes, no) between-subjects de- In addition, the process measures in these studies suggest that, an sign. Respondents were presented with a news story about a man identifiable decision recipient evoked stronger emotional reactions found guilty of stealing a car and sentenced to five years in prison, and a stronger endorsement of the compassionate gaze belief. Serial on the basis of a recidivism assessment conducted either by a judge/ mediations supported their roles underlying the observed effect (e.g., algorithm. In the no explanation condition, respondents were asked serial mediation in study 1: indirect effect = -.39,CI 95% = [-.56, -.25]). to rate to what extent they understood how the judge/algorithm Next, study 3 (N = 441) included the same mixed design with evaluated the risk that the defendant would re-offend (three items on another between-subjects factor of recipient prospect (good vs bad). 7-point scales, α=.94, averaged into a sense of understanding score). In a college admission scenario, the identifiable decision recipient In the explanation condition, respondents were asked to explain in was either a more qualified candidate with a good prospect of get- detail how the judge/algorithm evaluated the risk that the defendant ting admitted to a top university, or a less qualified candidate with would re-offend, before rating their understanding. a poor prospect of getting admitted. We found highly similar results A 2(decision-maker: algorithm, judge) x 2(explanation: yes, between conditions (more qualified recipient: Mstatistical = 4.89, SD = no) ANOVA revealed a significant main effect of explanation,

1.92 vs Midentifiable = 2.94, SD = 1.64, F(1,878) = 139.00, p < .001; F(2,397)=5.27, p=.022 and a marginally significant interaction less qualified recipient: Mstatistical = 4.68, SD = 1.96 vs Midentifiable = F(2,397)=3.73, p=.054. Participants in the no explanation condi- 2.44, SD = 1.31, F(1,878) = 187.33, p < .001), and no interactions tion indicated a lower sense of understanding when the decision- between identifiability and recipient prospect (F(1,878) = 1.54, p = maker was an algorithm (M=2.83, SD=1.61) rather than a judge .216). These results suggest that the effect is generalizable across (M=3.46, SD=1.80, p=.013). However, in the explanation condition, identifiable decision recipients and not attributable to participants participants’ sense of understanding did not differ significantly as perceiving the identifiable individual to be more deserving than the a function of decision-maker (judge: M=2.70, SD=1.88; algorithm: average recipient. M=2.76, SD =1.83, p=.819). Key to our hypothesis, asking to pro- Last, study 4 (N = 743) crossed the same mixed design with an- vide an explanation significantly reduced sense of understanding other between-subjects factor (recipient category: victim vs. villain), when the decision-maker was a judge (p=.003), but not when it was to further examine if the effect holds when the decision recipients an algorithm (p=.795). These results provide evidence that people evoke negative emotional reactions, or if participants (mis)perceive foster an illusion of understanding human decision makers better human decision makers to be more lenient towards any identifiable than algorithmic decision makers. individual. Again, we observed the main effect (Mstatistical = 4.27, SD Study 2. The goal was to provide converging evidence for our

= 2.61 vs Midentifiable = 3.31, SD = 2.54, F(1,1484) = 50.97, p < .001), illusionary understanding hypothesis, and test our proposed process and no significant reversal between recipient description and recipi- via moderation. Advances in Consumer Research (Volume 48) / 1167

442 undergraduate students were randomly assigned to a 2(de- abstraction (Kim and Duhachek 2020) and confuse a superficial un- cision-maker: algorithm, judge) by 3(mindset: how, why, control) derstanding of the process for a deeper understanding of how the between-subjects design. Respondents in the how/why mindset con- process actually works. ditions were instructed to consider how/why they would improve and maintain their health (Freitas, Gollwitzer, and Trope 2004). The How Will AI Affect Job Choice? control group did not complete this manipulation. Respondents then read the same news story described in the previous study, and an- EXTENDED ABSTRACT swered the same three items aimed to measure their sense of under- We explore how learning about the existence of AI in a given standing (α=.78). job alters students’ and workers’ willingness to pursue that job. We A 2(Decision-maker: algorithm, judge) x 3(mindset: how, why, show that perceiving AI as competing with humans in a given job de- control) ANOVA revealed a significant interaction, F(2,436)=5.11, creases willingness to pursue that job, while perceiving AI as helping p=.006. In the control condition, sense of understanding was higher humans in a job increases willingness. We also propose two modera- when the decision-maker was a judge (M=4.37, SD=1.51) rather than tors of this effect. an algorithm (M=3.68, SD=1.38, p=.003). This result replicates the The first is whether the job is seen as more objective or subjec- findings from Study 1. When the decision-maker was an algorithm, tive. Objective jobs require the use of quantifiable facts and data to sense of understanding did not differ across mindsets (algorithm con- perform them successfully, relative to subjective jobs which require trol: M=3.68, SD=1.38; algorithm how: M = 3.86, SD =1.15; algo- the use of intuition or personal opinion. People tend to see algorithm why: M=4.09, SD=1.18, p=.191), but when the decision-maker rithms as more capable of performing objective vs. subjective tasks. was a judge, sense of understanding differed across mindsets (judge Thus, emphasizing that a given job is best accomplished by subjec- control: M=4.37, SD=1.51; judge how: M=3.59, SD=1.52; judge tive (vs. objective) means may encourage more people to pursue that why: M=4.00, SD=1.34, p=.003). Importantly, when the decision- job even if AI is competing in that job. maker was a judge, participants in the how mindset reported a sig- A related factor is whether people focus on the standardized nificantly lower sense of understanding than participants in the con- vs. unstandardized aspects of a given job. Independent of whether trol condition (p=.001). These results provide further evidence that people believe that a job is best accomplished by objective or sub- people foster an illusion of understanding human decision-makers jective means, most jobs also consist of some tasks that are rela- better than algorithmic decision-makers, because they construe hu- tively rule-based, repetitive, routine, and standardized, as well as man decision-making in terms why and algorithmic decision-making some tasks that are relatively qualitative, non-repetitive, variable, in terms of how. and unstandardized. Emphasizing the more unstandardized aspects Study 3: The goal was to probe our proposed process via media- of a given job may therefore also encourage more people to pursue tion, test for consequential outcomes, and extend the investigation to that job even in the presence of AI. In Study 1,221 business students a different domain. read an article either about AI being used in accounting, or about a 198 MTurkers were presented with five activities in the HR neutral topic, before indicating their interest in accounting careers. domain (e.g., screening job candidates). Respondents indicated how Reading the article about AI in accounting significantly decreased much they would trust a manager/algorithm to perform each of these students’ interest in accounting careers (β = -.86, p = .002). In the activities on a 7-point scale (α=.89). Respondents then indicated the control condition, 70% of students scored at or above the scale mid- degree to which they understand how a manager/algorithm performs point on the dependent variable, indicating at least moderate interest each of these activities (α=.92). Finally, respondents completed a in accounting careers; in the AI condition, this figure was 50% (χ2 = task, designed in the spirit of the behavioral identification form (BIF; 8.34, p = .004). Vallacher and Wegner 1989), aimed to measure how they construed In Study 2, 151 Prolific participants chose to work on 1 of 2 pre- each of the five activities performed by a manager/algorithm. Final- diction tasks; after working on their chosen task for several rounds ly, to measure perceived accuracy, respondents rated the likelihood (i.e., making several predictions), we introduced an AI system to the of a manager/algorithm being inaccurate when performing each task task, which we told them would be either competing with them for (α=.84). bonus payments, helping them at the task, or we left the AI’s role ambiguous. We measured whether they chose to continue working on Respondents indicated higher trust (MManager=5.45, SD=.95; their chosen task with the AI or switch to the other task where there MAlgorithm=3.76,SD=1.27; t(196)=10.55, p<.001), higher sense of was no AI. Participants in the ambiguous AI condition chose to work understanding (MManager=5.32, SD=1.11; MAlgorithm=3.93, SD=1.55; alongside the AI 76% of the time, those in the AI help condition 86% t(196)=7.21, p<.001), and more why identifications (MManager=3.11, of the time, and those in the AI competition condition 65% of the SD=1.19; MAlgorithm=1.44, SD=1.39; t(196)=9.06, p<.001) in the manager versus the algorithm condition. We conducted a mediation time. The only significant two-sample proportions test was between analysis with decision-maker as independent variable and trust as de- the competition and help conditions (χ2 = 5.16, p = .023). pendent variable. Behavioral identification and sense of understand- 500 Prolific participants completed Study 3, choosing to work ing were specified as serial mediators, and accuracy as an alternative on a task either alone or with AI. We again manipulated whether the mediator. The results revealed a significant sequential indirect ef- AI would help them or compete against them. We also told them that fect (decision-maker -> abstraction -> understanding -> trust: b=.16, they could expect a higher bonus payment if they chose to work with CI95% [.06;.29]). The indirect effect through accuracy (decision- AI, since the per-round bonus payment would be higher. Neverthe- maker -> accuracy -> trust) was also significant (b=.19, CI95% less, 11% chose to work alone in the “AI help” condition, and 28% [.06;.36]). The direct effect (decision-maker -> trust) was reduced did so in the “AI competition” condition (χ2 = 20.6, p < .001). Thus, a after accounting for both processes (b=1.17, CI 95% [.82;1.51]), in- preference to avoid working with AI persists despite financial incen- dicating partial mediation. tives to the contrary. Across three studies, we show that people foster an illusion Study 4 manipulated both help vs. compete framing as in Study of understanding human better than algorithmic decision-making, 3 and framed the task itself as either objective or subjective. We because they construe human decision-making at a higher level of again measured the choice to work alone or with AI on the task. A 1168 / Being a Human in the Age of Artificial Intelligence

2x2 regression revealed that the choice of working alone was more processes raises a vital question about the future of the relationship likely when AI was competing vs. helping (β = .52, p < .001) and between humans and AI in creative industries. more likely when the task was framed as being subjective (β = .10, p Despite the widespread adoption of AI in creative industries, = .068); there was also a significant interaction (β = .24, p = .002). what is known in the CB academic literature is based on human cre- When the task was framed as being highly objective, 70% of par- ativity. Previous work on creativity predominantly studies the cogni- ticipants chose to work alone when AI was competing, compared tive processes that lead to an output that is not only different from to 18% choosing to work alone when AI was helping (χ2 = 62.32, what is already known but also is effective in solving the problem p < .001). When the task was framed as being highly subjective, at hand (Moreau and Dahl 2005). In their influential paper, Finke 80% chose to work alone when AI was competing, compared to 52% and colleagues (1992) propose that such cognitive processes can be when AI was helping (χ2= 21.41, p < .001). Thus, in both cases, there categorized into two critical phases that define creative cognition was a strong effect of AI helping vs. competing, but the effect was and present the “Geneplore Model” of creative thinking. The authors significantly smaller when the task was seen as subjective. argue that generative processes entail constructing mental repre- Finally, Study 5 manipulated 267 business students’ focus on sentations or preinventive cognitive structures of a desired creative the standardized vs. unstandardized aspects of accounting before solution (Murphy 1988). Once these preliminary mental structures reading one of the same articles from Study 1 (about either AI in ac- are generated, exploratory processes are engaged to interpret these counting or a neutral topic). We mentioned that accountants mainly preinventive structures in a meaningful way until a satisfactory cre- work on tasks like “calculating taxes owed and filling out tax filing ative solution is reached. forms” which are largely repetitive and rule-based, or like “interact- In the current work, we evaluate and assess how AI technolo- ing with clients in order to understand their short- and long-term fi- gies may redefine creative cognitive processes. Building on an in- nancial goals and obligations” which are non-repetitive and intuition fluential theory of human creative processes (Geneplore model by based. We again measured interest in accounting careers. Finke et al. 1992), we present the Cyber-Creativity Framework, that A 2 (AI vs. control article) x 2 (high vs. low job standardization) captures the dynamics of the relationship between humans and AI in ANOVA revealed a significant interaction,F (1,265) = 7.13, p = .008, creative processes. but no significant main effects,F ’s < 2.20, p’s > .139. When account- The creative cognitive process starts with the generative phase ing was described as consisting of highly standardized tasks, reading during which one constructs mental representations and preinventive about AI in accounting decreased students’ interest in accounting ca- structures that are evaluated in the exploration phase to reach a final reers (M = 2.84) relative to reading about a neutral topic (M = 4.17, creative solution (Murphy 1988; Perkins 1981). In this initial phase, t(126) = 3.04, p = .003). When accounting was described as consist- individuals retrieve mental information, form associations among ing of highly unstandardized tasks, reading about AI in accounting retrieved mental concepts, and then synthesize or transform these did not decrease interest in accounting careers (M = 3.38) relative concepts/information to generate preliminary mental structures. We to reading about a neutral topic (M = 3.02, t(124) = .79, p = .431). propose that AI can augment human performance during generative Students and workers are therefore discouraged from choosing processes by mitigating their possible weaknesses (e.g., limited com- jobs in which AI can also perform, especially when AI is seen as puting power, limited memory, mental blocks, unconscious plagia- competing; this effect can be reduced by highlighting the subjective rism; e.g., Baddeley 2013; Dahl and Moreau 2002; Miller 1956). or unstandardized nature of the job. Specifically, our model demonstrates that collaborating with AI improves humans’ performances by enhancing their memory, flex- Cyber-Creativity: Unraveling the Dynamics Between ibility, and comprehension. In contrast, we argue that AI shows a Humans and Artificial Intelligence in Creative Processes more secondary role during exploratory processes. Exploratory process entails interpreting the newly defined structures in meaningful EXTENDED ABSTRACT ways (Finke et al. 1992). Such evaluations of preinventive structures For decades, the role of advancing technology had been limited utilize cognitive processes that encompass metaphorical or theoreti- to helping humans perform tasks more efficiently and with higher cal interpretations of the generated concepts, exploration of different quality (van Bruggen et al. 1998). As technologies became more so- meanings and functional inferences of structures, as well as testing phisticated; however, they started taking over tasks from humans, and searching limitations of these interpretations. We suggest AI to performing them autonomously. Recent reports indicate that an in- have a secondary role during the exploratory stage as humans have creasing percentage of human labor has been replaced either par- an edge due to their (relative) advanced capabilities of intuition, tially or completely by new technologies in the last decade (Arntz adaptive decision-making and their ability to assess effectiveness et al. 2016; Brynjolfsson et al. 2018), and people who hold jobs that across contexts. Put differently, even though collaboration with AI are easy to automate (e.g., frontline service employees)– that cor- can enhance the overall effectiveness of exploratory processes, we responds to 30 million American workers– are expected to continue propose that AI is best to play an assistive role during this phase of being replaced (Muro et al. 2019). creative cognition. In addition to their adoption in easy-to-standardize business We believe that our work builds on the current comprehension processes, many companies have been utilizing artificial intelligence of human creativity and proposes a framework that captures the dy- (AI) technologies in business areas that are fundamentally driven namics of the relationship between AI and human creative processes. by high-order human cognition, such as creativity. Due to the wide- By theorizing how humans and AI can collaborate instead of replac- spread adoption of novel AI technologies, creativity is no longer a ing one another, we also contribute to literature on technology by process that is single-handedly achieved by humans. Today, many offering an alternative approach to studying a dark future that awaits companies (e.g., Google, IBM) increasingly employ AI in their cre- humanity (Castelo and Lehmann 2019; Granulo et al. 2019). Addi- ative processes (Ringel et al. 2019) to generate and execute ideas tionally, our work offers timely and actionable strategies for man- (e.g., medicine, food; Fry 2018), to design clothes (e.g., Glitch), and agers and creative professionals. We provide understanding on how even to compose songs (e.g., Watson Beat). This widespread adop- the integration of AI can impact the creative process and facilitate tion of AI technologies in tasks that require higher-order cognitive creative outcomes across different stages of the creative process. By Advances in Consumer Research (Volume 48) / 1169 doing so, we provide managers with insights into defining and de- Kim, Tae Woo and Adam Duhachek(2020), “The Impact of signing more successful creative processes (e.g., new product devel- Artificial Agents on Persuasion: A Construal Level Account, opment), and how to determine the effective distribution of tasks and Psychological Science, 31 (4), 363-380. resources shared between AI and humans. Logg, Jennifer M, Julia A Minson, and Don A Moore (2019), “Algorithm Appreciation: People Prefer Algorithmic to Human REFERENCES Judgment,” Organizational behavior and human decision Alter, Adam L., Daniel M. Oppenheimer, D. M., and Jeffrey C. processes, 151, 90-103. Zemla. (2010), „Missing the Trees for the Forest: A Construal Longoni, Chiara, Andrea Bonezzi, and Carey K Morewedge (2019), Level Account of the Illusion of Explanatory Depth. Journal “Resistance to Medical Artificial Intelligence,” Journal of of Personality and Social Psychology, 99, 436–51. Consumer Research, 46 (4), 629-50. Arntz, Melanie, Terry Gregory, and Ulrich Zierahn (2016), Miller, George A. (1956), “The magical number seven, plus or “The Risk of Automation for Jobs in OECD Countries: A minus two: Some limits on our capacity for processing Comparative Analysis,” OECD Social, Employment and information,” Psychological Review, 63 (2), 81. Migration Working Papers, No. 189, OECD Publishing, Paris, Moreau, C. Page, and Darren W. Dahl (2005), “Designing the http://dx.doi.org/10.1787/5jlz9h56dvq7-en solution: The impact of constraints on consumers’ creativity,” Baddeley, Alan (2013), Essentials of Human Memory (classic Journal of Consumer Research, 32 (1), 13-22. edition), Psychology Press. Muro, Mark, Robert Maxim, and Jacob Whiton (2019), Brynjolfsson, Erik, Tom Mitchell, and Daniel Rock (2018), “What “Automation and Artificial Intelligence: How Machines are can machines learn, and what does it mean for occupations and Affecting People and Places,”Metropolitan Policy Program at the economy?,” in AEA Papers and Proceedings, (108), 43. Brookings. Castelo, Noah, and Donald R. Lehmann (2019), “Be Careful What Murphy, Gregory L. (1988), “Comprehending complex concepts,” You Wish For: Unintended Consequences of Increasing Cognitive Science, 12 (4), 529-562. Reliance on Technology,” Journal of Marketing Behavior, 4 Perkins, David N. (1981), The Mind’s Best Work, Cambridge, MA: (1), 31. Harvard University Press. Castelo, Noah, Maarten W. Bos, and Donald R. Lehmann (2019), Pew Research (2018), “Public Attitudes toward Computer “Task-Dependent Algorithm Aversion,” Journal of Marketing Algorithms.” Research, 56 (6), 809-25. Ringel, Michael, Florian Grass, Michael Spira, Derek Kennedy, Dietvorst, Berkeley J, Joseph P Simmons, and Cade Massey (2015), and Ramon Baeza (2019), “AI Powers a New Innovation “Algorithm Aversion: People Erroneously Avoid Algorithms Machine” Most Innovative Companies 2019,” Available at: after Seeing Them Err,” Journal of experimental psychology: https://www.bcg.com/publications/2019/most-innovative- General, 144 (1), 114. companies-artificial-intelligence-innovation-machine.aspx Erlandsson, Arvid, Fredrik Björklund, and Martin Bäckström Rozenblit, Leonid and Frank Keil, (2002), “The misunderstood (2015), “Emotional Reactions, Perceived Impact and limits of folk science: An illusion of explanatory depth,” Perceived Responsibility Mediate the Identifiable Victim Cognitive Science, 26, 521–562. Effect, Proportion Dominance Effect and in-Group Effect Vallacher, Robin R., and Daniel M. Wegner (1989), “Levels Respectively,” Organizational behavior and human decision of Personal Agency: Individual Variation in Action processes, 127, 1-14. Identification,”Journal of Personality and Social Psychology, Finke, Ronald A., Thomas B. Ward, and Steven M. Smith (1992), 57 (4), 660-71. Creative cognition: Theory, research, and applications. Van Bruggen, Gerrit H., Ale Smidts, and Berend Wierenga (1998), Freitas, Antonio L., Peter Gollwitzer, and Yaacov Trope (2004), “Improving decision making by means of a marketing decision “The Influence of Abstract and Concrete Mind-sets on support system,” Management Science, 44 (5), 645-658. Anticipating and Guiding Others’ Self-regulatory Efforts,” Yeomans, Michael, Anuj Shah, Sendil Mullainathan, and Jon Journal of Experimental Social Psychology, 40, 739–52. Kleinberg (2019), “Making Sense of Recommendations,” Fry, Hannah (2018), Hello World: How to be Human in the Age of Journal of Behavioral Decision Making, 32 (4), 403-14. the Machine, Random House. Granulo, Armin, Christoph Fuchs, and Stefano Puntoni (2019), “Psychological Reactions to Human versus Robotic Job Replacement,” Nature Human Behaviour, 3 (10), 1062-9. Discrimination and Inclusiveness in the Marketplace Chair: Martina Cossu (Bocconi University)

Paper #1: Belief-based Discrimination: Beauty Premium and and Norton investigate how a company’s gender pay gap affects Beauty Penalty consumers’ willingness to pay. When firms’ gender pay gaps are re- Meng Zhu, Johns Hopkins University, USA vealed, consumers are found to be less willing to pay for their goods, Tingting Nian, University of California at Irvine, USA an effect driven by consumer perceptions of unfairness and moder- Joachim Vosgerau, Bocconi University, Italy ated by gender. Finally, Cossu, Estes and Vosgerau show that consumers are Paper #2: Racial Discrimination in Customer Service: A Field willing to reward companies for their inclusiveness, specifically Experiment for using disabled models in their advertisements. While disabled Alexandra Feldberg, Harvard Business School, USA models to some extent evoke negative emotions (e.g., pity), they are Tami Kim, University of Virginia, Darden School of Business, shown to more strongly trigger positive emotions such as admiration USA that transcend to the advertisement and the brand. These positive ef- Paper #3: Paying for Parity: Consumer Response to Gender Pay fects cannot be accounted for by social desirability alone. Gaps We expect the session to generate significant interest among Bhavya Mohan, University of San Francisco, USA conference attendees in general given the timeliness of the research Tobias Schlager, University of St. Gallen, Switzerland topic; it should specifically attract researchers interested in inclusive- Katy DeCelles, University of Toronto, Canada ness, discrimination, and minorities. Michael Norton, Harvard Business School, USA Paper #4: Uncommon Beauty: Physically Disabled Models Belief-based Discrimination: Beauty Premium and Positively Affect Consumers’ Attitudes and Choices Beauty Penalty Martina Cossu, Bocconi University, Italy Zachary Estes, Bocconi University, Italy EXTENDED ABSTRACT Joachim Vosgerau, Bocconi University, Italy The positive discrimination in favor of attractive workers—the so-called beauty premium—is a robust phenomenon. Explanations SESSION OVERVIEW for the beauty premium fall into three classes: taste-based accounts The public debate about discrimination and inclusiveness is in- (Becker, 1957) which maintain that beauty is in itself valued by con- tensifying in our societies. Social media is proliferating the means sumers, statistical-based explanations (Phelps, 1972; Arrow, 1973) to socially connect, thereby exposing consumers to more and more which argue that beauty serves as a reliable indicator for produc- information on different cultures and giving voice to vulnerable mi- tivity, and belief-based explanations (Mobius and Rosenblat 2005; norities. The #metoo movement has made it possible to openly talk Bohren, Imas & Rosenberg 2019) which suggest that the beauty pre- about sexual discrimination at the workplace, a topic that had been mium is caused by consumers’ mistaken beliefs about the relation- taboo just a few years ago. And the notion of discrimination itself is ship between beauty and productivity. changing and no longer covers only race and gender but is extended We conjecture that both beauty premium and beauty penalty to other dimensions such as body image and beauty. Given these en- can occur due to belief-based discrimination. The psychology litera- couraging developments, one may be optimistic about the market- ture has shown that attractive people are perceived as more sociable, place overall becoming more inclusive. This special topic session is dominant, sexually warm, mentally healthy, interpersonally compe- aimed at providing an insight as to how inclusive the marketplace has tent, confident, and better adjusted than unattractive people (Dion, become. The evidence is mixed. While consumers and service pro- Berscheid & Walster, 1972; Feingold 1992; Langlois et al., 2000). viders are found to still be discriminating in certain areas, consumers Based on this validated positive link between beauty and social are also observed to be willing to punish companies that discriminate skills, we conjecture that consumers positively discriminate in favor and to reward those that promote inclusiveness. of the attractive for activities in which social skills are important. In In particular Zhu, Nian and Vosgerau show that discrimina- contrast, a beauty penalty may occur for activities that require analyt- tion based on beauty―a beauty premium or a beauty penalty―can ical skills and extensive solitary training, because consumers believe occur because consumers hold incorrect beliefs about the relation- that less attractive individuals incur higher cost in social interactions ship between beauty and productivity. Consumers discriminate in than attractive individuals. favor of the attractive when social skills are deemed important, but In Study 1 (preregistered https://aspredicted.org/blind. discriminate against the attractive when analytical skills are deemed php?x=mf7pe7), 5704 MTurkers each rated a set of 30 photos of the important. same gender and race along one of the following dimensions: attrac- Feldberg and Kim investigate discrimination in the service tiveness, intelligence, responsibility, trustworthiness, social skills, sector in a national field experiment involving 6,000 hotels. Hotels analytical skills, and age. The set of 30 photos for each respondent received emails from one of twelve fictitious email accounts (vary- was created out of a total of 8 sets with varying gender (male vs. fe- ing race, gender, and education) asking for local restaurant recom- male) and race (Asian, African Americans, Hispanics vs. Caucasians; mendations. Hotel representatives’ email responses revealed racial photos were sampled from the Chicago Face Database: https://facul- discrimination along three dimensions of service quality: responsive- ty.chicagobooth.edu/bernd.wittenbrink/cfd/index.html). For each of ness, helpfulness, and rapport. the 240 photos, we averaged the standardized ratings across respon- While those first two papers document extant discrimination in dents within dimensions (attractiveness, intelligence, responsibility, the marketplace, the next two papers show that consumers are also trustworthiness, social skills, analytical skills, age). With photo as willing to punish companies that are still discriminating and to re- the unit of analysis, we then regressed the averaged and standardized ward those that promote inclusiveness. Mohan, Schlager, DeCelles social and analytical skill ratings onto the averaged and standardized

Advances in Consumer Research 1170 Volume 48, ©2020 Advances in Consumer Research (Volume 48) / 1171 ratings of attractiveness, intelligence, responsibility, trustworthiness, tels, investigating the manner in which their service representatives and age in two separate regressions. Consistent with our prediction, would treat potential customers based on perceived race, gender, and people (photos) perceived as more attractive were perceived to have class of customers. Through a simple email inquiry, we examined better social skills (standardized b = 0.53, t(235) = 10.76, p <.001) how much knowledge about restaurants—and the way in which— but worse analytical skills (standardized b = -0.26, t(235) = -5.95, p hotel service representatives were willing to share by asking whether < .001). hotel staff had recommendations for “unique” dining experiences in In Study 2 (N = 205), we manipulated the relative importance their cities. of social skills to analytical skills within the same profession (i.e., Prior to beginning data collection, we set the sample size for computer science professors). Half of the participants were asked to this experiment based on previous audit (e.g., Bertrand & Mullaina- predict the professors’ teaching performance (a domain where social than, 2004; Milkman et al., 2015; Tilcsik, 2011) and aimed for 500 skills are more important), and the other half were asked to predict hotels per condition. We randomly selected 6,000 hotels from OR- their research performance (a domain where analytical skills are BIS, a global company database, using a random number generator. more important). Consistent with our prediction, professors (photos) The experiment employed 12 treatment groups based on the perceived as more attractive by one set of subjects were rated to be perceived race (white, black, Asian), gender (male, female), and better teachers by another set (β = 0.219, SE = .059; t(35) = 3.71, p class of customers. Based on previous research (Milkman et al., < .001) but worse researchers by a third set (β = -0.294, SE = .065; 2012, 2015), we generated six different names to signal both the t(35) = -4.49, p < .001). race and gender associated with each name. To vary perceptions of Can such beauty penalties also be observed in the marketplace? class, we either included or excluded “MD/PhD” at the end of each To address this question, we investigated the influence of physical name. We validated the list of names and titles through a pretest on attractiveness on users’ evaluations of others’ on an online question- Amazon’s Mechanical Turk. We emailed hotels from one of 12 ficti- and-answer site for professional programmers, Stack Overflow. We tious accounts with a simple inquiry. The emails were identical ex- randomly sampled 126,573 profile images, and using Face ++ API, cept for two components: (1) the city that the customer planned to we calculated a physical attractiveness score for each image. Based visit (which corresponded to the city of the recipient hotel) and (2) on the belief-based discrimination account by Bohren, Imas and the customer’s signature at the end of the email. We worded emails Rosenberg (2019), we hypothesized that new programmers with low as follows: “Hello, I will be visiting [hotel’s city location inserted reputation scores should incur a beauty penalty because for program- here] in a few weeks. I’m interested in dining experiences unique ming, analytical skills are deemed more important than social skills. to your city—could you make some recommendations? Thank you, In contrast, among programmers with high reputation scores, attrac- [Customer’s signature inserted here].” Using an email help tool, we tive programmers must produce higher quality output than unattract- scheduled all emails to be sent at 10am in the time zone correspond- ive programmers to overcome the initial beauty penalty, resulting in ing to each hotel’s location. a beauty premium. Consistent with these predictions, among (new) We assessed information availability and quality on the follow-reputation programmers, the questions/answers of attractive lowing three dimensions: responsiveness, helpfulness, and rapport. programmers received more downvotes and fewer upvotes (a beauty Responsiveness refers to the accessibility of information from ho- penalty). In contrast, among experienced high-reputation program- tel representatives, which we measured in two ways: whether or not mers, the questions/answers of attractive programmers received each hotel responded and email length (i.e., the number of characters more upvotes than downvotes (a beauty premium). in each email). Helpfulness evaluates how directly and extensively The results of our three studies suggest that discrimination representatives addressed the customer’s inquiry, which we mea- based on beauty is belief-based, and hence context dependent. Con- sured in three ways: (1) number of restaurants recommended, (2) sumers may discriminate in favor of attractive individuals when so- whether or not hotels referred the customer to another source instead cial skills are deemed important, but may discriminate in favor of the of making any restaurant recommendations, and (3) whether or not plain-looking when analytical skills are more important. Likewise, hotels inquired after the customer’s dining preferences. Note that the discrimination against the attractive can switch to discrimination in act of referring to another source is not helpful, given that respond- favor of the attractive when objective performance criteria are pres- ers are not directly answering the customer’s question. Conversely, ent. inquiring about one’s restaurant preferences is helpful, since it represents an attempt to customize the content of a response. Rapport Racial Discrimination in Customer Service: A Field considers the extent to which representatives attempted to establish a Experiment relationship with the customer, for example, by conveying personal engagement and providing unsolicited information that goes beyond EXTENDED ABSTRACT responding to the inquiry. Finally, five variables measured efforts to Service interactions between service providers and customers establish rapport: whether or not hotels (1) provided extraneous in- are ubiquitous in everyday life; they are also rife with service pro- formation not about restaurants (e.g., attractions in the city, history viders’ discretion and subjective assessments and may be especially of the city), (2) acknowledged the customer as a guest, (3) encour- susceptible to discrimination. Yet, much of the existing research on aged the customer to stay at the hotel, (4) addressed the customer us- discrimination has tended to focus on formal transactions—such ing his/her name or honorific in the response (e.g., Dr. Washington, as the buying and selling of goods and services, or the interactions Mr. Andersen) and (5) signed the email with a complimentary close with organized bureaucracies. Here we study discrimination in the (e.g., Best, Sincerely). context of customer service—the manner in which service providers Patterns across these variables broadly reflect a white advan- (i.e., frontline employees) assist customers. tage in customer service. Compared to customers with white names We chose the United States hotel industry, which generated (42.9%), those with black names and Asian names were less likely over $170 billion in sales revenue and employed 1.9 million people to receive a response (Black: 39.7%, χ2(1)=3.91, p< 0.05; Asian: in 2015 (AHLA, 2015), as our experimental setting. We conducted a 32.0%, χ2(1)=16.29, p<0.001). We found consistent results of racial field experiment using the audit study methodology with 6,000 ho- discrimination for helpfulness and rapport. In terms of helpfulness, 1172 / Discrimination and Inclusiveness in the Marketplace compared to customers with white names (M=3.35, SD=3.72), cus- in a total of 93,627 tweets referencing the 158 firms in our dataset. tomers with black names and Asian names received significantly Next, we used the LIWC database to code all the tweets for positive fewer restaurant recommendations (MBlack=2.97, SD=3.65, p=0.05; versus negative emotional valence about the organization (Pfarrer et

MAsian=3.00, SD=4.16, p=0.10). Customers with black and Asian al. 2010). names were also more likely to be referred to another source (Black: We conducted a linear mixed model regression, using the gen- 14.0%, χ2(1)=3.84, p=0.05; Asian: 15.9%, χ2(1)=8.64, p<0.01) than der pay gap as the independent variable, publication date (before vs. customers with white names (10.7%). Furthermore, customers with after publication) as the moderator, and consumer sentiment as de- black and Asian names were less likely to receive preference inqui- pendent variable (Bpay gap = .001, t(1) = -.19, p = .31, Bpublicationdate_after = 2 2 ries (Black: 3.8%, χ (1)=10.25, p<0.01; Asian: 3.3%, χ (1)=13.34, .14, t(1) = 8.65, p = .003, Bpay gap×publicationdate_after = -.01, t(1) = -8.20, p p<0.01) than those with white names (7.9%). In terms of rapport, < .001; SDrandomintercept_company = .41). The negative interaction between compared to customers with white names (28%), customers with pay gap and publication date shows that after disclosure, a larger black and Asian names were less likely to receive extraneous in- gender pay gap led to more negative emotional tweets referencing formation from the hotel in advance of their visit (Black: 16.0%, the company on Twitter (see Figure 1). A follow-up pilot study re- χ2(1)=29.1, p<0.001; Asian: 4.0%, χ2(1)=158.9, p< 0.001). They vealed that women responded significantly less positively to gender were also less likely to be encouraged to stay at the focal hotel wage gap disclosure online than men, as indicated by a lower sen- 2 2 (Black: 4.7%, χ (1)=4.06, p<0.05; Asian: 4.7%, χ (1)=3.52, p<0.10) timent measure (Mfemale = .43, Mmale = .80, t(339) = 2.65, p = .01, than white customers (7.2%), suggesting that hotels were being se- Cohen’s d = .29). lective about which individuals to invite as hotel guests. These re- We next designed an incentive compatible experimental para- sults were robust to the inclusion of hotel and hotel-county charac- digm to examine whether the disclosure of gender pay gaps caus- teristics as controls. There were neither main nor interaction effects ally affects consumers’ preferences for a firm’s goods. Participants of gender or education. (N = 501, Mage = 36.6, 51% male) were randomized to one of two Through an audit study of 6,000 hotels, we found that hotel rep- conditions: the control condition versus gender pay gap disclosure. resentatives provided service in greater quantities and with greater Regardless of condition, participants saw descriptions of two com- enthusiasm to white customers than nonwhite customers, regardless peting ride share platforms, Uber and Lyft. In the gender pay gap of their gender or education level. The study advances understand- present condition, participants were given the following additional ings of inequality and knowledge-sharing in the marketplace, high- information about Uber, based on actual self-reported data on the lighting the need to focus not only on the quality of customer service United Kingdom’s publicly available Gender Pay Gap Service por- but consistency—as well as the need to identify organizational in- tal: “According to a recent report, women working for Uber have an terventions. average hourly rate that is 33% lower than men’s. In other words, comparing average hourly rates, women earn 67 cents for every $1 Paying for Parity: Consumer Response to Gender Pay that men make.” Participants indicated whether, if they won, they Gaps would like a $50 gift card for Lyft versus for Uber. In this study, we also examined whether consumers responded differently to such EXTENDED ABSTRACT disclosure based on their self-identified gender. Participants of both The gender pay gap—when women earn less money relative to genders were significantly less likely to choose the Uber gift card men for the same work—is a pervasive pattern worldwide (Blau and when they saw the company’s disclosed gender pay gap relative to Kahn 2017). One recent estimate suggests that globally comparing 2 when they did not (MDisclosure = 48.2%, MControl = 71.8%; χ = 27.95, differences in average annual pay, women earn 57% of what men p < 0.01). However, we observed a significant interaction between make (Harris 2017). We explore the effects of gender pay gaps from gender and pay gap (B = 0.78, p = 0.04); women were less likely a unique perspective—that of consumers. We build on an emerging than men to choose the Uber gift card when they were shown that the area of research exploring the influence of inequality on consumer company has a gender pay gap. behavior (e.g., Ordabayeva and Chandon, 2011; Yoon and Kim, Next, we examined perceived wage fairness as a mediator of 2018). Our investigation of the impact of gender pay inequality is the effect of disclosing a gender pay gap on product desirability, particularly timely as countries move towards compelling firms to again using stimuli directly informed by actual self-reported data. publicly disclosure pay gap information, such as the UK (Alderman Participants (N = 304, MAge = 34.8, 57.7% male) were all given a 2018). description of the clothing retailer Adidas. In the gender pay gap Building on past research, we first propose that disclosing gen- present condition, participants were told: “Adidas recently reported der pay gaps can adversely affect consumer purchase decisions by women working for the company have an average hourly rate that is undermining consumers’ perceptions of wage fairness (Benedetti 18% lower than men’s.” Willingness to buy when Adidas’ pay gap and Chan 2018; Mohan, Schlager, Deshpandé and Norton 2018). was revealed was significantly lower than when the pay gap was not Moreover, we propose that participant gender moderates the impact revealed for both genders (Mdisclosure = 3.82, SD = 1.72 vs. Mno_disclosure of pay gap disclosure on product desirability, given that gender plays = 5.42, SD = 1.42; t(298) = 8.76, p < .01). However, women were a role in how an individual takes fairness in account when assessing significantly less willing to buy then men when a pay gap was re- the compensation of others (Leventhal and Lane 1970; Witt and Nye vealed (t(155) = 3.27, p < .01).We conducted a moderated mediation 1992). analysis, with gender pay gap disclosure as the independent variable, In our pilot study, we used Twitter data on consumer sentiment wage fairness as the mediator, gender as the moderator, and willing- before and after mandated pay gap disclosure in the United Kingdom ness to buy as the dependent variable.. The indirect effect of pay gap to provide an initial examination of whether pay gap disclosure influ- disclosure on willingness to guy via wage fairness was significant for ences consumer responses. A research assistant monitored compa- men (B = -.73, CI 95% [-1.05, -.45]), and women (B = -1.56, CI 95% nies that published their gender pay gap leading up to the mandated [-1.97, -1.21]; Preacher and Hayes 2008). Given that the confidence April 4, 2018 deadline. We obtained all tweets between March 28 intervals between both genders did not overlap, the effect is signifi- and April 8 (i.e., five days prior to and after April 3). which resulted cantly more negative for women. Advances in Consumer Research (Volume 48) / 1173

Two pilot studies and two experimental studies show that gen- d = .97), greater admiration for the model (Mdisabled = 4.66, SD = 1.10, der pay gap disclosure can affect consumer purchase intentions. This Mcontrol = 3.15, SD = 1.07; F(1, 199) = 95.40, p < .001, Cohen’s d area of inquiry is particularly timely and important given that firms = 1.38), and were judged as more novel (Mdisabled = 4.74, SD = .90, are under increasing pressure to disclose gender pay gap informa- Mcontrol = 3.38, SD = .94; F(1, 199) = 107.82, p < .001, Cohen’s d = tion that was previously kept confidential. Thus, consumer-facing 1.18). Disabled models enhanced participants’ attitudes (F(1, 199) = firms with gender pay gaps could directly lose equity as a result of 13.79, p < .001, Cohen’s d = .53) and increased their willingness to greater customer awareness, and particularly when their customers buy (Mdisabled = 3.19, SD = .97, Mcontrol = 2.66, SD = .96; F(1, 199) = are women. 14.99, p < .001, Cohen’s d = .55). Study 3 (N = 298; lab; preregistered) tested if previous results Uncommon Beauty: Physically Disabled Models were driven by social desirability bias, using a response deadline Positively Affect Consumers’ Attitudes and Choices procedure. If social desirability is acting, participants should have an immediate negative reaction to the ad and then subsequently adjust EXTENDED ABSTRACT their answer more positively. If this is the case, we would expect to Recent research examined how advertisements that promote see a positive effect only if it has enough time to emerge, such as body positivity and multiethnic inclusiveness impact consumers’ going from one to two seconds. Participants were exposed to both responses. This paper studies the impact of promoting diversity us- ads with disabled models and controls and were asked to evaluate ing a different minority (people with a disability) on a more general them (on a 3-point scale) within either 1 or 2 seconds (between- audience. participants). Again, ads with disabled models were evaluated more The present research analyzes how the inclusion of physically- positively in the one second condition (Mdisabled = 1.80, SD = .49, disabled models in ads influences consumers’ attitudes, emotional 2 Mcontrol = 1.63, SD = .34; F(1, 149) = 14.10, p < .001, η = .08), and responses, and choices. Negative attitudes toward people with dis- in the two seconds condition (Mdisabled = 1.86, SD = .50, Mcontrol = 1.60, abilities originate from the perpetuation of negative stereotypes. The SD = .37; F(1, 149) = 37.17, p < .001, η2 = .20). The effect did not most common, the paternalistic stereotype, portrays people with a interact with the response deadline (p =.139), suggesting the absence disability as high in warmth and low in competence (Fiske et al. of social desirability bias. 2002). Following this line of reasoning, we would expect physically- Study 4, a field experiment (N = 300, preregistered), tested if disabled models to have a negative impact on consumers’ attitudes previous results were driven by participants’ impression manage- and choices, because the stigma associated with the disability should ment by manipulating anonymity of choice. Participants were re- be transferred from the model to the brand. However, according to cruited in the gym of a European university. A research assistant social categorization theory (Kunda, Miller, and Claire 1999) a dis- approached gym members while they were entering the gym and abled model belongs to two social categories: first, a person with a offered an energy drink for free. Both drinks were advertised, the disability and hence part of a stigmatized minority, and second, as a ad for one drink pictured a disabled model, while the other pictured model, an admirable subgroup. When looking at a disabled person the control. Participants were brought into a room where they could that is also a model, people infer that s/he must have been through choose the drink. In the private condition, participants were left many difficulties. What previously was a negative attribute becomes alone while choosing. In the public condition, they communicated a signal of the model’s strength and determination. Therefore, ads their choice to the research assistant. Participants preferred the drink with disabled models shift the negative stereotypical view of peo- advertised by the disabled model whether they made their choice in ple with a disability and increase consumers’ attitudes. We tested private (60.1%) or in public (68.4%; both percentages are greater whether disabled models have a positive impact on consumers’ at- than 50%, z’s > 2.38, p’s < .001) and choice did not differ across titudes (study 1), and emotional responses (study 2). We further in- conditions (χ2 (1) = 2.24, p = .134). vestigated whether the findings are due to social desirability (study The present study contributed to the literature in advertising by 3). Finally, in a field experiment, we tested whether energy drinks considering a specific minority that has not been studied before. Our advertised with disabled models are preferred in private as well as studies suggest that disabled models enhance attitudes, evoke posi- public choices (study 4). tive emotional responses and increase choice both in private and in Study 1, a between-participants study (N = 159; M-Turk), test- public. ed whether ads with disabled models enhance attitudes. Participants were assigned to one of two conditions (disability vs control). We Belief-based Discrimination: Beauty Premium and created two versions of two ads, one with a model (control) and an- Beauty Penalty other identical ad except the same model was missing a limb (dis- Arrow, Kenneth. (1973), “The theory of discrimination. Ashenfelter ability). Compared to the control condition, ads with disabled mod- O, Rees” A ed. Discrimination in labor markets, Princeton els were rated more favorably (Mdisabled = 4.30, SD = 1.09, Mcontrol = University Press, Princeton, 3-33. 3.87, SD = .85; F(1, 153) = 7.32, p < .01, Cohen’s d = .43). Becker, Gary. (1957),” The Economics of Discrimination,” Study 2, a between-participants study (N = 200; lab), replicated University of Chicago Press, Chicago. the previous result and, in addition, tested the effect of disabled mod- Bohren, J. Aislinn, Alex Imas, and Michael Rosenberg. (2019),” els on consumers’ emotional responses, perception of novelty and The dynamics of discrimination: Theory and evidence,” willingness to buy. Participants were randomly assigned to one of the American Economic Review, 109 (October), 3395-3436. two conditions (disability vs control) and saw four advertisements. Dion, Karen, Ellen Berscheid, and Elaine Walster. (1972), “What

Disabled models evoked more pity (Mdisabled = 3.90, SD = 1.39, Mcontrol is beautiful is good,” Journal of Personality and Social = 3.25, SD = 1.30; F(1, 199) = 11.55, p < .001, Cohen’s d = .48), and Psychology, 24 (March), 285-290. to a lesser extent also more negative emotions (Mdisabled = 2.43, SD = Feingold, Alan. (1992), “Good-looking people are not what we

1.01, Mcontrol = 2.17, SD = 1.10; F(1, 200) = 3.10, p = .08, Cohen’s d think,” Psychological Bulletin, 111 (February), 304-341.

= .25). They evoked also more positive emotions (Mdisabled = 3.94, SD

= 1.03, Mcontrol = 2.99, SD = .93; F(1, 199) = 46.72, p < .001, Cohen’s 1174 / Discrimination and Inclusiveness in the Marketplace

Langlois, Judith H., Lisa Kalakanis, Adam J. Rubenstein, Andrea Leventhal, Gerald S., and Douglas W. Lane. (1970), “Sex, age, Larson, Monica Hallam, and Monica Smoot. (2000), “Maxims and equity behavior “Journal of Personality and Social or myths of beauty? A meta-analytic and theoretical review,” Psychology, 15 (April), 312-316. Psychological Bulletin, 126 (March), 390-423. Mohan, Bhavya, Tobias Schlager, Rohit Deshpandé, and Michael I. Mobius, Markus M., and Tanya S. Rosenblat. (2006), “Why beauty Norton. (2018), “Consumers Avoid Buying From Firms With matters,” American Economic Review, 96 (January), 222-235. Higher CEO‐to‐Worker Pay Ratios,” Journal of Consumer Phelps, Edmund S. “The statistical theory of racism and sexism,” Psychology, 28 (February), 344-352. (1972), American Economic Review, 62 (April), 659-661. Ordabayeva, Nailya and Pierre Chandon. (2011), “Getting ahead of the Joneses: When equality increases conspicuous Racial Discrimination in Customer Service: A Field consumption among bottom-tier consumers,” Journal of Experiment Consumer Research, 38 (June), 27-41. AHLA. 2015. Lodging Industry Trends 2015. Retrieved from: Pfarrer, Michael D., Timothy G. Pollock, and Violina P. Rindova. https://www.ahla.com/sites/default/files/Lodging_Industry_ (2010), “A tale of two assets: The effects of firm reputation Trends_2015.pdf and celebrity on earnings surprises and investors’ reactions,” Bertrand, Marianne, and Sendhil Mullainathan. (2004), “Are Academy of Management Journal, 53 (May), 1131-1152. Emily and Greg more employable than Lakisha and Jamal? A Preacher, Kristopher J., and Andrew F. Hayes. (2008), “Asymptotic field experiment on labor market discrimination,”American and resampling strategies for assessing and comparing indirect Economic Review, 94 (April), 991-1013. effects in multiple mediator models,”Behavior Research Milkman, Katherine L., Modupe Akinola, and Dolly Chugh. Methods, 40 (March), 879–91. (2012),”Temporal distance and discrimination: An audit study Witt, L. Alan, and Lendell G. Nye. (1992), “Gender and the in academia,” Psychological Science, 23 (June), 710-717. relationship between perceived fairness of pay or promotion Milkman, Katherine L., Modupe Akinola, and Dolly Chugh. and job satisfaction,” Journal of Applied Psychology, 77 (2015), “What happens before? A field experiment exploring (June), 910-917. how pay and representation differentially shape bias on the Yoon, Sunyee, and Hyeongmin Christian Kim. (2018), “Feeling pathway into organizations,” Journal of Applied Psychology, economically stuck: The effect of perceived economic mobility 100 (June), 1678-1712. and socioeconomic status on variety seeking,” Journal of Tilcsik, András. (2011), “Pride and prejudice: Employment Consumer Research, 44 (May), 1141–1156. discrimination against openly gay men in the United States,” American Journal of Sociology, 117 (February), 586-626. Uncommon Beauty: Physically Disabled Models Positively Affect Consumers’ Attitudes and Choices Paying for Parity: Consumer Response to Gender Pay Deshpandé, Rohit, and Douglas M. Stayman. (1994), “A tale of two Gaps cities: Distinctiveness theory and advertising effectiveness,” Alderman, L. (2018). Britain aims to close gender pay gap with Journal of Marketing Research, 31 (January), 57-64. transparency and Shame. The New York Times, [available at Fiske, Susan T., Amy JC Cuddy, Peter Glick, and Jun Xu. (2002),”A https://www.nytimes.com/2018/04/04/business/britain-gender- model of (often mixed) stereotype content: Competence pay-gap.html]. and warmth respectively follow from perceived status and Benedetti, Arianna H., and Serena Chen. (2018), “High CEO-to- competition,” Social Cognition, 82 (January), 878-902. worker pay ratios negatively impact consumer and employee Kunda, Ziva, Dale T. Miller, and Theresa Claire. (1990), perceptions of companies,” Journal of Experimental Social “Combining social concepts: The role of causal reasoning,” Psychology, 79 (November), 378-393. Cognitive Science, 14 (April), 551–577. Blau, Francine D., and Lawrence M. Kahn. (2017),” The gender wage gap: Extent, trends, and explanations,” Journal of Economic Literature, 55 (March), 789-865. Difficult by Design: Choice Difficulty and Effort in Decision Making Chairs: Elina Yewon Hur, Cornell University, USA Kaitlin Woolley, Cornell University, USA

Paper #1: The Objectivity Illusion of Ranking Procedures: How “matters of taste.” Lastly, Hur, Woolley, and Tu (N=1,425) examine and Why Ranking Alleviates Decision Difficulty how choice architecture can affect consumers’ subjective evaluations Yonat Zwebner, Interdisciplinary Center (IDC), Israel of the effort they put into a decision. When holding actual effort in- Rom Y. Schrift, Indiana University, USA vested constant, consumers perceive their effort had a bigger payoff, that is, was better spent and more worthwhile, when they discover Paper #2: Roads or Rome? How Product Categorization Shapes the best option at the end (vs. beginning) in a choice set. As a result, the Attributions and Consequences of Choice Difficulty even though consumers all chose the same best option, those who Xiang Wang, University of Florida, USA found this option at the end (vs. beginning) valued it more and were Aner Sela, University of Florida, USA more committed to their choice. Paper #3: Nuanced Effects of Decision Effort on Decision Together, this session advances our understanding of how Confidence in Matters of Quality versus Matters of Taste choice difficulty and objective and subjective effort investment af- Nahid Ibrahim, University of Alberta, Canada fect consumers’ decision processes and search behavior. This session Gerald Hӓubl, University of Alberta, Canada offers broad appeal on how choice structure can enhance or mitigate Paper #4: Hidden Benefits of Hiding the Best Option: Consumers perceived effort and the consequences of choice difficulty on con- Value Their Selection More When Discovering it Later sumers’ decisions. Elina Yewon Hur, Cornell University, USA Kaitlin Woolley, Cornell University, USA The Objectivity Illusion of Ranking Procedures: How Yanping Tu, University of Florida, USA and Why Ranking Alleviates Decision Difficulty

SESSION OVERVIEW EXTENDTED ABSTRACT How difficult was your last decision? How did this difficulty The degree to which consumers experience decision difficulty during the decision process affect your choice? Consumers often is an important and fundamental aspect of the decision-making pro- face difficult decisions (Tversky and Shafir 1992). Choice difficul- cess. For example, previous research demonstrated that experienced ty shapes, and is shaped by, effort invested in the decision process difficulty may elicit negative emotional states (e.g., Luce, Payne and (Dhar 1997; Iyengar and Lepper 2000). The proposed session joins Bettman 1999), cause consumers to end up choosing nothing (e.g., together four papers offering a new perspective on these questions, Iyengar and Lepper 2000), delay their choice (Dhar 1997, Tversky examining how features of the choice context and decision process and Shafir 1992), increase the need to justify the decision (Shafir, affect perceived choice difficulty, with implications for consumers’ Simonson, and Tversky 1993), and increase the likelihood to com- information search. They further examine perceived and objective promise (e.g., Novemsky et al. 2007). Additionally, the subjective effort invested in choosing, with outcomes for choice confidence and experience of decision difficulty has been shown to also impact con- valuation of their chosen product. sumers’ tendency to engage in simplifying and complicating behav- The first two papers examine novel factors that influence con- iors (e.g., Schrift, Netzer and Kivetz 2011, Schrift, Kivetz, Netzer sumers’ perception of how difficult a choice is.Zwebner and Schrift 2016) and could even play an important role when observing others (N=955) identify how preference-elicitation modes (choice vs. rank- making decisions (e.g., Schrift and Amar 2015). ordering) affect perceived choice difficulty. Although rank-ordering Indeed, as the subjective experience of difficulty impacts vari- requires more effort than simply choosing a preferred option, these ety of different behaviors, understanding what factors increase or de- authors found that rank-ordering reduces perceived choice difficulty. crease it, is important from both theoretical and applied perspectives. This effect occurs because rank-ordering makes judgments feel more The current work focuses on how the preference elicitation mode objective. Wang and Sela (N=1,224) investigate how product cat- impacts consumers’ experience of difficulty. In particular, we focus egorization influences attributions of choice difficulty, and implica- on two prominent modes, namely, choice and rank-ordering. tions for consumers’ information search behavior. When products are When comparing a choice task to a ranking task, it is quite in- framed in different categories, consumers perceive that the options tuitive to expect that a ranking task will elicit greater experience of represent distinct goals and therefore attribute choice difficulty to difficulty (supported by a pilot study (N=100) where 76% of people preference uncertainty (i.e., to the self - “I am not sure what I want”). believed that choosing would be easier for them than rank-ordering). In contrast, when products are framed under the same category, That is, while choosing includes the selection of one option that of- consumers see the options as representing means to an overarching fers the greatest utility, rank-ordering includes the same selection goal and attribute choice difficulty to preference matching (i.e., to (i.e. ranked first), as well as the ordering of all the other options in the option - “These options are not what I want”). This differential the set. Thus, the operations required to make a choice are embedded attribution of choice difficulty to the self (vs. options) drives search in those required to complete the ranking procedure. However, in behavior; people seek information to discover their own preferences, the current research we repeatedly find that consumers’ experienced but seek alternative choices when they feel the option is to blame. difficulty is actually lower when asked to rank as opposed to choose The remaining papers examine how consumers’ effort evalua- among the options. That is, even though the outcome of the ranking tion shapes preferences and choice. Ibrahim and Hӓubl (N=2,335) procedure is identical to that of the choice (i.e., consumers are fully examine when and why exerting greater effort increases decision aware that the option they rank first is the one they will eventually confidence. Specifically, when consumers put more effort (e.g., time) receive), framing the process as ranking as opposed to choosing de- into the decision process, they feel less confident about decisions that creases the difficulty that consumers experience. are “matters of quality” but more confident about decisions that are

Advances in Consumer Research 1175 Volume 48, ©2020 1176 / Difficult by Design: Choice Difficulty and Effort in Decision Making Following recent research on consumer’s perceived objectivity Roads or Rome? How Product Categorization Shapes the vs. subjectivity in judgement (Berman et al., 2018; Goodwin and Attributions and Consequences of Choice Difficulty Darley 2008; Spiller and Belogolova 2017) we find that ranking compared to choosing increases perceptions of objectivity in judg- EXTENDED ABSTRACT ment and, therefore, reduces the associated decision difficulty. That Consumers often experience choice difficulty while making de- is, while choosing highlights subjective evaluation criteria (i.e., a cisions (Broniarczyk and Griffin 2014; Sela and Berger 2012), and matter of taste), ranking procedures increases the perceived objec- they often cope with such experiences by deferring choice and pro- tivity of the decision and eases the selection process. longing their information search (Corbin 1980; Dhar 1997; Dhar and Study 1 (N=217, lab), utilized an incentive-compatible para- Simonson 2003). But how do people select the specific strategy for digm in which participants viewed five products and either chose coping with choice difficulty? For example, when do they seek more their most preferred product, or rank-ordered the products according information on existing alternatives as opposed to new alternatives, to their preferences. Across both conditions we informed participants or try to figure out their own preferences? that their most preferred product (chosen or top-ranked) is the one Choice difficulty may be attributed internally, to preference un- they will receive should they win a lottery that was conducted. After certainty (e.g., “I am not sure what I want”), or externally, to quali- choosing or rank-ordering, participants rated the difficulty they ex- ties of the available options (e.g., “the available options are not good perienced using five items (Cronbach α=.88). Results confirmed that enough”). Each of these possible attributions has distinct implica- compared to the difficulty experienced in choice (M=3.42, SD=1.3) tions for information search, namely, seeking information about how ranking reduced the experienced difficulty (M=2.54, SD=1.02; the options differ in an attempt to form a preference, versus seeking t(215)=5.49, p<.001). Interestingly, this difference was despite the other, presumably better options, respectively. fact that participants spent more time ranking the options (p<.001). We argue that framing the choice options as representing dif- In Study 2 (N=199, MTurk) participants received only two op- ferent categories (e.g., “2-in-1”, “ultraportable”, “gaming”) leads tions and were asked to either choose or rank-order the two options consumers to construe the decision as choosing among different (between-subjects). Because the set consisted of only 2 options, the goals, whereas framing them as representing one overarching cat- actual task in both conditions, is virtually identical. Results con- egory (e.g., “laptops”) leads them to perceive the decision as choos- firmed that even when only two options were available, participants ing among different means to a goal. We argue this based on the still felt lower degree of difficulty in the rank (M=3.24, SD=1.49) notion that category labels often indicate the consumption goals compared to the choice condition (M=3.74, SD=1.5; t(197)=2.35; associated with products under the category (Lamberton and Diehl p=.02). Thus, these finding suggest that the effect is driven by how 2013). Goals are self-diagnostic, because they are seen as desired people construe the task and not by the actual cognitive operations end-states of the self (Carver and Scheier 1990, 1998; Chartrand et required to performing it. al. 2008; Sela and Shiv 2009), whereas means are more instrumen- In Study 3 (N=160, lab) we aimed to further examine why con- tal and situationally determined (Kruglanski et al. 2002; van Osse- sumers construe the same decision differently when choosing com- laer and Janiszewski 2012). Consequently, we propose that, to the pared to ranking. After either choosing or ranking, participants were extent that consumers experience choice difficulty, distinct product asked to indicate (on a 1-7 scale) the extent to which they found categories lead consumers to attribute choice difficulty to the self (as the (i) selection of a TV and (ii) differences among TVs, to be a opposed to the available options) and, therefore, seek information matter of taste or a matter of quality. As hypothesized, participants that may help them discover their own preferences. Conversely, a in the ranking condition indicated greater objectivity for their se- single product category frame leads them to attribute difficulty to the lected TV (M=5.29, SD=1.34) as well as for the differences among options available (as opposed to the self) and, therefore, seek other TVs (M=5.13, SD=1.33) compared to those in the choice condition choice alternatives. (M=4.53, SD=1.74; t(158)=3.12, p=.002, and M=4.53, SD=1.53; Experiment 1 examines the effect of categorization on informa- t(158)=2.66, p=.009, respectively). Further, 68.8% in the ranking tion search, using consequential choice. Participants saw three tea condition indicated that their chosen option is objectively better bag options from which to choose, each described on several attri- compared to only 51.2% in the choice condition (χ2(1)=5.1, p=.024). butes, framed as representing either the same category or different Study 4 (N=279, lab) replicated the effect observed in Stud- categories. Product attributes were the same across conditions and ies 1-2 (p<.001) and the perceived objectivity effect demonstrated in category labels provided no additional information not already con- Study 3 (p=.026) and extended these findings by testing the mediat- tained in the attributes. Presumably to help them with the decision, ing role of perceived objectivity on experienced difficulty. Employ- we offered participants the option to view either additional informa- ing a similar design to Study 3, a mediation analysis (using the macro tion about the existing options that can help them figure out what PROCESS, model 4, Hayes, 2013) confirmed that the perceived ob- type of tea bag they want or, instead, to view other options alto- jectivity mediated the effect of elicitation mode (choice vs. rank) on gether, corresponding to preference uncertainty and option quality participants’ experience of difficulty (B=-.09; 95% CI=[-.19, -.02]). attributions, respectively. They also rated their subjective feelings To summarize, we find that ranking, although takes longer and of choice difficulty. Participants were more likely to seek detailed requires greater cognitive operations, reduces experienced decision information about the existing options in the different-categories difficulty as it fosters greater perceived objectivity of the selection condition (M=70%) than in the same-category condition (M=47%; process. We discuss the applied contribution of these findings given χ2(1)=7.08, p=.008). Supporting our suggestion that this effect re- the prevalence of decision aids that could foster ranking procedures flects attributions of choice difficulty, it was pronounced when diffi- as opposed to direct choice. culty was moderate-to-high (greater than 3.13 on a 7-point scale) but attenuated when participants experienced no difficulty (moderation β=.57, Wald=4.43, p=.035). Categorization had no effect on difficulty itself. Experiment 2 examines how product categorization influences attributions of choice difficulty. Participants saw three laptop mod- Advances in Consumer Research (Volume 48) / 1177 els, framed as representing either one category (“laptops”) or differ- that exerting more mental effort increases metacognitive difficulty, ent categories (“ultraportable”, “performance”, “premium display”). resulting in low decision confidence (e.g., Dhar 1997; Iyengar and Participants rated choice difficulty and indicated whether the diffi- Lepper 2000; Tversky and Shafir 1992; see Anderson 2003 for a re- culty was due to preference uncertainty (e.g., “I am not sure what view), we theorize that, under some circumstances, exerting greater exactly I want”) or the options (e.g., “the options do not perfectly fit effort actually increases decision confidence. my wants”). As expected, participants were more likely to attribute We argue that whether exerting more effort in the decision pro- choice difficulty to preference uncertainty in the different-categories cess increases or decreases decision confidence (Parker et al. 2016; condition (M=4.22 on a 7-point scale: 1=attribute to preference un- Thomas and Menon 2007; Zakay 1985) is a function of consum- certainty, 7=attribute to option availability) than in the same-catego- ers’ effort sensitivity (the strength of the relationship between effort ry condition (M=4.99; F(1,98)=9.40, p=.003). There was no effect exertion and consumers’ judgment of metacognitive difficulty) in a on choice difficulty itself or perceived option similarity, variability, particular domain and its impact on the inferences consumers draw and comparability. from their effort exertion. Experiment 3 tests the mediating role of participants’ percep- We theorize that consumers tend to be more effort sensitive tion that they chose among goals versus means. Participants saw four when they consider effort exertion to be detrimental in achieving the mutual fund options, framed as representing either one or different desired decision outcome and less effort sensitive when they con- categories. Compared with the same-category condition, participants sider effort exertion to be beneficial in achieving the desired decision in the different-categories condition attributed difficulty to- prefer outcome. We demonstrate that consumers are systematically more ence uncertainty more (F(1, 195)=9.07, p=.003) and indicated that effort sensitive in domains that are considered “matters of quality” they thought more during the decision about which financial goal (where alternatives can be rank-ordered based on product features they would like to achieve rather than which option would best reflecting objective quality) than in domains considered “matters of serve their financial goals (F(1, 195)=3.90, p=.050). This measure of taste” (where alternatives cannot be rank-ordered in terms of objec- thought process mediated the effect of categorization on attribution tive quality). This divergence in effort sensitivity in turn differen-

(βindirect=.10, 95% CI=[.01,.26]) tially impacts two distinct aspects of decision confidence – prefer- Experiment 4 uses the same scenario to provide further insight ence clarity (i.e., to what extent consumers’ choices reflect their true into the process. If using multiple (vs. single) product categories af- inclination) and preference correctness (i.e., to what extent choices fects attributions by leading people to perceive that they are choos- can be validated or justified by normative preferences) (Olson et al. ing among goals (vs. means), then drawing participants’ attention, in 1983; Petrocelli et al. 2007). both conditions, to their personal goals before the decision should In quality domains, consumers tend to attribute the locus of attenuate this effect, because it should increase goal-based choice their evaluations to product features and consider normative prefer- regardless of product categorization. Consistent with our logic, the ences for these features (i.e., rank-ordering) to be diagnostic of the effect of product categorization on choice difficulty attributions rep- optimality of their own choices. In this case, exerting more (vs. less) licated in the baseline condition (F(1,690)=8.64, p=.003), but was effort in the decision processundermines inferences about preference attenuated when participants answered several questions regarding correctness, in turn diminishing decision confidence. By contrast, in their financial goals before seeing the four mutual fund options (F(1, taste domains, consumers tend to attribute the locus of their evalua- 690) < 1, NS; interaction F(1,690)=5.00, p=.026). Choice difficulty tions to personal preferences and consider normative preferences for did not differ across the four conditions. product features to be less diagnostic of the optimality of their own Experiment 5 further examines the downstream effect of attri- choices. In this case, exerting more (vs. less) effort in the decision butions on information search. Participants saw three wine options, process promotes inferences about preference clarity, in turn increas- framed as representing either one category (i.e., dinner party wine) or ing decision confidence. We present evidence from five studies that different categories (i.e., pre-dinner wine, table wine, dessert wine). were designed to test this conceptual model (see figure 1). Compared with the same-category condition, participants in the dif- In study 1 (N=245), we tested consumers’ predisposition toward ferent-categories condition attributed choice difficulty to preference exerting effort across 16 different product domains (adapted from uncertainty more (F(1,98)=5.35, p=.023). This, in turn, led them to Spiller and Belogolova 2017). We found that, on average, consumers indicate that they would prefer to view additional information that are willing to exert more effort (i.e., actively spend time; log-trans- could help them discover what they wanted rather than see other op- formed) in the decision process when they consider a domain to be 2 tions (72% vs. 52%; χ (1)=4.24, p=.039; mediation βindirect=.21, 95% more of a matter of quality versus more of a matter of taste (b=.086, CI=[.01,.69]). SE=.004, p<.001; see figure 2). This finding is consistent with our Taken together, our findings provide novel insights on how theorizing but not fully conclusive, particularly since the 16 domains product categorization influences consumer interpretations of choice differed considerably in economic significance (e.g., Chaiken and difficulty and information search. In addition to their practical sig- Maheswaran 1994; Petty and Wegener 1998). We controlled for this nificance, the findings also make theoretical implications for goal in subsequent studies by using quality and taste domains with similar systems and self-perception theories by showing the link between price ranges. Using product-choice paradigms, we manipulated the categorization, goal construal, and attribution to internal versus ex- amount of decision effort exerted by varying aspects of the choice ternal sources of choice experiences. architecture, such as the number of alternatives to be inspected, as well as the format and timing of information presentation. Nuanced Effects of Decision Effort on Decision Both studies 2 (N=488) and 3 (N=503) showed that, as hy- Confidence in Matters of Quality versus Matters of Taste pothesized, greater effort exertion reduced decision confidence in quality domains (e.g., electric toothbrushes) but increased decision EXTENDED ABSTRACT confidence in taste domains (e.g., coffee tables) (ps<.05). We esti- Effort, be it physical or mental, is quintessential to everyday mated a bias-corrected moderated mediation model (Hayes 2013; consumer decision-making, and it has a profound impact on consum- 5,000 bootstrap samples) in each study. Our analyses revealed that ers’ preference and choice appraisals. Although evidence suggests in quality domains, inferences about preference clarity and prefer- 1178 / Difficult by Design: Choice Difficulty and Effort in Decision Making ence correctness were interdependent and corresponded to decision The current research explores a hidden benefit of saving the effort similarly. In this case, greater effort exertion diminished both best option late: Presenting the best option at the end (vs. begin- preference clarity and preference correctness, in turn reducing deci- ning) increases consumers’ valuation of the selected option. This is sion confidence (ps<.05). By contrast, in taste domains, inferences because, when searching within a limited choice set, discovering the about preference clarity and preference correctness were indepen- best option at the end (vs. beginning) leads consumers to perceive dent and corresponded to decision effort distinctively. In this case, that the effort they put into finding the best option paid off more – greater effort exertion reduced preference correctness but enhanced that is, was well spent and worthwhile (Delleart and Hӓubl 2012). preference clarity, in turn increasing decision confidence (ps<.001). We propose that such benefits arise because consumers perceive that These effects were robust to chronic individual differences in effort they have made more progress in their search (Harkin et al. 2016). valuation (i.e., need for cognition, α=.96; Cacioppo and Petty 1982; As such, finding the best option at the end (vs. beginning) has key Cacioppo et al. 1984). consequences for consumers’ enhanced valuation of their chosen op- In study 4 (N=499), we tightly controlled for the hedonic versus tion, as the positive experience of the search process transfers to the utilitarian nature of the decision domain by asking participants to outcome of search (Fishbach et al. 2004). choose within a common product domain (i.e., electric toothbrush- Five studies (N=1,325) test our theory. Study 1 examined effort es). We manipulated the set of alternatives so as to make the choice payoff using an incentive-compatible design. Participants opened among them either a matter of quality (products varied in objective five gift boxes containing different financial payouts with the goal quality and price, but not in their ergonomic and aesthetic properties) of discovering the highest prize to claim as a bonus. Participants in or a matter of taste (products varied in their ergonomic and aesthetic the best-option-beginning condition viewed the highest prize first properties, but not in objective quality or price). The results showed ($0.22) before viewing four lower prize amounts; participants in the again that greater effort exertion reduced confidence in quality do- best-option-end condition viewed the four lower prize amounts first mains but increased confidence in taste domains (p=.032; figures and found the highest prize ($0.22) at the end. After selecting their 3a–3d). prize, we measured how much participants felt their effort paid off. In study 5 (N=603), we distinguished between instrumen- As predicted, those who found the best option at the end (vs. begin- tal decision effort (the effort exerted in considering and reasoning ning) felt their effort paid off more (Mend=5.03; Mbeginning=4.05; t(136) about products: 20 vs. 5 alternatives) and incidental decision ef- =3.75, p<.001, d=.64). fort (the effort exerted in merely obtaining product information: 40 Study 2 extended these results when searching for a rental on vs. 10 seconds). Using a 3 (decision effort: low instrumental–high Airbnb. Participants searched through five different rentals before incidental vs. low instrumental–low incidental vs. high instrumen- making their selection. We manipulated the order of the highest tal–low incidental) x 2 (domain: quality vs. taste) between-subjects rated option so that it appeared either at the beginning or end of the product-choice paradigm, we again found that exerting more effort choice set. Participants who found the best option at the end (vs. diminished confidence in quality domains (i.e., power banks) but in- beginning) felt their effort paid off more (Mend=5.92; Mbeginning=5.00; 2 creased confidence in taste domains (i.e., ceramic mugs) (p=.071; F(1,161)=13.99, p<.001, ηp =.08). figures 4a–4d). Collapsing the levels of exerted decision effort to Study 3 examined our proposed process, that people feel a high versus low instrumental effort yielded the predicted interaction greater sense of effort payoff when finding the best option at the end effect (p=.020), but collapsing the levels of exerted decision effort because they experience a sense of progress during their search. To to high versus low incidental effort did not (p=.378), suggesting that test this, we manipulated the presentation order of the highest rated instrumental decision effort is the key driver of the predicted effects. restaurant as in study 2, and manipulated the sense of progress (pres- This research advances our understanding of the psychological ent vs. absent) by presenting options sequentially versus simulta- forces that govern how decision effort informs decision confidence. neously. In the progress-present-condition, participants viewed all Importantly, it pinpoints conditions under which choice architec- options one at a time, as in studies 1-2; in the progress-absent-condi- tures can benefit both firms and consumers by either promoting or tion, participants viewed all options at once, such that they were un- restraining the exertion of decision effort. able to feel a sense of progress during search. We found a significant 2 interaction (F(1, 340)=3.85, p=.051, ηp =.011; figure 5); finding the Hidden Benefits of Hiding the Best Option: Consumers best option at the end (vs. beginning) increased perceived effort pay-

Value Their Selection More When Discovering it Later off only for those who could experience search progress (Mend=6.23; 2 Mbeginning=5.60; F(1, 340)=11.00, p=.001, ηp =.03), which attenuated EXTENDED ABSTRACT when there was no search progress experience (p=.58). Consumers often suffer from choice overload. They at times Study 4 provided additional process evidence by measuring face endless options in stores and online that they need to sort consumers’ perceptions of progress during their search and we exam- through to find a product that fits their need. To address this, com- ined a consequence of this effect - that people value their selected op- panies have begun offering curated products that limit the search tion more when they feel greater effort payoff. Participants searched process and help consumers narrow in on the best option for them through a set of environmental non-profits with the goal of choosing (Cha et al. 2018). For example, online dating apps limit the number one charity to donate to. Participants who found the highest rated of potential relationship partners users can view and connect with non-profit at the end (vs. beginning) of their search felt greater ef- per day, and Book of the Month, a book subscription service, recom- fort payoff (Mend=5.39; Mbeginning=4.91; t(308) =2.81, p=.007, d=.31). mends five books every month with reviews, and consumers read Finding the best option at the end also increased valuation of the cho- each review from curators before ordering a book to read. sen option; these participants were more willing to recommend their When searching through a limited choice set, consumers often selected non-profit to others (Mend=5.97; Mbeginning=4.79; t(308)=7.24, learn about options in a specific order. When arranging such a choice p<.001, d=.82). A serial mediation analysis supported our predic- set, should marketers present the best option - the one that consum- tions; finding the best non-profit at the end (vs. beginning) increased ers are likely to prefer the most in the set - at the beginning of the perceived progress in search, which increased effort payoff, leading sequence or save it for last? participants to be more likely to recommend their selected non-profit Advances in Consumer Research (Volume 48) / 1179 to others (βindirect=.10, 95% CI=[.01,.21]; PROCESS model 6; Hayes Dellaert, Benedict GC, and Gerald Häubl (2012), “Searching in and Preacher 2014) (figure 6). Choice Mode: Consumer Decision Processes in Product Our final study demonstrated another consequence of presen- Search with Recommendations,” Journal of Marketing tation order. Participants viewed five apartment options and chose Research, 49 (1), 277-88. one to rent. When the participants found the best apartment at the Diehl, Kristin, and Gal Zauberman (2005), “Searching Ordered end (vs. beginning), they were more willing to submit an applica- Sets: Evaluations from Sequences Under Search,” Journal of tion to rent the apartment sooner without viewing additional options Consumer Research, 31 (4), 824-32.

(Mend=6.49; Mbeginning=6.06; t(197)=2.92, p=.004, d=.51). This effect Dhar, Ravi (1997), “Consumer Preference for a No-Choice was mediated by greater perceived effort payoff when the best option Option,” Journal of Consumer Research, 24 (2), 215–31. was at the end (vs. beginning) (βindirect=.43, 95% CI=[.23, .68]). Dhar, Ravi, and Itamar Simonson (2003), “The Effect of Forced Overall, we find that showing the best option at the end (vs. Choice on Choice,” Journal of Marketing Research, 40 (2), beginning) can increase perceived payoff from search and increase 146-60. valuation of the chosen option. This research contributes to literature Drolet, Aimee, Mary Frances Luce, and Itamar Simonson (2009), on search (Delleart and Häubl 2012) and effort perception (Kim and “When Does Choice Reveal Preference? Moderators of Labroo 2011). Additionally, we connect literature on goal (Harkin et Heuristic Versus Goal-Based Choice,” Journal of Consumer al. 2016) to demonstrate that perceived progress in search drives this Research, 36 (1), 137-47. effect. Lastly, these results offer managerial implications, as finding Goodwin, Geoffrey P., and John M. Darley (2008), “The the best option at the end increased willingness to recommend the Psychology of Meta-Ethics: Exploring Objectivism,” selected option and increased choice commitment. Cognition, 106 (3), 1339-66. Harkin, Benjamin, Thomas L. Webb, Betty PI Chang, Andrew REFERENCES Prestwich, Mark Conner, Ian Kellar, Yael Benn, and Paschal Anderson, Christopher J. (2003), “The Psychology of Doing Sheeran (2016), “Does Monitoring Goal Progress Promote Nothing: Forms of Decision Avoidance Result from Reason Goal Attainment? A Meta-Analysis of the Experimental and Emotion,” Psychological Bulletin, 129 (1), 139–66. Evidence,” Psychological Bulletin, 142 (2), 198-229. Berger, Jonah (2016), “Does Presentation Order Impact Choice Hayes, Andrew. F. (2013), An Introduction to Mediation, After Delay?” Topics in Cognitive Science, 8 (3), 670-84. Moderation, and Conditional Process Analysis: A Regression- Berman, Jonathan Z., Alixandra Barasch, Emma E. Levine, and Based Approach, New York, NY: Guilford Press. Deborah A. Small (2018), “Impediments to Effective Altruism: Hayes, Andrew F., and Kristopher J. Preacher (2014), “Statistical The Role of Subjective Preferences in Charitable Giving,” Mediation Analysis with a Multicategorical Independent Psychological Science, 29 (5), 834-44. Variable,” British Journal of Mathematical and Statistical Broniarczyk, Susan M., and Jill G. Griffin (2014), “Decision Psychology, 67 (3), 451-70. Difficulty in the Age of Consumer Empowerment,”Journal of Iyengar, Sheena S., and Mark R. Lepper (2000), “When Choice is Consumer Psychology, 24 (4), 608-25. Demotivating: Can One Desire Too Much of a Good Thing?,” Cacioppo, John T. and Richard E. Petty (1982), “The Need for Journal of Personality and Social Psychology, 79 (6), 995- Cognition,” Journal of Personality and Social Psychology, 1006. 42(1), 116–31. Kim, Sara and Aparna A. Labroo (2011), “From Inherent Value to Cacioppo, John T., Richard E. Petty, and Chuan F. Kao (1984), Incentive Value: When and Why Pointless Effort Enhances “The Efficient Assessment of Need for Cognition,”Journal of Consumer Preference,” Journal of Consumer Research, 38 (4), Personality Assessment, 48(3), 306–07. 712-42. Carver, Charles S. and Michael F. Scheier (1990), “Origins and Kruglanski, Arie W., James Y. Shah, Ayelet Fishbach, Ron Functions of Positive and Negative Affect: A Control-Process Friedman, Woo Young Chun, and David Sleeth-Keppler View,” Psychological Review, 97 (1), 19–35. (2002), “A Theory of Goal Systems,” in Advances in ——— (1998), On the Self-Regulation of Behavior, New York: Experimental Social Psychology, Vol. 34, ed. Mark P. Zanna, Cambridge University Press. San Diego: Academic Press, 331–78. Cha, Hoon S., Soeun You, KwangWook Gang, and Minseok Park Lamberton, Cait Poynor, and Kristin Diehl (2013), “Retail Choice (2018), “The Value and Risk of Curated Shopping: Online Architecture: The Effects of Benefit-and Attribute-Based Consumer’s Choice,” International Journal of Business and Assortment Organization on Consumer Perceptions and Information, 13 (3), 321-49. Choice,” Journal of Consumer Research, 40 (3), 393-411. Chaiken, Shelly and Durairaj Maheswaran (1994), “Heuristic Luce, Mary Frances, John W. Payne, and James R. Bettman (1999), Processing can Bias Systematic Processing: Effects of Source “Emotional Trade-Off Difficulty and Choice,”Journal of Credibility, Argument Ambiguity, and Task Importance Marketing Research, 36 (2), 143-59. on Attitude Judgment,” Journal of Personality and Social Miller, Joanne M. and Jon A. Krosnick (1998), “The Impact Psychology, 66 (3), 460–73. of Candidate Name Order on Election Outcomes,” Public Chartrand, Tanya L., Joel Huber, Baba Shiv, and Robin Tanner Opinion Quarterly, 62, 291-330. (2008), “Nonconscious Goals and Consumer Choice,” Journal Norton, Michael I., Daniel Mochon, and Dan Ariely (2012), of Consumer Research, 35 (2), 189–201. “The IKEA Effect: When Labor Leads to Love,”Journal of Corbin, Ruth M. (1980), “Decisions That Might Not Get Made,” in Consumer Psychology, 22 (3), 453-60. Cognitive Processes in Choice and Decision Behavior, ed. T. Novemsky, Nathan, Ravi Dhar, Norbert Schwarz, and Itamar S. Wallsten, Hillsdale, NJ: Erlbaum. Simonson (2007), “Preference Fluency in Choice,” Journal of Marketing Research, 44 (3), 347-56. 1180 / Difficult by Design: Choice Difficulty and Effort in Decision Making

Olson, James M., Robert J. Ellis, and Mark P. Zanna (1983), Sela, Aner, and Jonah Berger (2012), “Decision Quicksand: How “Validating Objective Versus Subjective Judgments Interest in Trivial Choices Suck Us In,” Journal of Consumer Research, Social Comparison and Consistency Information,” Personality 39 (2), 360-70. and Social Psychology Bulletin, 9(3), 427–36. Shaddy, Franklin, and Ayelet Fishbach (2018), “Eyes on the Prize: Parker, Jeffrey R., Donald. R. Lehmann, and Yi Xie (2016), The Preference to Invest Resources in Goals Over Means,” “Decision Comfort,” Journal of Consumer Research, 43 (1), Journal of personality and social psychology, 115 (4), 624-37. 113–33. Shafir, Eldar, Itamar Simonson, and Amos Tversky (1993), Petrocelli, John V., Zakary L. Tormala, and Derek D. Rucker “Reason-Based Choice,” Cognition, 49 (1-2), 11-36. (2007), “Unpacking Attitude Certainty: Attitude Clarity and Spiller, Stephen A., and Lena Belogolova (2017), “On Consumer Attitude Correctness,” Journal of Personality and Social Beliefs About Quality and Taste,” Journal of Consumer Psychology, 92 (1), 30–41. Research, 43 (6), 970-91. Petty, Richard. E. and Duane T. Wegener (1998), “Attitude Change: Thomas, Manoj and Geeta Menon (2007), “When Internal Multiple Roles for Persuasion Variables,” in The Handbook of Reference Prices and Price Expectations Diverge: The Role of Social Psychology, Vol.4, ed. Daniel Gilbert, Susan Fiske, and Confidence,” Journal of Marketing Research, 44 (3), 401–09. Garner Lindzey, New York: McGraw-Hill, 323-90. Touré-Tillery, Maferima, and Ayelet Fishbach (2012), “The End Schrift, Rom Y., Oded Netzer, and Ran Kivetz (2011), Justifies the Means, but Only in the Middle,”Journal of “Complicating Choice,” Journal of Marketing Research, 48 Experimental Psychology: General, 141 (3), 570-83. (2), 308-26. ——— (2015), “It was (n’t) me: Exercising Restraint When Schrift, Rom Y., and Moty Amar (2015), “Pain and Preferences: Choices Appear Self-Diagnostic,” Journal of Personality and Observed Decisional Conflict and the Convergence of Social Psychology, 109 (6), 1117-31. Preferences,” Journal of Consumer Research, 42 (4), 515-34. Tversky, Amos, and Eldar Shafir (1992), “Choice Under Conflict: Schrift, Rom Y., Ran Kivetz, and Oded Netzer (2016), The Dynamics of Deferred Decision,” Psychological Science, “Complicating Decisions: The Work Ethic Heuristic and the 3 (6), 358-61. Construction of Effortful Decisions,”Journal of Experimental Van Osselaer, Stijn MJ, and Chris Janiszewski (2012), “A Goal- Psychology: General, 145 (7), 807-29. Based Model of Product Evaluation and Choice,” Journal of Sela, Aner, and Baba Shiv (2009), “Unraveling Priming: When Consumer Research, 39 (2), 260-92. Does the Same Prime Activate a Goal Versus a Trait,” Journal Zakay, Dan (1985), “Post-Decisional Confidence and Conflict of Consumer Research, 36 (3), 418-33. Experienced in a Choice Process,” Acta Psychologica, 58, 75–80. Working Papers

30% Human is Better Than 10% Human: Consumers’ Sensitivity to Human Versus Machine Involvement in Production

Almira Abilova, Erasmus University Rotterdam, The Netherlands Gizem Yalcin, Erasmus University Rotterdam, The Netherlands Stefano Puntoni, Erasmus University Rotterdam, The Netherlands

Prior research shows that people prefer human-made to machine-made products but has not investigated preferences for products made with more vs. less human involvement. Using a newly developed paradigm we find that consumer preferences for products decrease as less human labor is involved and that this decrease is linear.

It Tastes Better For Me Just Because You Don’t Like It: The Effect of Cognitive Dissonance on Product Enjoyment

Aya Aboelenien, HEC Montreal, Canada Caroline Roux, Concordia University, Canada

Prior research on cognitive dissonance has mostly focused on understanding consumers’ attitude and behavior after having to process counter attitudinal information. However little is known about the effect of cognitive dissonance on subjective consumption experiences. Across two experiments we investigate the effect of cognitive dissonance on subjective product enjoyment.

Markets as Contested Assemblages: Comparative Case Study

Aya Aboelenien, HEC Montreal, Canada Jack Sadek, Concordia University, Canada

Through studying meat vape and cryptocurrency markets we investigate actors contesting the legitimacy of these markets and their strategies. Involved actors include industry associations specialized media consumers religious institutions federal agencies financial and scientific communities. Actors participate in strategies of problematization demarcation validation and affirmation.

When Indulgent Choice Can Promote the Next Indulgent Choice

Naoki Akamatsu, Meiji Gakuin University, Japan Reo Fukuda, Asia University, Japan

This study discussed the impact of the result of the prior indulgent choice (buy or restraint) on subsequent indulgent choices. Whereas the conventional idea is that “restraint” of a prior choice promotes subsequent indulgent choices we clarified the conditions where “exercising” a prior choice facilitates subsequent indulgent choices.

The Vulnerable Refugee Mother: Consumption Responses to Identity Threats

Roua Alhanouti, University of Lille, France

This paper attempts to summarize a research investigating the experience of refugees’ families in “transition phases” associated to forced migration. It considers strategies developed by mothers in consumer behavior to deal with identity threats. Research was conducted based on ethnographic approach. The study focuses on family identity and transitional challenges.

Catia Carvalheiro Alves, Nova School of Business and Economics, Portugal Irene Consiglio, Nova School of Business and Economics, Portugal Luis Martinez, Nova School of Business and Economics, Portugal

I propose that brand preferences affect creative performance as well as accuracy of self-reported creativity. Professional creatives and consumers produce more creative work when they work for brands they like (vs. dislike). Also I observe a brand preference bias on successful forecasting of their own creative work.

Using Consumer Neuroscience to Explain Product Choice From Experience

Oriana Rachel Aragón, Clemson University, USA

Academics and practitioners increasingly are expending valuable resources studying consumer behavior through neuromarketing. Elec- troencephalogram (EEG) investigations have thus far gauged consumers’ preferences for presented products. This is the first EEG investigation (N=135) to successfully capture consumers’ nonproduct-related internal states to predict consumers’ product choices made moments after EEG data collection.

Perceived Age of a Name: Adopting Age-based Names in Over-the-Top Media Content

Sunny Arora, S P Jain Institute of Management and Research, Mumbai, India M. G. Parameswaran, S P Jain Institute of Management and Research, India Unnati Dogra, Sardar Patel Institute of Technology,Mumbai, India Sarah Hawa, Sardar Patel Institute of Technology,Mumbai, India

With media content exploding there is a need for character names befitting to their age. Names from the Indian voters’ list and over-the- top media were analyzed using sound symbolism. When vowel and consonant sounds are of consistent frequency names with high-frequency sounds are perceived younger than names with low-frequency sounds.

Curious Abstraction: How Message Abstraction and Financial Scarcity Affect Processing of Mortgage Ads

Esra Asif, University of Leeds, UK J. Josko Brakus, University of Leeds, UK Alessandro Biraglia, University of Leeds, UK

Buying a house is one of the most important decisions consumers make in their lifetime weighing reality (e.g. their financial resources) against imagination (e.g. owning a mansion). We show that abundant - opposed to scarce - mindsets trigger consumer’s curiosity when exposed to an abstract (vs. concrete) mortgage advertising frame.

Soul Inside the Machine: Product Morphology Influences Consumer Valuation

Sumitra Auschaitrakul, University of the Thai Chamber of Commerce, Thailand Dan King, University of Texas Rio Grande Valley, USA

Many products such as luxury wristwatches have a “creviced morphology” with a crevice bored into the product structure. We demonstrate the “crevice effect” in which products with a crevice (versus without) elicit a higher valuation underpinned by increased perceptions of a high level of materialized expression.

Experiences as Ends (versus Means) Bring More Happiness and Meaning

Hankyul Bae, University of Minnesota, USA Kathleen Vohs, University of Minnesota, USA

We tested whether pursuing an experience for intrinsic reasons (ends approach) versus extrinsic reasons (means approach) would differently affect consumer well-being. Two experiments revealed that people report more happiness and meaning when they think about their experiences as ends versus means to another end. Advances in Consumer Research (Volume 48) / 1183 Better With Diverse Contexts? The Effects of Contextual Diversity of Background on Consumption Imagery

Eunsoo Baek, Hong Kong Polytechic University, China Zhihong Huang, Hong Kong Polytechic University, China

The study explored the effect of contextual diversity of background images on consumption imagery with a moderating effect of processing style. Results showed that a multiple (v.s. a single and none) contextual background generated consumption imagery of a product for holistic (v.s. analytic) processors leading to positive product evaluation.

A Good Cry or A Gratifying Revenge? Desired Mixed Emotional Experience in Movie Consumption Across Cultures

Aaron J. Barnes, University of Louisville, USA Jie(Doreen) Shen, University of Illinois at Urbana-Champaign, USA Jennifer L Stoner, University of North Dakota, USA Carlos Torelli, University of Illinois at Urbana-Champaign, USA

Prior research suggests that content with mixed emotions appeals to collectivists but not individualists. In contrast the current research suggests that both individualists and collectivists desire movies with mixed emotional themes as long as the mixed emotions align with their cultural values.

Can’t See You! The Effect of Packaging on Calorie Perceptions

Sara Baskentli, Western Washington University, USA Tracy Rank-Christman, University of Wisconsin - Milwaukee, USA

This research investigates how packaging material impacts consumers’ perceptions of calories. Results from two studies show that packaging material has an impact on consumers’ caloric perceptions. The authors suggest that the material of package impacts calorie perceptions which in turn could have implications on product consumption.

Augmented Reality Effects on Attitude and Memory

Juliana M. Batista, EAESP-FGV, Brazil Annaysa Salvador Muniz Kamiya, Centro Universitário FEI, Brazil Delane Botelho, EAESP-FGV, Brazil

Little is known about the underlying mechanisms on the effect of augmented reality (AR) on consumer´s attitude and memory. In two experiments we demonstrate that positive attitude toward the brand increases and memory recall decreases when consumers are exposed to AR. Perception of media´s innovation and entertainment mediate such effect.

Understanding Recovery From Compulsive Consumption With Practice Theory

Larissa Carine Braz Becker, University of Turku, Finland Melissa Archpru Akaka, University of Denver, USA Hope Schau, University of Arizona, USA Elina Jaakkola, University of Turku, Finland

This paper analyzes how consumers reconfigure a system of practices during recovery from compulsive consumption. A study with recovering alcoholics reveals how consumers disengage from non-supportive practices and how new practices relate to each other. These findings contribute to the literature by using a sociological perspective to study behavioral change. 1184 / Working Papers Making Home Away From the Established Models of Ownership: Preliminary Findings From a Housing Cooperative

Ons Belaid, IAE Lyon, France Mariam Beruchashvili, California State University Northridge, USA Sonia Cappelli, University of Lyon, France William Sabadie, University Jean Moulin Lyon 3, France

Based on an ethnographic study conducted in the context of a large cooperative house located near the metropolitan area in France where residents practice consciously elected collective cohabitation we examine how consumers make home (Douglas 1991) in a physical space of a house deliberately devoid of conventional forms of ownership.

Digital Platforms and Market Intermediation

Kristin Bentsen, University of South-Eastern of Norway Eileen Fischer, York University, Canada Per Egil Pedersen, University of South-Eastern of Norway

This paper explores how market intermediation may be affected when market actors rely on digital platforms. It draws on participant observation conducted in “REKO” markets in which transactions and interactions are primarily conducted via groups organized on Facebook. The paper identifies four actor/platform intersections that may profoundly shape intermediation.

Another Advertising Stereotype: Effects of Non-Stereotyped Portrayals of Older Women in Advertising

Hanna Berg, Stockholm School of Economics, Sweden Karina T. Liljedal, Stockholm School of Economics, Sweden

This paper examines the effects of featuring non-stereotyped portrayals of older women in advertisements. In two empirical studies we demonstrate positive ad effects of using non-stereotyped portrayals of older models for female consumers. Furthermore these effects are explained with signaling theory and the social effects of advertising.

Commitment Contracts: The Effect of Commitment-Based Pricing on Customer Acquisition and Retention

Katja Berger, University of Hamburg, Germany Klaus Wertenbroch, INSEAD, Singapore Christina Schamp, University of Mannheim, Germany

Nowadays an increasing number of firms incorporate commitment into their pricing schemes and business models. Our research shows how to design and present commitment-based pricing optimally to acquire and establish long-term relationships with customers.

Feeling Pressured to Disclose Personal Information? Understanding The Role of Social Pressure on Consumers’ Information Disclosure Intention

Alena Bermes, Heinrich Heine University Düsseldorf, Germany Maximilian Alexander Hartmann, Heinrich Heine University Düsseldorf, Germany Peter Kenning, Heinrich Heine University Düsseldorf, Germany

This research examines if consumers feel social pressure to disclose personal information for mobile application usage. Grounded in the theory of self-determination the results prove that social pressure (being of multifaceted structure including fear of missing out subjective norms and image) is positively related to information disclosure intention. Advances in Consumer Research (Volume 48) / 1185 Stop It! Consumer Resilience as a Buffer Against Psychological Conflicts in the Digital Age

Alena Bermes, Heinrich Heine University Düsseldorf, Germany Nikita Maleev, Heinrich Heine University Düsseldorf, Germany Peter Kenning, Heinrich Heine University Düsseldorf, Germany

Drawing on psychological resilience theory this research is the first to define conceptualize and test the phenomenon of consumer resilience against the background of today’s digital environment which confronts consumers with rising psychological conflicts (e.g. fear of missing out). The preliminary results proof consumer resilience’s buffering effect against such stressors.

Dimensions of Believability of Brand News: An Exploratory Study

Kshitij Bhoumik, Texas Tech University, USA

This research explores the factors that shape consumer believability toward brand news on digital platforms. What are the cues consumers look to evaluate the authenticity of any brand communication? What kind of brand message is considered as highly believable? We identified eight dimensions that shape consumers’ perception of brand news.

Sequential Sensory Cues in Retailing and Food Packaging: Managerial and Ethical Implications

Dipayan Biswas, University of South Florida, USA Lauren I Labrecque, University of Rhode Island, USA Donald Lehmann, Columbia University, USA

Sensory cues are often encountered sequentially (than simultaneously) in many retailing food packaging and consumption contexts. A series of experiments demonstrate how the sequence in which sensory cues are encountered might influence food taste perceptions food choices and consumption volume.

It’s Written All Over Your Face—Applying Deep Neural Networks to Explain Human Mating Based on Facial Characteristics

Daniel Boller, University of St. Gallen, Switzerland

This research investigates human mating in online dating markets based on facial characteristics by utilizing high-dimensional profile section transaction and facial image data of 640000 users of an online dating platform. The results of this research assist software developers and marketing practitioners in designing algorithms for human matching.

Counterfeit Consumption Eliciting Group-Based Emotions

Delane Botelho, EAESP-FGV, Brazil Fernando A. Fleury, EAESP-FGV, Brazil Ramona de Luca, EAESP-FGV, Brazil

This paper investigates positive and negative emotions as consequences of the consumption of counterfeits in the context of soccer supporters. Specifically the perceived personal in-group responsibility for consuming counterfeits leads to negative affect (e.g. guilt and anger) while the responsibility for consuming original products leads to positive affects (e.g. pride).

A Contingency Theory of Artificial Intelligence: Consumer Beliefs, Value Creation, And Resistance to Creative AI

Anna Bouwer, University of St. Gallen, Switzerland Christian Hildebrand, University of St. Gallen, Switzerland

This paper examines the unexplored role of ‘creative AI’ by studying the perception of machine-generated ‘creative output’. It provides conceptual foundations on the notion of creative AI and demonstrate the consumer associations with AI along with the downstream consequences of creative AI for consumers’ perception of aesthetic quality and valuation. 1186 / Working Papers “’Sweet, Tasty Evil’ Or ‘Healthy Bliss’? Lay Theories On Food Products For Children”

Raphaela Elisabeth Bruckdorfer, University of Duisburg-Essen, Germany Oliver B. Büttner, University of Duisburg-Essen, Germany Gunnar Mau, Deutsche Hochschule für Gesundheit und Sport (DHGS), Germany

Based on qualitative and quantitative data we developed an instrument to assess consumers’ lay theories on child food products. In line with experts lays expected such products to contain high amounts of fat sugar and calories but at the same time perceived them as being moderately healthy and rather useful.

Unveiling the Color Matching Effect in Product Displays

Daniel Erik Brylla, Friedrich Schiller University Jena, Germany Gianfranco Walsh, Friedrich Schiller University Jena, Germany

Consumer researchers have long been interested in explaining the effects of individual and combined colors on consumer perceptions and behaviors. Across four studies the authors show that product colors that match a moderate amount of colors in the shopping environment increase consumers’ aesthetic perception and purchase intentions.

Uninformed But Unaware: How the Number of Likes Triggers a Knowledge Illusion on Social Media

Andrea Bublitz, University of Zurich, Switzerland Anne Scherer, University of Zurich, Switzerland René Algesheimer, University of Zurich, Switzerland

Drawing from theory on the group mind this research uncovers how likes on social media trigger a knowledge illusion: If individuals are not motivated or able to read but content received many likes they rely on external knowledge in the group mind and feel better informed than they objectively are.

How Can We Make Fat Women Feel Miserable?

Ana Julia Büttner, ESPM, Brazil Suzane Strehlau, ESPM, Brazil

The objective is to comprehend how plus-size women find well-being through fashion inside the symbolic violence. The qualitative method used semi-structured interviews using photo-elicitation reveals that fatshionistas play with fashion rules and present well-being. Al- though at the same time they reveal a subordination to the cult of thinness.

The Visual Impaired Tourist: Facilitators and Inhibitors in Choosing Tourist Destinations

Aline Delmanto Capone, ESPM, Brazil Vivian Iara Strehlau, ESPM, Brazil

This paper aims to understand what factors influence leisure travel destination choice for the visually impaired. A qualitative approach using in-depth interviews with three different groups preliminarily found that acquired blinds have more constraints than congenital or low vision. The origin of blindness affects their well-being and how they travel.

The Exquisite Exotic: Revalorization of Culinary Ingredients and Traditional Food Practices in the Global Marketplace

Flavia Cardoso, Universidad Adolfo Ibáñez, Santiago, Chile Angela Gracia B. Cruz, Monash University, Australia Pilar Ximena Rojas Gaviria, University of Birmingham, UK

This theoretical piece offers an integrative framework to theorize how marketing actors invest and participate in the international re- signification of culinary ingredients and food practices. It goes beyond localized case studies synthesizing the current theorizations on the internationalization of food and charts an agenda for future research. Advances in Consumer Research (Volume 48) / 1187 Chatbots an Exploratory Analysis on the Impact of NLP and Customer Sentiment Analysis

Lilian Carvalho, FGV/EAESP, Brazil Eusebio Scornavacca, University of Baltimore, USA

We partnered with a startup that uses NLP (natural language processing) to assist different firms to set up their virtual assistants. We analysed 4320 interactions and found that chatbot-human interaction had a better score than human-to-human interactions.

Let Me Split the Donation, I Will Donate More: How Installment Payment Reduces Perceived Cost and Increases Willingness to Donate .

Marta Caserotti, University of Padova, Italy Enrico Rubaltelli, University of Padova, Italy Paul Slovic, University of Oregon, USA

Do people prefer a donation split into several instalments or a lump sum? Do they prefer to pay now or later? Four experiments (and a pretest) investigated how payments options and intertemporal choice affect donations. We found that the opportunity of giving through instalments increased donations decreasing the perceived cost.

A Reimagining: Prefiguring Systems of Alternative Consumption

Katherine Casey, University of Kent, UK Maria Lichrou, University of Limerick, Ireland Lisa O’Malley, University of Limerick, Ireland

This ethnographic research reveals how an ecovillage prefigures consumption via a repertoire of alternative consumption and production systems designed to challenge neoliberal notions of choice value and ownership; explores how community members participate in broader changemaking and how the community engages the broader institutional framework to further environmental education.

Understanding Overconsumption – Symbolic and Spatial Transgressions in Public Places

Cecilia Cassinger, Lund University, Sweden Jörgen Eksell, Lund University, Sweden

Overconsumption refers to harmful or excessive consumption. This study demonstrates how experiences of overconsumption are constructed in transgressions of boundaries within places. Results are supported by a study of citizens’ experiences of tourism-consumption in touristified cities. They underscore the importance of considering experiences of overconsumption practices for sustainable consumption.

Causa Sui (Cause of Itself): How Self-benefit Appeal Framing Interacts With Situational and Dispositional as Donation Cause-types

Chia-Chi Chang, National Chiao Tung University, Taiwan Joseph Iesue, National Chiao Tung University, Taiwan Chia Hua Lin, National Chiao Tung University, Taiwan

Research provides insight on charitable giving by testing the interaction between donation cause-types (situational vs. dispositional) and self-benefit appeals (‘karma’ - intentional but dependent ‘feel-good’ - intentional and independent ‘humanity’ - unintentional). Results show ‘karma’ appeal elicits higher donations for situational causes with ‘feel-good’ and ‘humanity’ appeals benefiting dispositional causes.

Social Influence and Personalization in Behavioral Retargeting Advertising

Yaping Chang, Huazhong University of Science and Technology, China Shaowei Chai, Huazhong University of Science and Technology, China Jun Yan, Huazhong University of Science and Technology, China

Through a serial of field experiments on Facebook this research examined the impacts of social and personalized messages in retargeting advertising on customer responses. Results indicate geographic proximity is more effective than number of likers on social media. Personal- ized message works negatively. Product knowledge works as a moderates. 1188 / Working Papers How Social Functioning Ability and Crowdedness Impact Consumer Behavior

Shiyun Chen, University of Iowa, USA Gary Gaeth, University of Iowa, USA Irwin Levin, University of Iowa, USA

In the current research we investigate the impact of social functioning on consumers under varying levels of social crowdedness. Our finding suggests that people with lower social functioning abilities feel less comfortable more anxious and dislike the more crowded environment compared to people with higher social functioning abilities.

Securing Non-Touristy Shops-- Claiming Local Collective Identities in Political Consumption Against Inbound Tourists’ Shopping Spree

Wei-Fen Chen, University of Leicester, UK Tin-yuet Ting, The Hong Kong Polytechnic University, China

Departing from behavioral perspectives that examine how consumers’ political consumption relates to personal characteristics this study investigates how collective identities of local consumers are shaped in response to overwhelming inbound shopping tourism informing the contemporary practices of citizen consumers and manifesting a novel approach to understand political consumption.

Utilizing Matte Packaging to Communicate Social Warmth

Yu-Shan Athena Chen, Eindhoven University of Technology, The Netherlands Wei-Ken Hung, National United University, Taiwan

Two studies demonstrate that matte packaging conveys haptic information in the same manner as interpersonal touch (Study 1); and matte packaging is preferred over glossy packaging when people have a goal to reduce coldness (Study 2). Together these results suggest that matte packaging is instrumental in communicating warmth.

Waiting For A Download: The Effects of Congruency Between Anthropomorphic Cues and Shopping Motivation on Consumer Patience

Siyun Chen, Jinan University, China Xinliang Wei, Temple University, USA

People tend to be impatient while waiting for a webpage to download which might be negatively correlated with their satisfaction with the online providers. In particular we propose a new framework of how the congruence of anthropomorphic messengers and shopping motivation enhances patience (i.e. perceived quickness of a download).

What to Buy When the American Dream Fails? Understanding the Taste of Downwardly Mobile Consumers

Wei-Fen Chen, University of Leicester, UK Xue Wang, Chinese University of Hong Kong, China Ying-yi Hong, Chinese University of Hong Kong, China

This study examines how consumer behaviors are influenced by upward and downward economic mobility. Findings indicate that facing downward mobility consumers lessen the economic components in their self-identities and thus are less interested in purchasing products framed in advertisements that appeal to economic capital compared to their upwardly mobile counterparts. Advances in Consumer Research (Volume 48) / 1189 The Effect of Identical or Distinct Service Failure Recovery on Customer Satisfaction

Yin-Hui Cheng, National Taichung University of Education, Taiwan Shih-Chieh Chuang, National Chung Cheng University, Taiwan Ya-Ju Shen, National Chung Cheng University, Taiwan

This paper study showed that providing the same type of service is worse than providing a different service in terms of customer satisfaction. Two boundary conditions were examined in which the severity of the failure and consumer participation mitigate the effect of identical/ distinct service recovery on customer satisfaction.

I Can Be Perfect! Growth Mindset Moderates the Effect of Perfectionism on Maladaptive Eating Behaviour

Jennifer Chernishenko, University of Lethbridge, Canada Rhiannon MacDonnell Mesler, University of Lethbridge, Canada Debra Z Basil, University of Lethbridge, Canada

To clarify the link between perfectionism and eating behaviors we proposed that implicit mindset would moderate the relationship between perfectionism and maladaptive eating behaviour. Whereas a growth mindset increased use of dieting strategies and consequently maladaptive eating behavior fixed mindset reduced this effect. Implications for research and practice are discussed.

When Technology Environment is More Indulging: Impact of Accessibility to Technology-Equipped Environment on Self-Control Behaviors

Yunjia Chi, Huazhong Agricultural University, China Fue Zeng, Wuhan University, China Li Huang, Hofstra University, USA

The mere perception of accessibility to technology-equipped environment without actually using technologies triggers consumers’ effort conservation intention which results in lapses of self-control.

The Effects of Majority/Minority Source Status and Argument Quality of Online Reviews on Product Perceptions: The Moderating Role of Product Type

Yi-Wen Chien, National Taiwan University, Taiwan Shian-Ko Liu, National Taiwan University, Taiwan

This study posits that for different products (search vs. experience goods) source status and review quality may have different impacts. Experience goods rely more on source status whereas search goods rely more on review quality. Marketers are suggested to develop effective promotion strategies based on product type.

Strong As An Ox: Usage Of Storytelling By Vegan Athletes To Reduce Consumers’ Meat-Health Associations

Ziad Choueiki, Ghent University, Belgium Maggie Geuens, Ghent University, Belgium Iris Vermeir, Ghent University, Belgium Adriaan Spruyt, Ghent University, Belgium

This paper explores the implicit/explicit association between meat consumption and health. We propose that storytelling by vegan athletes (whose personal experiences challenge the prejudice of veganism being an inadequate diet) can help reducing both consumers’ implicit meat-health association and the explicit belief that meat is necessary normal natural and nice 1190 / Working Papers Friendship Alleviates Sense of Regret

JungHan Chung, Sogang University, South Korea Young-Won Ha, Sogang University, South Korea

The authors investigated whether psychological relatedness affects consumers’ sense of regret in consumption failure settings. Through three experiments the authors demonstrated that close relationship can alleviate sense of regret and this friendship effect was mediated by locus of causality and was moderated by individuals’ level of subjective well-being.

The Making of Healthy, Wealthy, and Happy Consumers: Practices and Politics of Nudging in For-Profit Firms

Leonardo Conte, Università della Svizzera italiana, Switzerland Lena Pellandini-Simanyi, Università della Svizzera italiana, Switzerland

This paper extends research in Consumer Culture Theory on the making of the consumer subject by examining (1) the performative role of nudging practices in businesses (2) their underlying normative visions and (3) the concrete processes through which they shape consumers within the Foucauldian framework of neoliberal governmentality.

Life Transition to Terminality: Dynamic Regimes of Representation, Permanent Liminality, and Coping Consumption

Michele Corengia, Università della Svizzera italiana, Switzerland Luca M. Visconti, Università della Svizzera italiana, Switzerland

Through the lenses of representation theory this (n)ethnographic research contributes to literature on consumer permanent liminality and coping consumption. Unpacking life and market transition of patients to terminality accounts for how they use representational agency and consumption to counterbalance etic representational regimes and the tensions arising from their liminal experience.

Salience Over Centrality? How Environmental Identity Drives Pro-Environmental Consumption

Sandor Czellar, HEC Lausanne, Switzerland Leila Rahmani, HEC Lausanne, Switzerland Simona Haasova, University of Lausanne, Switzerland Valentina Clergue, HEC Lausanne, Switzerland Christian Martin, National University of Ireland Maynooth, Ireland

We investigate the relationships between environmental identity centrality its salience and pro-environmental behavior. Five studies reveal a distinct influence of environmental identity salience and centrality but no interaction between the two on several types of self-reports and actual behaviors. Our findings bear implications for research aiming to promote pro-environmental consumption.

A Permanently Loose End: Constructing the Intersectional Identity of British South Asians

Ofer Dekel Dach, De Montfort University, UK Amandeep Takhar, De Montfort University, Leicester, UK

Moving away from dominant acculturation models this research study draws on intersectional theory to better understand how individuals from the British South Asian community in the U.K. respond to social complexity and engage in identity construction. Our findings present a picture of the way participants mix and juxtapose identity categories. Advances in Consumer Research (Volume 48) / 1191 Leaving the Field: Problematizing Temporality in and of Ethnographic Consumer Research

Christian Dam, University of Gothenburg, Sweden Mathias Sosnowski Krabbe, University of Southern Denmark, Denmark

This paper argues that the temporality of ethnographic consumer research is inadequately illuminated which prevents the conceptualization of exiting the field. Thus a stronger focus on the various methodological and ethical aspects of exiting the field is needed particularly due to the changing notions of what constitutes the field.

Group Buying of Experiential (vs . Material) Purchases

Gopal Das, Indian Institute of Management Bangalore, India James Agarwal, University of Calgary, Canada Mark T Spence, Bond University, Australia

This research examines how experiential vs. material purchases influence consumer preference for group buying (vs. individual buying). Specifically consumers while purchasing experiential (vs. material) goods anticipate more enjoyment which in turn results in higher intent to participate in group buying behavior. These effects hold for interdependent (vs. independent) self-construal.

Marketing’s Role in Promoting Common Good: A Systematic Examination and an Agenda For Future Inquiry

Benét DeBerry-Spence, University of Illinois at Chicago, USA Lez Ecima Trujillo Torres, University of Illinois at Chicago, USA Rumela Sengupta, University of Illinois at Chicago, USA Jia Chen, University of Illinois at Chicago, USA Kohei Matsumoto, University of Illinois at Chicago, USA

We examine the role marketing has played in promoting the ‘common-good’ by conducting a systematic multi-method multi-journal inquiry. Marketing consistently contributes to a dialogue of common-good by publishing impactful articles in areas like health and CSR. Opportunity exists for greater contributions to important areas such as gender-empowerment and education.

Does Alphago’s Victory Over Human Champions Reduce The Interest Of People In Go? The Negative Effect Of Artificial Intelligence On Competition Events

Zhongzhun Deng, Huazhong University of Science & Technology, China Donghong Zhu, Huazhong University of Science & Technology, China

Four studies (N=1214) confirm that how and why losing to robots (vs human) negatively affected consumer attitude toward these events. We contribute to the literature on Robot AI and goal theory from another perspective. In addition research help companies and governments to avoid negative effects when advertising robots.

Sacrificing Pleasure For Status. Which Customers Are Attracted by Healthy Luxury Goods?

Perrine Desmichel, Northwestern University, USA Goedele Krekels, IÉSEG School of Management, France

Luxury brands have recently entered the market of healthy products although healthy benefits can reduce the hedonic benefits of luxury consumption. Across two preliminary studies we propose and show that consumers who did not earn (vs. earned) their wealth present stronger desire for healthy (vs. non-healthy) luxury goods. 1192 / Working Papers Atemporal Nostalgia as Artefact in Post-Colonial Non-Western Context

Amina Djedidi, Université Paris-Est Créteil, IRG, France Nacima Ourahmoune, Kedge Business School, France

Often pictured as a passive longing for the past this paper conceives nostalgia as atemporal: it creates states and spaces where time is reconfigured to reach selective accounts of bridges between the past present and future. Nostalgia atemporality sutures complex and heterogeneous sociocultural references in Post-Colonial Non-Western context Algeria

Keeping Track of What’s Deserved When Personal Finances Grow

David Dolifka, University of California Los Angeles, USA Stephen A Spiller, University of California Los Angeles, USA

All financial growth first requires principal to invest. In this project we explore the relationship between the perceived deservingness of both principal and downstream growth. Initial studies suggest consumers downplay whether or not their principal was deserved when judging the deservingness of subsequent financial growth.

Does a Flaw Outweigh a Flaw? The Bright Side of Negative Comments in Online Product Reviews

Shen Duan, Renmin University of China, China

The research find that two-sided information reviews are more persuasive to consumers than one-sided information reviews and the importance proportion and position of negative information play a moderating role in this relationship. This study enriches the relevant research on the impact of online reviews on consumer product purchases.

Recovering Cultural Authority: Marketplace Abandonment and Return

Toni Eagar, Australian National University, Australia Andrew Lindridge, Newcastle University, UK Diane Martin, RMIT University, Australia

This project investigates how a brand can recover its former position in a marketplace after losing legitimacy. We apply a process data approach to qualitatively analyse 614 media reports and 31 interviews. We extend the brand longevity concept and reveal brand recovery as a process involving return dormancy and resurrection.

Goal Conflict Undermines Self-Continuity and Leads to Greater Preference For Renting Vs . Buying

Mahdi Ebrahimi, California State University Fullerton, USA Anoosha Izadi, University of Massachusetts, USA

The present research identifies a novel antecedent for consumer renting behavior. We propose that consumers who experience goal conflict are more likely to rent vs. buy a product or service. We also show that the effect is mediated by lower perceptions of future self-continuity.

The “Common Good” Phenomenon in Country of Origin Effects

Martin Egger, University of Vienna, Austria Arnd Florack, University of Vienna, Austria Adamantios Diamantopoulos, University of Vienna, Austria Georgios Halkias, University of Vienna, Austria

The “common good” phenomenon describes that products share positive attributes but are distinguished according to negative attributes. In three studies we demonstrate that this phenomenon applies to products from different product categories and countries and show that it decreases the differentiating power of favorable country of origin labels. Advances in Consumer Research (Volume 48) / 1193 Coopetition and Creativity in Idea Crowdsourcing: Investigating the Mediating Role of Emotional and Motivational Ambivalences

Mehdi Elmoukhliss, TSM Research, France

We investigate why coopetition-based idea crowdsourcing has a positive effect on consumers’ creativity. We study two potential mediators: emotional ambivalence and motivational ambivalence. We found a significant indirect effect of motivational ambivalence. We contribute to the consumer research literature by introducing coopetition as a new model for engaging creative consumers.

The Effects of Mindfulness on Healthy Food Choice in Childhood Socioeconomic Groups

Amy Errmann, University of Auckland, New Zealand Yuri Seo, University of Auckland, New Zealand Felix Septianto, University of Auckland, New Zealand

The effects of mindfulness on healthy food selection is receiving attention in academia and practice. The current research examines which childhood socioeconomic groups benefit the most from mindfulness as it pertains to food choice. We propose that mindfulness pro- vokes low childhood socioeconomic groups to make healthier food choices.

Risk on the Edge: The Effect of Socio-spatial Location on Consumer Preferences

Sina Esteky, Miami University, Ohio, USA Amar Cheema, University of Virginia, USA

This paper explores how the relative spatial location of consumers in a group affects consumer preferences. We find that being located close to the center versus edge of a group primes the concepts of safety versus risk respectively. As a result being farther away from the center heightens risk-seeking.

How the Different Wordings of Message in Pop-Up Window Affect Consumers’ Behavior?

Yafeng Fan, Tsinghua University, China Jing Jiang, Renmin University of China, China

This research classifies the message wording approaches in pop-up windows into “guilt appeal” “delaying appeal” and the “no appeal” and examine how the different wording affect consumers’ behavior. Compared with not using any appeal “guilt appeal” and “delaying appeal” in pop-up windows can promote consumers’ retention rate and customer satisfaction.

When Social Media Word-of-mouth Catches Envious Eyes: How Feelings of Being Envied Influence Self-brand Connection

Wenting Feng, Hainan University, China Irina Y. Yu, Chinese University of Hong Kong, China Morgan X. Yang, The Hang Seng University of Hong Kong, China

We examined the ambivalent impacts of being envied experiences triggered by sharing luxury consumption on social media. Results from three experiments show that feeling being maliciously (benignly) envied decreases (increases) self-brand connection mediated by anxiety (pride). The effects are more salient when consumers interact with peers with lower social-distance.

Give Me the Fish Vs. Teach Me How to Fish: The Effect of Socioeconomic Status on Recipients’ Willingness to Accept Help

Maura Ferreira, Catholic University of Rio Grande do Sul, Brazil Erick M. Mas, Vanderbilt University, USA Kelly Goldsmith, Vanderbilt University, USA

This research investigates an overlooked side of prosocial behavior the recipient’s willingness to accept help. We show that recipients’ socioeconomic status can decrease and increase willingness to accept low-order (money/food) and high-order (scholarships) help as these types of help can represent both self-threat and self-support depending on recipients’ socioeconomic status. 1194 / Working Papers Mitigating Uncertainty in Consumer Adoption of Medical Artificial Intelligence

Darius-Aurel Frank, Aarhus University, France Polymeros Chrysochou, Aarhus University, Denmark Panagiotis Mitkidis, Aarhus University, Denmark Dan Ariely, Duke University, USA

This pre-registered study investigates the potential of explicit uncertainty quantification information in mitigating consumers’ reluctance to trust and depend on recommendations by medical artificial intelligence agents’ in decision scenarios characterized by uncertainty.

Brand Familiarity Shows in Consumers’ Eyes

Léon Franzen, Concordia University, Canada Amanda Cabugao, Concordia University, Canada Aaron P. Johnson, Concordia University, Canada H. Onur Bodur, Concordia University, Canada Bianca Grohmann, Concordia University, Canada

Although brand familiarity has been linked to product recognition speed and prior experience the objective measurement of consumers’ brand familiarity remains challenging. We integrated an extensive product image validation study with novel pupillary response measure- ments to demonstrate reliable dissociation of familiar and unfamiliar brands without an overt response.

What Makes Services Luxurious? Insights From a Qualitative Study

Martin P Fritze, University of Cologne, Germany Jochen Wirtz, National University of Singapore, Singapore Jonas Holmqvist, Kedge Business School, France

We address the research question ‘What constitutes a luxury service experience’? and report results from exploratory interviews conducted with luxury service consumers to understand how they perceive differences between luxury and non-luxury services. Our results reveal key luxury perception dimensions in the search decision consumption and post-consumption states.

Re-imagining Subcultures: A Comparative Study of Generational Preconceptions of Music Consumption, Identity, and Belonging

Olivia Fulvio-Mason, University of Gloucestershire, UK Richard Warr, University of Gloucestershire, UK

As the existence of subcultures within today’s youth culture have begun to be questioned this study focuses on changes surrounding music consumption that have occurred between Millennials and their parent generations. Aims include the exploration of youth subcultures musical identity and the transformation of generational preconceptions towards popular music consumption.

Seeing Through Your Eyes: The Influence of Interface on Advertisement Processing

Lin Ge, Chinese University of Hong Kong, China Hao Shen, Chinese University of Hong Kong, China

Consumers are more likely to take a first-person perspective while viewing advertisements on a smartphone than on a personal computer. Consequently they might react more favorably to an ad with content generated from the first-person perspective if they view it on a smartphone (vs. a PC). Advances in Consumer Research (Volume 48) / 1195 Mixed Emotions Inspire Missions (Im)possible – The Role of Emotions in Setting and Achieving High Aspirations

Jana-Verena Gerhart, Johannes Gutenberg University Mainz, Germany Oliver Emrich, Johannes Gutenberg-University of Mainz, Germany Michael Norton, Harvard Business School, USA Jordi Quoidbach, ESADE Business School, Spain Ann-Kathrin Haderdauer, Johannes Gutenberg University (graduate), Germany

A longitudinal study of New Year’s Resolutions examines the effect of emotional states on goal setting and goal attainment. Participants (N=1004) assigned to experience mixed emotions when setting goals increased their aspiration levels more than those assigned to experience positive emotions. Higher aspirations in turn predicted resulting change.

How Incongruent Products Drive Brand Engagement: The Role of Curiosity

Maximilian H. E. E. Gerrath, University of Leeds, UK Alessandro Biraglia, University of Leeds, UK

Brands seek new ways to engage consumers on social media. For example McDonald’s recently generated buzz by announcing an incongruent product—Big Mac scented candles. Across three experiments we examine the link between (low vs. high) congruence and brand engagement. Moreover we identify curiosity as a driver of this effect.

You Don’t Need To Be Me To Win My Heart: The Impact of Cognitive Appeals on Empathy In Pro-Social Behaviors

Mahsa Ghaffari, university of Portsmouth,UK Giovanni Pino, University of Chieti-Pescara, Italy Daniel Nunan, University of Portsmouth, UK

Despite the widespread conviction that empathy and congruency between self and the target increase pro-social behaviours this research underscore the importance of marketing communication techniques in boosting altruistic behaviour when the self and target are incongruent.

The Effect of Warmth and Competence Perceptions on Algorithm Preferences

Zohar Gilad, Technion University, Israel Liat Levontin, Technion University, Israel Ofra Amir, Technion University, Israel

Perceptions of warmth (intent) and competence (ability) strongly affect people’s judgment of individuals groups and organizations. We find that surprisingly not only competence perceptions affect people’s preferences for artificial intelligence algorithms. Moreover similarly to the judgments of other people warmth perceptions tend to have a stronger effect on preferences.

Which Autonomous Vehicle (AV) Will be More Pro-Social? The One You Own or Rent .

Tripat Gill, Wilfrid Laurier University, Canada

Recent research reveals that consumers expect AVs to be less pro-social than regular car drivers. Current work examined how ownership would impact the latter moral shift. It was found that owned AVs are expected to be more pro-social than the ones that are rented due to self-extension into the AV. 1196 / Working Papers With God on My Side, I Have No Fear of Ultra-processed Foods and Lack of Exercise

Ali Gohary, Monash University, Australia Hean Tat Keh, Monash University, Australia Eugene Chan, Monash University, Australia

This research investigates the impact of God salience on consumers’ food choice and physical inactivity. Three studies (secondary data chronic and primed God salience) show that God reminders lead consumers to exhibit greater willingness to choose ultra-processed (vs. unprocessed) food products and being more physically inactive.

When Recycling Seems Risky: Are Consumers Afraid to Recycle Sensitive Information?

Gabriel E. Gonzales, SUNY New Paltz, USA Matthew D. Meng, Utah State University, USA Christopher Berry, Colorado State University, USA R. Bret Leary, University of Nevada, Reno, USA

The current research proposes that consumers are uncomfortable recycling items containing private information (e.g. “junk” mail bank statements etc.) despite seeing such items as recyclable. Two initial studies provide evidence for this effect suggesting a potentially untapped source of highly-recyclable material while generating insights for both scholars and practitioners.

Behavioral Option Value and Product Set Choice

Paniz Gorji, Erasmus University Rotterdam, The Netherlands Benedict Dellaert, Erasmus University Rotterdam, The Netherlands Bas Donkers, Erasmus University Rotterdam, The Netherlands

We propose a behavioral theory of consumers’ option value for sets from which they consume a product in the future. For example consumers may choose between health insurance networks from which they select a provider for treatment later. Results from two experiments support our hypotheses and illustrate the proposed approach.

The Injustice of Envy

R Justin Goss, Colorado State University-Pueblo, USA

This research investigates the impact that status hierarchies within brand communities have on the occurrence of Envy. I examine the antecedents and consequences of Malicious and Benign Envy. Specific attention is paid to Deservingness. Findings show that Deservingness affects feelings of both Malicious and Benign Envy. Implications are discussed.

Consumer Deceleration Through Cultural Borrowing and Critical Reflection

Sarah Grace, University of Arkansas, USA

This research extends previous findings on consumer deceleration by exploring how cultural borrowing facilitated through marketing activities in a globally connected consumer culture can provoke critical reflection. The findings of this study empirically ground a theoretical process that marketers can use to constructively engage in the phenomenon of social acceleration.

Is That Brand Relevant to Me? Concept, Measurement and Antecedents of Brand Relevance

Amélie Guèvremont, École des Sciences de la Gestion, UQAM, Canada Fabien Durif, École des Sciences de la Gestion, Université du Québec à Montréal, Canada Cindy Grappe, Écoles des Sciences de la Gestion, Université du Québec à Montréal, Canada

This research explores the meaning of brand relevance according to consumers and its creation from a branding perspective. Across four studies it defines brand relevance and its dimensions (utilitarian/symbolic/social/environmental) develops a measurement scale identifies brand relevance’s antecedents and validates them through an in-depth analysis of relevant brands’ social media content. Advances in Consumer Research (Volume 48) / 1197 Investigating the Effects of Uncertainty Avoidance on Customer Loyalty Intention

Bingxuan Guo, University of Texas at San Antonio, USA Karen Page Winterich, Pennsylvania State University, USA Yinlong Zhang, University of Texas at San Antonio, USA

This research explores how cultural orientation of uncertainty avoidance affects customer loyalty intention. Secondary country level data and primary individual level experiment provide convergent evidence showing a negative effect of uncertainty avoidance on customer loyalty intention and this effect is mitigated by desire to change and need for structure.

Social Class and Risk Taking: The Effects of Perceived Benefits and Decision Domain

Yafei Guo, University of Illinois at Urbana-Champaign, USA Xue Wang, University of Hong Kong, China

The present paper is the first to systematically examine the class-related differences in consumer risk-taking behavior. It proposes a moderation of risk domain to reconcile the differential risk patterns among different socioeconomic consumers. The class-related risk patterns provide insights for managers who target consumers with social class differentiation.

Upskilling Communication and Self-Production: How to Communicate the Value of Cocreated Innovation?

Ajmal Hafeez, University of South-Eastern Norway, Norway Marit Engeset, University of Southeast Norway, Norway

Although much research have focused on how consumers derive value from engaging in self-production less focus has been given to how such offerings can be communicated to consumers. We posit that messages focuses on upskilling or ways in which engaging in self-production can improve consumers’ skills increases adoption of such products.

How Estimating One’s Minimum or Maximum Spend Affects Total Expected Expenditure on a Shopping Trip

Eunha Han, Monash University, Australia Harmen Oppewal, Monash University, Australia Eugene Chan, Monash University, Australia Luke Greenacre, Monash University, Australia

We investigate the effect of considering one’s possible minimum and/or maximum spend on the expected total spend during a grocery shopping trip. We propose that considering one’s maximum will increase the effect of unpacking on magnitude estimation while consideration of the minimum spend will decrease the effect.

Social Exclusion Causes People to Share Blatantly False Stories in the Service of Affiliation

Hyerin Han, University of Minnesota, USA Hyun Euh, University of Minnesota, USA

The present study demonstrated that participants who felt excluded (vs. accepted) were more likely to share fake news which was high in emotional intensity even though they knew that these news stories were highly likely to be false but only when doing so boosted their chances to communicate with others.

Do Happiness and Meaning in Life Uniquely Predict Consumer Well-Being?

Xianyu Hao, University of Minnesota, USA Kathleen Vohs, University of Minnesota, USA

We tested the influence of happiness and separately meaning on consumer well-being once their overlapping variance was removed. Two studies testing life views psychological needs and interpersonal outcomes indicated that happiness was a stronger predictor than meaning — thereby suggesting doubts about the influence of meaning on well-being. 1198 / Working Papers Consumer Reactions to Social Media Brand Blunders

Ceren Hayran, Ozyegin University, Turkey Melis Ceylan, Bilkent University, Turkey

This research explores how brand blunders - humorous and accidental business mistakes - that take place on social media influence consumers’ attitudes and relationship with the brand. Two studies show that loyal (vs. non-loyal) consumers are more negatively influenced by blunders as they think blunders are not funny but offensive.

The Impact of Gratitude on Consumer Maximization Tendency in Decision-Making

Dongjin He, Hong Kong Polytechic University, China Brent McFerran, Simon Fraser University, Canada Yuwei Jiang, Hong Kong Polytechic University, China

This research shows that gratitude can increase consumers’ maximization tendency evidenced by searching for more options spending more time on decision-making and choosing stores with larger assortment sizes. This effect appears because gratitude activates a self-improvement motivation. Furthermore we show the situations in which the proposed effect is weakened.

Feeling Justified to be a Jerk: Consumers Are More Likely to Complain For Someone Else Than For Themselves

Patrick Matthew Healey, Washington University, USA Cynthia Cryder, Washington University, USA Sydney Scott, Washington University, USA

Consumers’ participation in behaviors that lead to negative social perceptions increases when they are engaging in the behavior on behalf of another person. The present research is focused in the realm of customer complaints and nagging.

Communities of Stigmatized Knowledge: Social Exclusion, Political Sovereignty, and Globalized Capitalism

Tim Hill, University of Bath, UK Stephen Murphy, University of Essex, UK Robin Canniford, University of Melbourne, Australia

Prior studies show that consumers can be sceptical of the truth claims produced by the State universities and scientific communities. The proposed study investigates consumers who develop alternative claims to truth that are ignored or rejected by such dominant knowledge-making institutions.

Fractal Agency: Fetishization of Algorithms and the Quest For Transcendence

Soonkwan Hong, Michigan Technological University, USA

This research discusses critical topics related to algorithms and transhumanism to explicate and theorize the entanglement between big data marketable transcendence and fractal agency. Such discussion helps identify future research topics that highlight more specific characteristics of algorithms such as opacity messiness inscrutability hyper-legitimacy hyper-functionality and equal distribution of accountability.

The Impact of Social Crowding on Consumers’ Sensitivity to Price Magnitude

Yuansi Hou, Queen Mary University of London, UK Ke Zhang, Shanghai University, China

Crowding is a widely observed phenomenon. Through a correlational study and two experiments the research provides evidence suggesting consumers are less sensitive to the magnitude of service prices and reveals the underlying mechanism of reliance on feelings (vs. cognitions) when making judgments in the more (vs. less) crowded conditions. Advances in Consumer Research (Volume 48) / 1199 Adding Human Cues in a Green Advertisement Makes it More Engaging

Junhui Huang, Tsinghua University, China Maggie Wenjing Liu, Tsinghua University, China

Making a green advertisement more engaging has always been a great concern for marketers and environmentalists. With two studies we show that subtle human cues in advertisements can increase the audience’s engagement with the advertisement.

“Brain Damaged” Celebrity Followers: An Exploration Of The Effect Of Celebrity’s Brand Endorsement On Followers’ Social Media Engagement

Mei Huang, Xihua University, China Xiaodan zhang, Peking University, China Fang Wan, University of Manitoba, Canada Yifan Chen, University of Manitoba, Canada

The current research explores the impact of celebrity’s brand-related posts on followers’ social media engagement. We find that brand endorsement types in celebrity’s social media post impacts their fans’ engagement behaviors (i.e. likes comments and shares) differently.

The Closeness Buffering Effect: How Closeness Weakens Negative WOM Impacts

Li Huang, Hofstra University, USA Hang Nguyen, Michigan State University, USA

Word of Mouth (WOM) has great impacts on consumer decisions. Contrary to the common belief of greater impacts from close friends over strangers we found that negative WOM from close (distant) others were viewed more subjective (objective) and thus become less (more) influential on consumers’ decisions.

Wine Attributes For Purchase Satisfaction: an Exploration of Gender Difference

Hyowon Hyun, Hanyang University, South Korea Dongyoup Kim, Hanyang University, South Korea Jungkun Park, Hanyang University, South Korea

This research explores how numerous wine decision attributes affect satisfaction. The wine attributes were categorized into three wine factors (i.e. basic intrinsic extrinsic) and positively associated with satisfaction. The results of hierarchical regression and experiment showed the gender difference in the relationship between wine factors and satisfaction.

Semantic Congruity of Price and Name: The Effect of Precise or Round Numbers and Sound Symbolism

Hiroaki Ishii, Aoyama Gakuin University, Japan Jaewoo Park, Musashi University, Japan Taku Togawa, Sophia University, Japan

Previous studies show that consumers associate precise or round numbers with specific concepts. In this study we explore the semantic association of precise or round prices with the concept of smallness or largeness and the effect of semantic congruity between the price and sound of brand name on product evaluation.

Money Can Corrupt Others But Not Me: Asymmetrical Perception of Future Self-Continuity

Anoosha Izadi, University of Massachusetts, USA Mahdi Ebrahimi, California State University Fullerton, USA

The present research examines how people perceive change in their (vs. others) fundamental characteristics after facing with hypothetical situational changes in their (vs. others) future. We predict that the relationship in not symmetrical meaning that people perceive more change when the situational change is related to others (vs. themselves). 1200 / Working Papers The Effect of Implicit Bias on Marketing Practitioners’ Decisions For Minority Consumers

Jorge Rodrigues Jacob, Columbia University, USA Martin Davidson, University of Virginia, USA Valerie Purdie-Greenaway, Columbia University, USA Tatianna Dugue, Columbia University, USA

The mismatch between the identity background of those that define products’ marketing strategies and those who purchase them may increase the likelihood of biased marketing strategies that target stigmatized consumers.

“We” Don’t Always Like Copycats: How Self-construal Influences Evaluation of Product Imitation

Seongun Jeon, Vrije Universiteit Amsterdam, The Netherlands Femke van Horen, Vrije Universiteit Amsterdam, The Netherlands Peeter Verlegh, Vrije Universiteit Amsterdam, The Netherlands Keith Wilcox, Columbia University, USA

Three experiments provide converging evidence that self-construal (interdependent versus independents) affects the evaluation of copycats (products imitating the trade-dress of leading brands). Interdependents evaluate high similarity imitations more negatively than independents. Such copycat evaluation is moderated by norms regarding copycatting for interdependents but not for independents.

Setbacks as Self-Control Replenishment in Consumer Goal Pursuit

Lan Jiang, Menlo College, USA Stephanie Dellande, Menlo College, USA Miranda Canniff, Menlo College, USA

This research investigates the effect of setbacks in goal pursuit. Extending the literature on self-control as a resource we propose that occasional setbacks can be seen as a chance for replenishment resulting in longer term goal-consistent behaviors. The frequency and duration of the setback will be studied as potential moderators.

“Scarce- Insincere” Heuristic: Signaling Effect of Limited Quantity in Scarcity Appeal

Yufang Jin, Dalian University of Technology, China Guang-Xin Xie, University of Massachusetts Boston, USA

This research reveals a novel “scarce - insincere” heuristic when consumers make inferences of limited quantities in scarcity appeals. Product scarcity can reduce perceived sincerity of suppliers and deter consumers from purchase. Real-world transaction data demonstrate the counterproductive effect of product scarcity and experimental data suggest the underlying process.

Do Majority and Minority Always Act Differently? How Relative Group Size Affects Utilitarian Consumption

Yu Jinjun, City University of Hong Kong, China Yijie Wang, Hong Kong Polytechic University, China Nakaya Kakuda, City University of Hong Kong, China

This research investigates how relative group size affects consumers’ utilitarian consumption. We propose that consumers prefer utilitarian products more when they are situated in a relatively smaller group which is mediated by the need for self-improvement. Two experiments are conducted to test our hypothesis. Advances in Consumer Research (Volume 48) / 1201 Fostering Progressive Literacy Through New Media Consumption: How Memes, Podcasts, And Political Comedy Spur The Political Imaginary

Hunter Jones, Aalto University, Finland

This netnography of the increasingly popular digitally active and politically charged ‘Dirt Bag Left’ presents an original model depicting how consumption of political comedy in new media is used to foster progressive literacy and reignite stagnant socio-political imaginaries. In doing so it makes substantial contributes to Transformative Consumer Research.

I Hope Therefore I Save: The Positive Effect of Hope on Financial Decision Making

Stephen Juma, Virginia Tech, USA Mario Pandelaere, Virginia Tech, USA

Understanding what motivates people to save is critical in a time when most people do not have enough savings. We tested the prediction that hope motivates saving through an increased focus on the future. Results from three studies show that hope leads to savings and a boundary condition is identified.

“Others Seem to Move Further Than I Do”: How Perceived Change in Distance May Differ Depending on Who Creates the Distance

Wonsuk Jung, University of Wisconsin - Madison, USA Joann Peck, University of Wisconsin - Madison, USA

Our judgement on distance may be influenced by whether the distance in question is created by the self or the other. We propose that when the distance is created by the other the impact of that distance may seem greater compared to when the distance is created by the self.

Bibliometric Analysis of Immersive and NeurophysiologicalTools in Retailing

Shobhit Kakaria, University of Valencia, Spain Aline Simonetti, University of Valencia, Spain Enrique Bigne, University of Valencia, Spain

Given the evolving nature of marketing research and the prominence of neurophysiological and virtual tools this bibliometric analysis showcases the representative research trends in the past two decades and prospective directions in the domain of retailing.

Can Taste Predict Product Evaluation?: The Role of Variety-Seeking on Innovative Products .

Nakaya Kakuda, City University of Hong Kong, China Xiaolei Su, City University of Hong Kong, China Cheng Gao, Nanjing University, China

This research investigates the impact of taste sensation on innovative product evaluations. Specifically this research found spicy taste recall having a positive influence on product evaluation and bitter taste recall having the opposite effect. This effect is mediated by variety-seeking intention moderated by innovation types: radical innovation vs. incremental innovation.

The Effect of Gratitude Expression on Word-of-Mouth Intentions

Christine Kang, California State University Long Beach, USA

This research finds consumers’ tendency to reciprocate firm’s favor is modulated when they were given a chance to express gratitude. Participants who were guided to express their gratitude reported significantly lower intention to generate WOM offline or online compared to those who expressed happiness or those in control condition. 1202 / Working Papers The Effect of Nostalgia on the Ideal Self and Moral Identity

Hedieh Karachi, Deakin Business School, Australia Jeffrey Rotman, Deakin Business School, Australia Andrea Vocino, Deakin Business School, Australia

While past research has demonstrated that nostalgia leads to self-continuity (Sedikides et al. 2015; Sedikides et al. 2016) the current research suggests that self-continuity is linked to just thinking about the distant past. However only nostalgia reduces the discrepancy between one’s ideal-self and heightens moral identity.

The Effect of Assortment Categorization and Construal Level on Consumer Satisfaction

Bianca Kato, University of Guelph, Canada Juan Wang, University of Guelph, Canada

Our work explores the interplay between construal levels and assortment categorization on consumers’ satisfaction. We found that consumers are more satisfied with an assortment with fewer (vs. more) categories when under a concrete construal and that this effect is mediated by fluency perceptions attenuated by consumers’ familiarity with the product.

What Makes a Product Cute: Infantile Attributes Influence Perceived Cuteness of Products

Carolyn Wells Keller, Northwestern University, USA Neal Roese, Northwestern University, USA

In four studies we investigate the influence of infantile attributes on perceived cuteness in anthropomorphized products. We find that roundness of a product’s base or ‘body’ has a particularly consistent impact on perceived cuteness. Infantile attributes also had a stronger effect on perceived cuteness in utilitarian (vs. hedonic) products.

What Makes an Advertisement Offensive? The Interplay of Prescriptive and Descriptive Norms

Saeid Kermani, York University, Canada Peter Darke, York University, Canada

This research explores the joint impact of prescriptive and descriptive norms on consumers’ intention to complain about an offensive advertisement. Two experiments provide support to the proposition that descriptive norms influence intention to complain via perceptions of harm when a prescriptive norm has been violated.

A Time-Oriented Explanation Of Desire For Scarce Item In Online Retailing

Dongyoup Kim, Hanyang University, South Korea Jungkun Park, Hanyang University, South Korea

Throughout series of four studies this research examines the effect of scarcity on consumers’ purchase intention mediated by perceived control over time. The effect was moderated by two different online retail strategies (i.e. pressurizing by blinking the limited quantity and price discount promotion) and consumers’ initial time availability for shopping.

Consumers’ Responses to Sustainable Smart Home Services: The Role of Regulatory Focus

Moon-Yong Kim, Hankuk University of Foreign Studies, South Korea

This research examines the role of consumers’ regulatory focus in their responses to sustainable smart home services. The results indicate that consumers who are not familiar with smart home services have more favorable attitudes toward hedonic smart home services when they are promotion-focused (vs. prevention-focused). Advances in Consumer Research (Volume 48) / 1203 Consuming Beauty in Place of Power: Gender and Culture Moderation

Sahoon Kim, University of Illinois at Urbana-Champaign, USA Carlos Torelli, University of Illinois at Urbana-Champaign, USA

How would consumers’ psychological sense of power – which can be influenced by numerous factors such as seller-buyer interaction – affect their apparels and beauty consumption? We theorize and demonstrate that men (women) consume more when feeling powerful (powerless). This tendency is stronger for individuals with certain cultural orientations.

From Powerlessness To Variety-Seeking When Choosing For Others: Making Gift Choices For A Romantic Partner

Sahoon Kim, University of Illinois at Urbana-Champaign, USA Sharon Shavitt, University of Illinois at Urbana-Champaign, USA Carlos Torelli, University of Illinois at Urbana-Champaign, USA

Many goods are purchased for others (e.g. gifts). When choosing a gift for a romantic partner would one’s sense of power affect the choice? We demonstrate that feeling powerless increases variety seeking in choices made for the partner mediated by the reduced certainty in one’s partner’s preferences associated with powerlessness.

“I’m Cold But Feeling Warm”: Effect of Cold Temperature on Evaluations of Ad Message

Myung Joo Kim, Ewha Womans University, South Korea Jisoo Shim, Ewha Womans University, South Korea Eunice Kim, Ewha Woman’s University, South Korea

The present study investigates whether the temperature individuals physically experience influences their perception of social affiliation and evaluations of ad messages. This study shows that coldness generates more positive attitudes toward psychologically close (vs. distant) ad messages. The implications of our findings are discussed.

The Interactive Effects of Identity Salience and Accountability on Product Judgment

Claire Heeryung Kim, McGill University, Canada

A salient identity motivates consumers to think and behave consistently with that identity. Expecting to be required to explain a judgment to others later might increase motivation to engage in comprehensive information processing. Thus this research proposes that the act of providing reasons will diminish the identity salience effect.

When Is a Good Thing a New Thing? How Upcycling Product Display Affects Consumer Novelty Perception and Advertising Evaluation

Junghyun Kim, NEOMA Business School, France Youngju Kim, NEOMA Business School, France Junbum Kwon, University of New South Wales, UK

How does the position of source materials influence attitudes toward a repurposed product? We investigate the impact of product communication strategy on consumer attitude toward an upcycled product. Two studies demonstrate that how a company presents the past identity of an upcycled product influences consumers’ novelty perception and advertising evaluation.

Generation Z Consumers’ Luxury Goods Ownership Types and Purchase Intention Of Luxury Goods Using Collaborative Redistribution Platforms

Jihyun Kim-Vick, Kent State University, USA

The findings of this study extend our understanding of the young adult consumers’ channel choice intentions toward purchasing brand new or previously owned personal luxury goods via traditional retail vs. collaborative redistributional platforms. Pragmatic implications for the luxury brand marketers as well as the online resale industry were provided. 1204 / Working Papers Syllabic Fluency

Dan King, University of Texas Rio Grande Valley, USA Sumitra Auschaitrakul, University of the Thai Chamber of Commerce, Thailand

We show that the syllabic structure of brand slogans and medical claims can influence consumers’ judgments of truth. When the number of syllables of the brand name matches the number of syllables in the outcome consumers feel a metacognitive sense of processing fluency which increases judgments of truth.

It is Better to be Unknown Than Known: Mixed Use of Less Known and Well-Known Luxury Brands Can Elicit Higher Status Inference

Min Jeong Ko, Seoul National University, South Korea Kyoungmi Lee, Seoul National University, South Korea

Across four studies we find that mixed use of less known and well-known brands compared to use of all well-known or all less known brands increases perceived status when well-known brands are in luxury(vs. non-luxury) domain. This is because observers infer user’s desire to dissociate from lower class luxury users.

The Three-component Multidimensional Model of Self-image Congruence

Magdalena Kolanska, University of Zielona Gora, Poland Oleg Gorbaniuk, University of Zielona Gora, Poland Wilczewski Michał, University of Warsaw, Poland

The Three-Component Multidimensional Model makes it possible to assess and distinguish between the values of the brand incorporated into the consumer’s self-image which correspond to self-enhancement self-protection and self -verification motives (Sedikides Gregg 2008). In order to build the Model a series of 3 tests was carried out.

Losing Possessions and Subsequent Spending

Shruti Koley, Portland State University, USA Christina Kan, University of Connecticut, USA Chiraag Mittal, Texas A&M University, USA

This research examines how losing one’s possessions affects consumers’ subsequent spending on unrelated purchases. We show that the effect depends on consumers’ perceived wealth and their manner of losing. After losing their possessions wealthy (less wealthy) consumers spend more (less) subsequently but only when they blame themselves for the loss.

Smart Homes From a Consumers’ Perspective

Monika Koller, Vienna University of Economics and Business, Austria Fanny Springer, Vienna University of Economics and Business, Austria

Smart homes play a major role in the world of digital change. The aim of this paper is to shed light on smart homes from a consumers’ perspective beyond focusing on technological solutions. We explore major associations with smart homes and dig deeper into a comprehensive understanding of their acceptance.

Small-Talking Brands: Exploring Phatic Brand Communication on Social Media

K.B. Koo, University of Alberta, Canada Sarah G Moore, University of Alberta, Canada Jennifer Argo, University of Alberta, Canada

We introduce and develop a new construct: phatic brand communication. The main function of such communication is to express sociability rather than information to consumers (e.g. “Happy Tuesday!”). Using Twitter data and experiments we propose and test a conceptual framework of phatic brand communication including its components and potential outcomes. Advances in Consumer Research (Volume 48) / 1205 Striving For Social Media Reduces the Ability to Filter Out Visual Distractors

Christina Kößmeier, University of Duisburg-Essen, Germany Oliver B. Büttner, University of Duisburg-Essen, Germany

This research examined how visual distraction is influenced by consumers’ striving for social media. In three laboratory experiments we showed striving for social media – measured as craving for social media situational social exclusion and situational fear of missing out – decreased the ability to filter out visual distractions.

The Role of Visual Perspective in Influencing Donation Giving

Sining Kou, Renmin University of China, China

Based on construal level theory this research demonstrates that when the consumer’s mindset (i.e. concrete vs. abstract) activated by visual perspective (i.e. first vs. third person perspective) is aligned (vs. misaligned) with the construal level of the message donation intentions and behaviors increase (vs. decrease).

Immunizing Against Diversity: Feeling Disgusted Reduces Positive Diversity Beliefs

Afra Koulaei, Innland Norway University of Applied Sciences, Norway Luk Warlop, Norwegian School of Management, Norway Bjørn Ove Grønseth, University of South-Eastern Norway

This study concerns with the impact of disgust on the evaluation of social groups and more specifically on the reduction of positive beliefs about increasing diversity in social groups. Three studies provided the test of our prediction that disgust (both chronically and state) reduces positive diversity beliefs.

The Moral Deviant: Norm Violators Are Expected to Communicate Concretely

Afra Koulaei, Innland Norway University of Applied Sciences, Norway

Violation of social norms may signal that the target lacks morality. Two studies suggest that morality have a leading role over other basic dimensions of human social cognition (i.e. sociability and competence) in the impression formation process of norm violators. To signal their morality norm violators are expected to communicate concretely.

Friend or Foe? A Long-term Multi-Country Perspective on Consumer Animosity

Tinka Krüger, Kiel University, Germany Robert Mai, Grenoble Ecole de Management, France Wassili Lasarov, Kiel University, Germany Stefan Hoffmann, Kiel University, Germany

The Brexit or the US-Chinese trade conflict mark the resurgence of deglobalization and consumer animosity. Based on existing data from six countries from 2010 this study lays the foundation to examine whether animosity context changes result in altered individual consumption behaviors and hence evaluate the scope of animosity for practitioners.

Emotionality and Language Norms in Consumer Reviews: The Curious Case of Emoji

Polina Landgraf, IE Business School, IE University, Spain Nicholas Lurie, University of Connecticut, USA Antonios Stamatogiannakis, IE Business School, IE University, Spain Susan Danissa Calderon Urbina, University College Dublin, Ireland

How do emoji impact the persuasiveness of consumer reviews? We argue that effects of emoji in reviews are stronger for utilitarian than hedonic products. Two experiments show that effects of emoji on persuasion are consistent with an account based on emoji as a language rather than as markers of emotionality. 1206 / Working Papers Healthy or Not Healthy ? That’s Not the Main Question

Sarah Lasri, University Paris-Dauphine, PSL, France

Even if researchers tried to understand healthy consumption it was often through a normative lens. They try to figure out triggers and consequences without questioning how healthiness is constructed. We attempt to show that it is a social myth where consumers play a huge part to define constantly new moralisms.

Consumer Perceptions of Floating Signifiers in Spurious Health-related Product Claims

Olga Lavrusheva, Aalto University, Finland Alexei Gloukhovtsev, Aalto University, Finland Tomas Falk, Aalto University, Finland

Our study looks at the use of ambiguous product health claims in food marketing. In particular we focus on consumer interpretations of “floating signifiers”: claims that are ambiguous enough to escape precise definition and regulation while at the same time sufficiently “loaded” with health-related meanings from the consumers’ perspective.

The Truth is in the Tweet: Exploring the Impact of the Linguistic Authenticity of Politician’s Tweets on Their Personal Brands

Andrew Lee, University of Manitoba, Canada Mehmet Yanit, University of Manitoba, Canada Fang Wan, University of Manitoba, Canada

Politicians take to Twitter as a vehicle for connecting with voters and making their opinions known. Donald Trump became infamous for his outrageous tweeting style during the 2016 election. By using LIWC and experiments this research aims to unearth the mechanisms of authenticity with politicians on Twitter through their language style.

Truth Or Dare: Do People Behave Differently In Social Risk Vs. Financial Risk? The Inherent Difference Between Financial And Social Risk

Sunme Lee, University of Iowa, USA Catherine Cole, University of Iowa, USA Dhananjay Nayakankuppam, University of Iowa, USA

This paper examines systematic differences in risk preferences between financial and social contexts. People tend to be loss averse in the financial domain but gain averse in the social domain i.e. more (less) risk-seeking in the gain frame than in the loss (gain) frame in the social (financial) domain.

5th Best Hotel in America Vs . 2nd Best Hotel in New York: When Using Narrower Sets in Rank Claims Improves Consumer Judgments

Wei Li, Huazhong University of Science and Technology, China Yaping Chang, Huazhong University of Science and Technology, China Jun Yan, Huazhong University of Science and Technology, China

This research shows consumers evaluate a product or brand more favorably when it appears front in a narrow ranking list than later in a broad list. This effect arises only when consumers are maximizers. Perceived risk cognitive load and significance of the decision also moderate the effect. Advances in Consumer Research (Volume 48) / 1207 Imperfect Boundaries: The Effects of Boundaries on Perceived Healthiness of Healthy Food

Ruiqin Li, Renmin University of China, China

How do boundaries influence perceived healthiness of healthy food? The present research shows that expressing product information with boundaries fixes attention of consumers on the information which induces motivated reasoning and finally reduces perceived healthiness. In addition the effect of boundaries only exists in healthy (vs. unhealthy) food domain.

Over-Inference in Predicting Others’ Preferences

Xilin Li, University of Chicago, USA Christopher Hsee, University of Chicago, USA

Through five studies we document an “over-inference bias” whereby predictors over-rely on predictees’ preferences in one domain to infer their preferences in another domain. We also explore the underlying mechanism of this bias and articulate boundary conditions under which this bias can be attenuated or reversed.

Recall the Past: How Actors’ and Observers’ Perspective Affect Word of Mouth

Ran Li, Chinese University of Hong Kong, China

This research examines how the visual perspectives that people use to recall their past consumption experience (actor vs. observer) affect their subsequent sharing. Two experiments showed that taking an observer perspective reduces consumers’ sharing of negative word of mouth through the mediating role of attention-seeking motivation.

The Awed Holistic Thinker: The Effect of Awe Experience on Individuals’ Thinking Styles

Ran Li, Chinese University of Hong Kong, China Tao Tao, Hong Kong Baptist University, China

This research examines the effect of awe on people’s thinking styles. Seven studies showed that awe could promote holistic thinking (studies 1 2A-2D) through the mediating role of self-diminishment tendency (study 3). Moreover we demonstrated how the proposed effect influences consumers’ reactions to service failures (study 4).

Beauty Camera Makes You More Emotional!: The Effect of Beauty-filtered Selfies on Consumer Decision Making

Shuyu Liang, Lingnan University, China Tingting Wang, Sun Yat-Sen University, China

This research reveals a positive effect of beauty-filtered (vs. not) selfies on consumers’ reliance on affective considerations in decision making. Drawing on extant research on self-presentation motivation and confidence we propose and demonstrate in two studies that taking beauty-filtered (vs. regular) selfies makes consumers prefer affectively (vs. cognitively) superior options.

Can Gratitude Increase Individuals’ Sustainable Consumption Behavior? — the Mediating Effect of Time Discounting

Jianping Liang, Business School, Sun Yat-sen University, China Leilei Guo, Sun Yat-sen University, China

This research examines the influence of gratitude on individuals’ sustainable consumption behaviors. Results from a longitudinal study and three between-subjects experiments showed positive influences of trait gratitude and priming gratitude on consumers’ willingness to use green products through reduced time discounting. Connectedness to future-self moderated these effects. 1208 / Working Papers Social Media Amplifies Gift-givers’ Conspicuous Consumption Motivation

Shuyu Liang, Lingnan University, China Tingting Wang, Sun Yat-Sen University, China

This research reveals an interesting misalignment between gift giver’s and receiver’s considerations. Specifically we found that facing receivers who are social-media active (vs. inactive) givers prefer conspicuous (vs. not) gifts out of anticipation of receivers’ sharing the gifts on social media; however gift conspicuousness does not affect receivers’ sharing intention.

Consumer Knowledge and the Psychology of Opposition to Scientific Consensus

Nicholas Light, University of Colorado, USA Philip M. Fernbach, University of Colorado, USA

Communicating scientific evidence is a major challenge. We report three studies on the relationships between knowledge type and anti-scientific attitudes across seven scientific issues. We find that as extremity of attitudes increases objective knowledge decreases but subjective knowledge increases. However several issues show inconsistencies worthy of further examination.

The Influence of Fake News on Consumer Spending in Response to COVID-19

Nicholas Light, University of Colorado, USA Justin Pomerance, University of Colorado, USA Lawrence E Williams, University of Colorado, USA

Fake news has entered the global consciousness. While previous work examines interventions and direct consequences we examine how it changes consumer spending intentions during the COVID-19 pandemic. Two studies indicate that fake news induces uncertainty which increases both the desire to save money and the intention to spend it compensatorily.

How Legacy Motivation Influences Financial Risk-Taking

Wang Lin, Sun Yat-Sen University, China Chen Zengxiang, Sun Yat-Sen University, China Yun He, Sun Yat-sen University, China

The creation of legacy is one of the most fundamental motivations human beings possess. The current research investigates the impact of legacy motivation on consumers’ financial decision-making. Results of three experiments revealed that legacy motivation reduced consumers’ financial risk-taking tendency and explored why and when this effect occurred.

#Favoritethings: How Posting Your Favorite Possessions on Social Media Increases Happiness

Jingshi Liu, Business School City, University of London, UK Amy Dalton, Hong Kong University of Science and Technology, China Anirban Mukhopadhyay, Hong Kong University of Science and Technology, China

Posting favorite possessions on social media increases people’s happiness. This is because favorite possessions are unique and of personal meanings and thus resistant to the upward social comparisons prominent on social media. The happiness associated with posts of favorite possessions in turn increases liking of the social media platform. Advances in Consumer Research (Volume 48) / 1209 How Consumers React to Culturally Mixed Products: The Effects of Cultural Domain and Mixing Direction

Yingyue Liu, Sun Yat-Sen University, China Zhimin Zou, Sun Yat-Sen University, China Yimin Zhu, Sun Yat-sen University, China

This paper conducts two studies to examine the joint effect of cultural domain and mixing direction on consumers’ attitudes to culturally mixed products as well as the mediating effect of perceived cultural threat and the moderating effect of polycultural mindset which give implications to better understand the cultural mixing phenomenon.

Lonely Heart? Warm it up With Love: The Effect of Loneliness on Singles’ and Non-singles’ Conspicuous Consumption

Wei Liu, Xiamen University, China Zhaoyang Guo, Xiamen University, China Rui Chen, Xiamen University, China

Three studies show that loneliness increases singles’ conspicuous consumption but yields divergent effect for non-singles depending on their sociosexual orientation (SOI). In particularly loneliness has decrease conspicuous consumption among non-single people with low SOI but increase conspicuous consumption among those with high SOI. Mating motive mediates the above effects.

Will Highlighting Low Stock Level Speed Up or Impede Consumers’ Purchase Intention? An Exploratory Study

Mengmeng Liu, Chinese University of Hong Kong, China Jessica Y. Y. Kwong, Chinese University of Hong Kong, China

How would consumers respond when they see an item that they are considering has “only a few left”? Would this highlight of low- stock level speed up or impede their purchase decisions? Our findings suggest that the effect depends on the consumption type (utilitarian vs. hedonic).

The Effect of Cuteness on New Product Adoption: Moderating Role of Thinking Style

Hongyan Liu, Jinan University, China Choong W. Park, University of Southern California, USA

This research examines how exposure to cute product design influences new product adoption. We propose that exposure to cuteness increases perception of hedonic benefit (fun) and new product adoption. While this effect is moderated by thinking style. Three experiments were conducted to test the propositions. The results support our predictions.

Hiding in Plain Sight- Low Self-Concept Clarity and Ironic Consumption

Karan Pratap Singh Lohan, Hong Kong University of Science and Technology, China

Ironic consumption allows consumers to signal different meanings- one to people who detect the irony and one to those who don’t. Low self-concept clarity consumers are attracted to this dual (ambiguous) signal since it insulates them from the risk of self-instability as well as disapproval from others.

Does It Matter Whose Fault It Is? The Moderating Effects of Similarity and Product Substitution on Consumer Boycotts of Parent and Subsidiary Corporations

Yung-Chien Lou, National Chengchi Uniersity, Taiwan Wei-Chih Tseng, National Chengchi Uniersity, Taiwan

This study demonstrates that consumer boycott behavior (and consumer perceptions) as a response to a conglomerate’s wrongdoing is affected by the similarity between and degrees of product substitution of the parent and subsidiary corporations. 1210 / Working Papers Cultural Difference In Self-Consciousness On Empathy Toward Socially Responsible Consumption

Chi-Cheng Luan, National Chiao Tung University, Taiwan

According to Triandis’s theory people with individualistic cultures tend to use private self-consciousness whereas people with collectiv- istic cultures use public one. A shopping scenario was provided to examine participants’ responses. The result shows that cultural difference is found in the distinct effects of self-consciousness on empathy for socially responsible consumption.

Discussing The Moderating Effect of Construal Level On The Relationship Between Emotional Appeals And Sustainable Consumption

Chi-Cheng Luan, National Chiao Tung University, Taiwan Yun-Hui Wang, National Chiao Tung University, Taiwan

Moral emotions such as pride and empathy are found to relate to sustainable consumption. This research aims for the interaction between these two emotions and construal level based on the construal level theory. The findings can provide another direction of promotional strategies by simulteneously considering emotional appeals and construal level.

Priming Pleasure Through Smells: the Effect of Positive Odors on Categorization of Affectively Congruent Objects

Ramona de Luca, EAESP-FGV, Brazil Delane Botelho, EAESP-FGV, Brazil

Studies demonstrated that pleasant scents induce pleasurable experiences and choices. However the unconscious mechanism through which individuals make mental associations between odors and unrelated objects is unclear. This research explores the effect of odors on mental processes (categorization) when the odor and stimuli in other modalities match on their valence.

The Effect of Ambient Odors on Categorization and Semantically-Congruent Product Choices

Ramona de Luca, EAESP-FGV, Brazil Delane Botelho, EAESP-FGV, Brazil

Research in sensory and scent marketing has prioritized cognitive approaches in which consumer choices result from deliberate decisions. However consumers’ choices are largely regulated by pleasurable experiences that arise automatically. This research explores the unconscious process underlying the effect of ambient scent on consumers’ categorization and semantically congruent choices.

How Hope of Breaking Even Makes People Cling on Losing Stocks

Siria Xiyueyao Luo, Vrije Universiteit Amsterdam, The Netherlands Femke van Horen, Vrije Universiteit Amsterdam, The Netherlands Kobe Millet, Vrije Universiteit Amsterdam, The Netherlands Marcel Zeelenberg, Vrije Universiteit Amsterdam, The Netherlands

Four studies found that hope explains the disposition effect: people’s hope predicts the tendency to keep the losing stock; people who hold losing (vs. not-losing) stocks had a stronger hope of breaking even and thus were more likely to keep the losing stock; reducing the hope attenuates the disposition effect.

Closeness and Purchase Strategy Independently Influence Gift Choice

Dong Lyu, University of Nottingham, UK Jia Jin, Ningbo University, China

This research aims to examine whether gift givers process social distance and purchase strategy independently. We plan to adopt drift diffusion model to simulate the decisional dynamics of givers then to check whether the drift rates of the processing of social distance and purchase strategy choice are correlated. Advances in Consumer Research (Volume 48) / 1211 The Experiential Brain of Smartphone Users: How Smartphone Use Reshape Thinking Style

Jingjing Ma, Peking University, China Ying Zhu, University of British Columbia, Canada Jiajia Liu, Peking University, China Jingjing Wang, Peking University, China

Through two large-scale nationwide surveys with 24180 participants and an Implicit Association Test this study identifies an undocu- mented positive correlation between smartphone use/apps and experiential thinking. We revealed that this effect is driven by pleasure and social activities (e.g. watching movies and using social media) people perform on their smartphones.

Exploring the Concept of Sustainable Food Consumption

Natalia Maehle, Western Norway University of Applied Sciences, Norway

The purpose of this study is to explore how the concept of sustainable food consumption emphasizes the four aspects of sustainability. We conclude that it is a multidimensional concept focusing on the nature the human being the society and the optimal use of economic resources in the food chain.

Collaborative Consumption in a Subsistence Marketplace

Bhupesh Manoharan, Masters Union School of Business, India Krishanu Rakshit, Indian Institute of Management Calcutta, India

The absence of lateral exchange markets (LEM) has impeded researchers from understanding collaborative consumption in subsistence marketplace contexts. This work aims to understand how an interactive LEM shapes interdependent peer-to-peer engagement in a subsistence marketplace.

Contesting Stigma in the Online Space: An Institutional Perspective

Bhupesh Manoharan, Masters Union School of Business, India Krishanu Rakshit, Indian Institute of Management Calcutta, India

We explore how marginalised consumers tackle stigma in the online space by conceptualising the socio-political online consumption space as an institutional field and studying the dynamics between the upper and lower caste members and their efforts to distort the current institutional logic to establish their desired logics respectively.

The Effect of Transparent Packaging on Psychological Ownership and Preference

Eva Marckhgott, Wirtschafts University, Austria Bernadette Kamleitner, Vienna University of Economics and Business, Austria Zachary Estes, Bocconi University, Italy

Transparent packaging removes the visual barrier between the consumers and the product thereby allowing consumers to develop a more intimate relationship with the product. This in turn leads to a preference for products in a transparent packaging over products in an opaque packaging.

(No) Offense Taken: Value Creation vs Value Destruction by Brand-Offended Consumers

Ereni Markos, Suffolk University, USA Lauren I Labrecque, University of Rhode Island, USA Mujde Yuksel, Suffolk University, USA

Consumers often post negative comments on brand social media with intentions to harm (value-destruction). We examine if this action backfires instead strengthening observing consumers’ relationships with the brand. We illustrate that consumers attack brands on social media for different reasons and also investigate why consumers defend brands against others’ attacks. 1212 / Working Papers Patterns of Use, Essence Transfer, and Value: How Patterned Usage Impacts Object Valuation .

Nathanael S. Martin, University of Cincinnati, USA Noah VanBergen, University of Cincinnati, USA

Does how an object is used impact consumers’ valuation of the object itself? The present research seeks to investigate whether objects whose usage is illustrated by a specific pattern are deemed more valuable to consumers due to possessing more of the original user’s essence. Three studies offer initial support.

Are Ingroup Recommendations Always Better? The Role of Outgroup Credibility on Parental Purchases and Attitudes

Fábio Miguel Ferrony Varela Martins, Universidade Europeia, Portugal Márcia Maurer Herter, Universidade Europeia, Portugal Raquel Reis Soares, Universidade Europeia, Portugal Diego Costa Pinto, Nova IMS - Information Management School, Portugal

This paper uncovers how outgroup credibility can reduce ingroup bias. In particular this paper examines how credibility acts as a boundary condition to social influence of parents (ingroups) and pediatricians (outgroups) turning outgroups more influential than ingroups on parental purchase intentions and brand attitudes.

Customer Citizenship Behavior in Customer Networks: An Analysis in the Lens of Network and Consumer Value Theories

Renata Martins, Fundação Getúlio Vargas, Brazil Sofia Batista Ferraz, EAESP-FGV, Brazil Felipe Zambaldi, FGV/EAESP, Brazil

Most studies that investigate customer citizenship behavior (CCB) in customer networks neglect the impact of networks’ structures on it. Our work is the first to fulfill this gap proposing a framework showing the extent to which CCB depends on the network’s structure and on the purposes customers expect to achieve.

Eating More to Save the Planet: Political Ideology, Food Waste, and Overconsumption

Erick M. Mas, Vanderbilt University, USA Kelly Haws, Vanderbilt University, USA Kelly Goldsmith, Vanderbilt University, USA

When facing the tradeoff to waste or overconsume liberals may engage in maladaptive eating behavior by overconsuming to offset food waste because they are motivated to reduce environmental impacts. Conservatives who are less concerned with these issues do not make consumption choices predicated on reducing environmental harm.

The Impact of Indulgent Consumption Images on In-Store Decision Making

Maxine Materne, University of Hamburg, Germany Mark Heitmann, University of Hamburg, Germany Christina Schamp, University of Mannheim, Germany

Marketers assume that displaying indulgent consumption facilitates mental simulation and increases purchase intentions. One field experiment and two simulated shopping experiments suggest that showcasing indulgent consumption is context-dependent such that the positive effect found in laboratory experiments for advertisement related settings does not generalize to images displayed in-store. Advances in Consumer Research (Volume 48) / 1213 When and Why an Entertaining Story Triggers Brand Buzz

Koji Matsushita, Chuo University, Japan

This research proposes that an entertaining story with no brand appearance can boost brand word-of-mouth (WOM). When highly transported consumers are aware that a brand might utilize a story to strengthen the brand image (salient persuasion knowledge) they become motivated to engage in brand WOM activity.

Feeling Hopeful in Response to High/Low Prototypical Identity-linked Appeals: The Distinctive Roles of Affective and Cognitive Social Identity

Miriam McGowan, University of Birmingham, UK Louise May Hassan, Bangor University, UK Edward Shiu, Bangor University, UK

Research shows identity-linked messages need utilize established social group characterizations to be effective. We find priming consumers’ identity using less/more established (low/high prototypical) characterizations has differential effects on purchase intention. The emotion hope mediates this process. Three experimental studies with US MTurk samples using gender/nationality identities supported our hypothesized effects.

The Role of Fluency in Identity-Linked Marketing

Miriam McGowan, University of Birmingham, UK Edward Shiu, Bangor University, UK Louise May Hassan, Bangor University, UK

Across three studies we examine low identifiers in the context of identity-targeted messages. Drawing on fluency literature we explore how low identifiers process identity relevant information differently to high identifiers. Low identifiers process identity-linked messages less fluently than primes depressing their product evaluation. The opposite occurs for high identifiers.

The Effect of Perspectives in Food Pictures on Unhealthy Food Choices

Eva Meersseman, Ghent University, Belgium Iris Vermeir, Ghent University, Belgium Maggie Geuens, Ghent University, Belgium

Consumers choose less unhealthy food when seeing pictures of food shot in a top perspective vs. a diner’s eye perspective. We show that lower familiarity with seeing food in top view decreases product vividness and subsequently lowers the need for instant gratification. Hence less unhealthy food is chosen.

The Impact of Dynamic Advertising With Zoom Lens on Consumers’ Comparative Decision-Making Preferences

Lu Meng, Renmin University, China

This research proposes the impact of merchants ‘zooming and pulling lenses on consumers’ comparative decision-making methods in dynamic advertising on online shopping platforms and the perception distance’s mediator role in this process. 1214 / Working Papers “Show Me Your Basket, I Show You What Drives Your Food Waste” Identifying Drivers of Shoppers’ Food Waste and Ways to Reduce It

Sybilla Merian, University of Zurich, Switzerland Petra Tipaldi, University of Zurich, Switzerland Klaus Fuchs, ETH Zurich, Switzerland Martin Natter, University of Zurich, Switzerland

The increasing debate on food waste puts pressure on consumers and sellers. We propose a scalable approach to identify food waste predictors based on consumers’ automatically logged loyalty card data. Identifying characteristics triggering food waste we propose a non-invasive intervention informing consumers and help them to reduce the problem.

Consumer Experience Of Working From Third Places

Laetitia Mimoun, City University of London, UK Adèle Gruen, Goldsmiths, University of London, UK

This research contributes to the literature on consumer experience in third places. We document the emerging value that a growing body of working consumers derive from third places as work accelerators. We also show how this evolution may generate conflicts and confusion despite sustaining the social ideology of third places.

Any Takers For Male Alex Vs . Female Alexa? The Impact of the Gender of Voice Technology on Consumer Choices

Nirajana Mishra, Boston University, USA Geeta Menon, New York University, USA

In the future consumers’ interactions with voice technology will only increase. Most of these voices are currently female. Through an initial set of studies we find that gender of voice technology brings stereotypes to bear on consumers’ judgments that translates to products.

Masculine Servicescapes? Examining the Intersections Between Men’s Identity Construction and Built Commercial Environments

Risto Moisio, California State University Long Beach, USA Mariam Beruchashvili, California State University Northridge, USA

We examine how servicescapes built commercial environments partake in the process of masculine identity construction. Using men’s recreational training at the Mixed Martial Arts dojos/gyms we demonstrate how a physical space of a dojo/gym contains and constructs a specific identity project the fighter ideal that would be difficult to realize elsewhere.

The “Why” of Spending Matters: The Role of Extrinsic and Intrinsic Goals on Consumer Spending Choices and Hedonic Value Estimations

Olaya Moldes, Cardiff University, UK

Past research suggests that what we buy (experiential vs. material; prosocial vs. proself) impact one’s well-being. We propose that why— the motivation behind—not only determines the spending choices made but also the well-being experienced. Three studies test the influence of dispositional and situational intrinsic and extrinsic goals in consumption. Advances in Consumer Research (Volume 48) / 1215 The Effects of Virtual Compared to Real Eating Companions on Unhealthy Food Intake.

Saar Mollen, University of Amsterdam, The Netherlands Sindy Resita Sumter, University of Amsterdam, The Netherlands Nadine van der Waal, Tilburg University, The Netherlands Pascale Kwakman, University of Amsterdam, The Netherlands Nynke van der Laan, Tilburg University, The Netherlands

In the current experiment we investigated whether social models in VR have a similar influence on eating behavior as real-life models and whether similar processes underlie their influence. Participants’ intake was influenced by that of the social model irrespective of whether this person was present in real-life or VR.

Misperception of Multiple Risks: The Role of Categorical Reasoning

Mehdi Mourali, University of Calgary, Canada Zhiyong Yang, University of North Carolina at Greensboro, USA

How do consumers combine multiple risk items when forming overall risk impressions? Our study shows an intriguing pattern where an option with multiple risks is judged to be less risky than the same option with only one of these risks. We attribute the findings to categorical reasoning about risk.

Is There Room For Social Stimuli in Green Ads?: An Investigation of the Use of Human Face in Green Advertisements

Brishna Nader, Fontys University of Applied Science, The Netherlands Meng-Hsien (Jenny) Lin, California State University Monterey Bay, USA Iris van Hest, Fontys University of Applied Science, The Netherlands

This paper explores the effectiveness of human face (a social cue) used along with green message (an informational cue) in green advertisements. We find that while green ads are generally preferred over non-green ads it is less preferred when a face (vs. no face) is present.

“When First is Best”: Why Brand-First is Superior to Attribute-First During Sequential Presentation of Product Information

Priya Narayanan, Indian Institute of Management Kozhikode, India Arvind Sahay, Indian Institute of Management Ahmedabad, India

Four studies including a mall intercept show that consumers’ willingness to pay (WTP) for a product is higher when brand information is presented before attribute information an effect mediated by brand-elicited affect. This research contributes to feelings-as-information ordered presentation of product features and derives implications for marketing communication and packaging.

“Practices As Institutional Fractals: Zooming in on Fika, Zooming Out to Lagom”

Angeline Nariswari, California State University Monterey Bay, USA Kaisa Koskela-Huotari, Karlstad University, Sweden Teea Palo, University of Edinburgh, UK

This paper studies Swedish coffee drinking ritual fika to understand institutional maintenance by looking beyond micro-level practices to examine the role of broader level institutional arrangements. The study identifies how fika is supported by other practices that display self-similarity regardless of scale—fractals—that altogether refract a common societal-level value. 1216 / Working Papers Mouse Tracking in E-Commerce: Assessing Choice Difficulty Via Consumer’s Hand Movement

Patrick Wolfgang Neef, University of Innsbruck, Austria

Pre-decisional information processing is strongly linked to consumer choice. Since consumers increasingly purchase through online channels the question arises whether other computer-based technology can be applied to better understand consumer decision-making. One purpose of the present inquiry is to understand how difficulty of the decision-making process affect hand movements.

I Enjoy Being Mii: How Knowing Yourself Impacts Avatar Enjoyment

Emma Neybert, University of Cincinnati, USA Rashmi Adaval, University of Cincinnati, USA

Often consumers assemble avatars to resemble themselves. However avatars often fail to fully capture all aspects of their creator resulting in “poor copy syndrome” lowering enjoyment. In one provocative study we find that contingent on one’s self-concept clarity perceived personality similarity between the creator and their avatar impacts avatar enjoyment.

The Effects of Harmonics on the Tactile Perceptions of Products

Mayuko Nishii, Waseda University, Japan

Congruence between product elements and sensory cues elicits positive responses from consumers. However we show that when listening to soft-toned music that seems appropriate for products such as soft blankets consumers tend to be less sensitive to their haptic perceptions. Therefore we propose the role of comfort.

Cultures of Product Reviewing

Mikkel Ørholm Nøjgaard, University of Southern Denmark, Denmark Niklas Woermann, University of Southern Denmark, Denmark

How do expert reviews and online user reviews construct product quality information? This paper explores the social processes of product reviewing. It conceptualizes expert reviewing and user reviewing as two distinct ‘epistemic cultures’ and shows that these cultures vary across three epistemic processes: generating evidence evaluating evidence and presenting evidence.

Examining the Structural Relationship Between Socioemotional Comparisons and General Risk Preference

Dallas Novakowski, University of Calgary, Canada Mehdi Mourali, University of Calgary, Canada

Past research suggests that feelings of relative disadvantage (i.e. envy relative deprivation) have a causal role in eliciting domain-general risk-taking behaviours. This study used structural equation modelling in a crowdsourced sample (n=804) and found that feelings of relative disadvantage account for significant variance with a general factor of “risk preference.”

Entitled to Be Served: Does Self-Checkout Make Customers Feel Less Rewarded?

Farhana Nusrat, Drexel University, USA Yanliu Huang, Drexel University, USA

We examine how self-checkout service impacts customer loyalty compared to regular checkout service. In two studies we show that self-checkout service compared to regular check out make customers feel less rewarded less satisfied with their purchases and less likely to return to the store. Advances in Consumer Research (Volume 48) / 1217 The Effects of Similarities- vs Differences-focus on Fake News Processing

Gergely Nyilasy, University of Melbourne, Australia Bernice Plant, Monash University, Australia

The fight against fake news currently lacks working interventions. This study explores individual cognitive differences to pave the way towards such interventions. Relying on comparison theory hypotheses test the effects of similarities- and differences proneness. Findings suggest that these factors indeed influence cognitive processing – if not summative truth and confidence judgments.

Just Teasin’: Why Poking Fun at Consumers Makes Brands Feel More Human

Demi Oba, Duke University, USA Holly S Howe, Duke University, USA

Can consumers handle being teased by a brand? Across three experiments (on three different advertising platforms) brands who tease consumers are consistently shown to be more anthropomorphic than purely funny or serious brands. We demonstrate that this anthropomorphism has positive downstream consequences for brand liking and self-brand connection.

Indulge Every Now and Then: Anticipating Indulgence Increases Indulgent Food Preferences Among Restrained Eaters

Ga-Eun (Grace) Oh, Open University of Hong Kong, China

This research examines the interactive effects of anticipating indulgent consumption and dietary restraint on indulgent food preferences for immediate consumption. Three experiments reveal that anticipating indulgent food consumption increased restrained eaters’ indulgent food preference for immediate consumption whereas it often reduced unrestrained eaters’ indulgent food preference.

The Status Implications of Age Perception and Conspicuous Consumption

Ga-Eun (Grace) Oh, Open University of Hong Kong, China

This research examines how age perception has different implications regarding status among men and women. As a result of the different status implications of age towards men versus women subjective age perceptions have different effects on the conspicuous consumption of men and women.

Impact of Inverted Packaging on Brand Memory

Nur Yazgan Onuklu, Temple University, USA Maureen Morrin, Rutgers University, USA

In this paper we explore the influence of upside down packaging on brand recall and recognition. With two studies we show that unaided recall and recognition memory are improved when there is an upside down bottle presented among regular bottles compared to an all regular bottle presentation.

Paying For Free Products Stops Consumers From Committing to Other Brands: Role of Perceived Control

Maria Ortiz, Concordia University, Canada Arani Roy, McGill University, Canada

Consumers are susceptible to encounter non-zero costs for products which they think should be available for free (e.g. online gaming apps) leading them to a reduced feeling of control. We examine the effect of reduced perceived control on consumers’ commitment to brands they subsequently encounter using predictive control theory. 1218 / Working Papers “Run, Forrest, Run: How Intense Athletic Pursuits Induce Therapeutic Feelings”

Tatsiana Padhaiskaya, Aalto University, Finland

This research aims to uncover how consumers’ understanding of their own bodies informs the production of therapeutic feelings. My ethnographic inquiry in the context of ultra-running illustrates a multi-stage process of in-depth engagement with the body that results in the therapeutic consequences of various magnitudes.

Objectification and the Acceptance of Inequality

Helena Palumbo, Pompeu Fabra University, Spain Gert Cornelissen, Pompeu Fabra University, Spain

Media and marketing communications often represent people in an objectified manner. The systematic exposure to such objectifying messages might affect various types of judgments that people make. In this paper we test whether it increases people’s tolerance towards economic inequality. We hypothesize that the effect is mediated by reduced empathy.

Recipe Ingredients Predict Digital Engagement With Food Media

Ethan Pancer, Saint Mary’s University, Canada Matthew Philp, Ryerson University, Canada Maxwell Poole, Saint Mary’s University, Canada

Two field studies (Buzzfeed’s Tasty & Yummly) demonstrate that food media based on recipes rich in saturated fats receive more reactions comments shares and recipe bookmarks even when controlling for meal complexity and meal occasion (e.g. dessert). Experimental evidence suggests this effect is driven by positive affect.

When the Light Bulb Goes Off, Goal Progress Increases

Mivena Panteqi, Vilanova University, USA

We propose that lighting ambience influences goal completion via perceptions of goal distance. That is the progress towards achieving a goal is perceived through the reduction in distance towards the targeted goal.

Does Anticipated Future Self-Regulation Increase Present Self-Regulation? Counteractive Construal of Intertemporal Choice

Jihye Park, Seoul National University, South Korea Youjae Yi, Seoul National University, South Korea

This research examines how anticipated future behavior can encourage current self-regulation. Findings show that consumers are more likely to engage in self-regulatory behaviors when they anticipate resisting (vs. succumbing to) future temptation. Results also show that this effect is accounted for by counteractive construal and stronger for non-impulsive consumers.

Personal Trainers’ Physical Appearance and Service Registration Intention: Perceived Competence as an Underlying Mechanism

Sangchul Park, Texas A&M University, USA Shinhyoung Lee, Sookmyung Women’s University, South Korea Hyun-Woo Lee, Texas A&M University, USA

This study investigates the effect of personal trainers’ physical appearance on fitness service registration intention via perceived competence. Specifically we hypothesize that consumers perceive muscular (vs. well-shaped) trainers as more competent which in turn boosts their registration intention for personal training services. Advances in Consumer Research (Volume 48) / 1219 The Impact of Natural Versus Attained Beauty on Service Evaluations

Jooyoung Park, Peking University, China Nathasya Pricilia B Kristianto, Peking University HSBC Business School, China

This research examines how women customers differently judge a service provider with natural or attained beauty and in turn evaluate their service. Two field experiments showed that obtained beauty is judged less favorably decreasing overall service evaluations.

Double Anchors in an Online Charitable Giving Platform: The Limitation of Small Suggested Amount Solicitation

Sohyeon Park, Korea University, Korea

This research proposes a two-stage anchoring process model that analyzes the anchoring behavior of consumers in charitable giving when a solicitation message contains double anchors such as suggested amount and target amount. Specifically transition of anchoring behavior under manipulation of amount size in suggested amount and target amount is observed.

“Turning the Other Cheek”: The Role of Face Orientation on Perceived Model Attractiveness and Product Evaluation

Jaewoo Park, Musashi University, Japan Charles Spence, University of Oxford, UK Hiroaki Ishii, Aoyama Gakuin University, Japan Taku Togawa, Sophia University, Japan

The present research demonstrates that people perceive a model showing their left (vs. right) cheek as more attractive even when the images happen to be otherwise identical. We also show that in an advertising context a model apparently showing their left (vs. right) cheek leads to more favorable product evaluation.

The Inescapable Quest of Happiness: Exploring How the Ideology of Happiness Shapes Contemporary Consumer Society

Gabrielle Patry-Beaudoin, Queens University, Canada Jay Handelman, Queens University, Canada

This paper draws attention to the ideology of happiness and examines how it influences a marketplace. We contribute to the literature on happiness in marketing by highlighting two cultural discourses of happiness that consumers and producers constantly negotiate: happiness as a virtuous project and happiness as a momentary pleasure experience.

When Technology Fails: Rage Against the Machine or Self-Control? Investigating Customer’s Negative Emotions and Coping Strategies in AI-Service Failure Scenarios

Giulia Pavone, Toulouse School of Management TSM Research UMR 5303 CNRS, France Lars Meyer-Waarden, Toulouse School of Management TSM Research UMR 5303 CNRS, France Andreas Munzel, University of Montpellier - Montpellier Research in Management (MRM), France

We investigate the impact of service failures on customers’ anger frustration and coping strategies when interacting with an artificial intelligence based chatbot compared to a human. Results show that in the same aversive situation emotional responses differ when customers are aware of interacting with a chatbot rather than a human. 1220 / Working Papers Sounds Healthy: Modelling Healthy Food Choices Through Music-Evoked Eye-movements in Consumers From Denmark and China-Tracking Study in Danish and Chinese Consumers

Danni Peng-Li, Aarhus University, Denmark Derek Byrne, Aarhus University, Denmark Raymond Chan, Chinese Academy of Sciences, China Qian (Janice) Wang, Aarhus University, Denmark

The present study explores how custom-composed soundtracks associated with the notions of “healthiness” and “unhealthiness” can modulate Danish and Chinese consumers’ eye-movement patterns and choice of specific food items. The findings of this study provide a better understand how specific auditory components collectively can nudge consumers in different market segments.

What is Transmedia Narrative Experience For Consumers of Arts and Heritage?

Christine Petr, Université de Bretagne Sud, France Elodie Jarrier, Université d’Angers, France Dominique Bourgeon-Renault, Université de Bourgogne Franche-Comté, Dijon, France Maud Derbaix, Kedge Business School Bordeaux, France

New technologies allow Transmedia Narrative to become a successful strategy to enlarge consumers’ experience with brands. We investigate the impact of transmedia on audience attachment considering arts and heritage. Conclusions offer additional conceptualizations to narrative transportation suggest dimensions for further research on consumers’ engagement and on narrative transportation measurement.

Influence of Childhood Socioeconomic Status on Indulgent Consumption: A Life History Theory Approach

T. Andrew Poehlman, Clemson University, USA Denny Huynh, Duke University, USA Oriana Rachel Aragón, Clemson University, USA Gavan Fitzsimons, Duke University, USA

Why do some people indulge now whereas others delay? The evolutionary framework of life-history theory predicts that preferences for indulgences should be influenced by stress and resource scarcity. The effect of stress depended on whether people grew up in a resource-scarce or resource-plentiful environment.

Model Citizens of The Empire: Mythology, Ethnic Identification, And British Indian Women

Anuja Pradhan, University of Southern Denmark, Denmark Margaret Kathleen Hogg, Lancaster University, UK Hayley Cocker, Lancaster University, UK

This study explores how macro-level identity myths such as the ‘model minority’ manifest and influence the lived experiences of ‘privileged’ ethnic migrant consumers such as professional British Indian women. It challenges our accepted norms of ‘dominant’ and ‘dominated’ consumers and shows the transformation of ethnicity across migrant generations.

How Religious Minorities Resolve Cultural, Market, and Religious Ambiguity

Theeranuch Pusaksrikit, Chulalongkorn University, Thailand Amna Khan, Manchester Metropolitan University, UK Andrew Lindridge, Newcastle University, UK

This research aims to explore how British Muslims engage with the religious cultural and market-derived consumption narratives surrounding Christmas festival of the religious dominant group. Utilizing qualitative method the findings identify three interconnected themes explaining how these participants engage (or not) with Christmas. Advances in Consumer Research (Volume 48) / 1221 How to Convince Consumers That “You Do Not Always Get What You Pay For”

Vahid Rahmani, Rowan University, USA Elika Kordrostami, Rowan University, USA

Findings of three studies showed that provoking suspicion could eliminate consumers’ reliance on price to judge quality. Furthermore findings showed that activating persuasion knowledge is likely to result in a schematic thinking style a greater primacy effect of the readily available information and higher price-quality perceptions.

Effects of Visual Attention on Intertemporal Choice

Jairo Ramos, University of Colorado, USA Kellen Mrkva, Columbia University, USA Leaf Van Boven, University of Colorado, USA

Consumers discount delayed outcomes. We hypothesized that this happens partially because people attend primarily to immediate outcomes; and that shifting attention towards future outcomes reduces discounting. Across three experiments participants cued to visually attend to future rewards discounted future rewards less than participants cued towards immediate rewards and control participants.

Managing Multiple Identities: A Case of Meat Consumption

Louise Randers, Department of Management, Aarhus University, Denmark John Thøgersen, Aarhus University, Denmark Alice Grønhøj, Aarhus University, Denmark

The study identifies that consumers can hold multiple hierarchically ordered identities with implications for consumption reduction and avoidance of various protein products. Identities can correspond or conflict. In the latter case strategies of managing identity salience or changing protein consumption are used and these mechanisms can be utilized in practice.

Do as I Say, Not as I Do: Figuring the Child as Digital Native Through Technology Ideology and Caregiver Consumption

Sophie Alexandra Reeves-Morris, Liverpool John Moores University, UK Shona Bettany, University of Huddersfield, UK

This early stage work seeks to add to the qualitative understanding of how children are becoming differentially socialised consumers through their smartphone engagement. To gain a holistic view the nature and dynamism of a carer’s technology ideology is being explored to see how this may create disparities in children’s smartphone socialisation.

The Friday Payday Effect: The Impact of Intraweek Payday Timing on Deserving Justifications and Discretionary Spending

Wendy De La Rosa, Stanford University, USA Broderick Turner, Northwestern University, USA Jennifer Aaker, Stanford University, USA

More than half of Americans are paid on Fridays. However little is known about how intraweek payday timing (e.g. Monday vs. Friday paydays) impacts consumer spending. Across a series of studies and an analysis of real-world spending we demonstrate that Friday paydays increase consumers’ deserving justifications and thus their discretionary spending.

Pain of Philanthropy: When Does Cash Help Charitable Donations?

Spencer Ross, University of Massachusetts Lowell, USA Sommer Kapitan, Auckland University of Technology, New Zealand

Businesses have found their costs of accepting cash payments increasing as consumers have shifted toward noncash payments. Given this societal decrease in cash we examine the role of payment mechanisms in donation behaviors. We find while noncash payments are beneficial in certain donation transactions cash remains more beneficial in others. 1222 / Working Papers Digital Resource Scarcity Affects Online Purchase Intent: The Moderating Role of Message Framing

Arani Roy, McGill University, Canada Ashesh Mukherjee, McGill University, Canada

Consumers often face scarcity of digital resources such as low battery-life or low data availability. Two experiments are conducted to examine the effect of digital resource-scarcity on consumers’ online purchase intent. Using regulatory focus theory we propose that the effect of scarcity on purchase intent is moderated by advertisement message-framing.

Algorithmic Decision-Making, Agency and Autonomy in a Financial Decision Making Context: An Experiment

Laszlo Sajtos, University of Auckland, New Zealand Benjamin G. Voyer, ESCP Europe, France Marion Sangle-Ferriere, ESCP Europe, France Billy Sung, Curtin University, Australia

Algorithmic decision making (ADM) plays an increasing role in consumers’ life. ADMs rely on customer information and can contribute to reducing consumers’ perceived autonomy which consumers may be reluctant to accept. Using a financial decision-making scenario we explore how human ADM and joint decision-making affects consumers’ agency and autonomy.

Towards a Better Understanding of Consumer Online Browsing (COB)

Fabio Shimabukuro Sandes, Sao Paulo School of Business Administration FGV-EAESP, Brazil Delane Botelho, EAESP-FGV, Brazil Yuliya Komarova, Fordham University, USA

Through in-depth interviews and four experiments we investigate the Consumer Online Browsing (COB) as a distinct type of search behavior. There are different effects of COB and similar constructs on dependent variables and they are differently moderated by variables that arose from the qualitative phase of the research.

Guilty Pleasures: Consumers Choosing Identity Conflicting Behaviors

Nikkita Sarna, University of Texas at Austin, USA Susan Broniarczyk, University of Texas at Austin, USA

We examine a real-world phenomenon where consumers knowingly and repeatedly act in an identity conflicting way for their own personal pleasure— guilty pleasures. Four experiments (n=~1000) demonstrate how guilty pleasures are behaviors that individuals choose to engage in even though they report identifying with the behaviors less than other behaviors.

Enjoying the Sale: The Case of Itinerant Retailers in Festivals

Prakash Satyavageeswaran, Indian Institute of Management Udaipur, India Sridhar Samu, Great Lakes Institute of Management, India Rajesh Nanarpuzha, Indian Institute of Management Udaipur, India Jossin Shaji, Indian Institute of Management Udaipur, India

Retailers have been seen primarily as commercial entities; itinerant retailers included. We consider itinerant retailer as consumer in the context of temple festivals where they are motivated by the consumption of the festival. We also consider their role as informants and suppliers of products associated with the festival to consumers. Advances in Consumer Research (Volume 48) / 1223 Digital Service Consumption, Blessing or Curse? Exploring the Effect of Persuasive Design Features on Mindfulness and Consumer Well-Being

Dorothea Schaffner, University of Applied Sciences and Arts of Northwestern,Switzerland Uta Jüttner, Lucerne University of Applied Sciences and Arts, Switzerland Anja Bruggmann, University of Applied Sciences and Arts Northwestern, Switzerland Tobias Véron, University of Applied Sciences and Arts Northwestern, Switzerland Philipp Wyss, University of Applied Sciences and Arts Northwestern, Switzerland Martina Bracher, University of Applied Sciences and Arts Northwestern, Switzerland Linda Lingg, University of Applied Sciences and Arts Northwestern, Switzerland

This study explores effects of mindfulness in digital service consumption. The findings of a diary research study provide mixed evidence: Consumer mindfulness in the digital service experience is negatively related to stress and positively related to satisfaction and well-being. However persuasive design features are not found to impact consumers’ mindfulness.

You Get What You Ask For: How the Question Asked Shapes Person Perception and Discussion of Controversial Topics

Ann Schlosser, University of Washington, USA Evelyn Olivia Smith, University of Washington, USA

Our research contributes to prior research by demonstrating that more important than comparing “how” to “why” is whether the question directs the respondent’s focus inward. We find that an internal (e.g. “how improve”) versus external (e.g. “how work”) question inflated people’s understanding of a topic and perceptions of disagreement.

I’m Not Too Generous: The Implications of a Novel Moderator of the Better-Than-Average Effect

Shoshana Segal, New York University, USA Yonat Zwebner, The InterDisciplinary Center (IDC Herzliya), Israel Alixandra Barasch, New York University, USA

The finding that people often rate themselves as better than the average (BTA) is widely understood to apply to positive traits such as intelligence and resourcefulness. We show however that the BTA effect is attenuated for the generosity trait: people do not (like to) think of themselves as too generous.

Consumer Experiences of Responsibilization in the Context of Digital Healthcare Services

Henriikka Seittu, Aalto University, Finland Alexei Gloukhovtsev, Aalto University, Finland Henri Weijo, Aalto University, Finland Tomas Falk, Aalto University, Finland

We study consumers’ experiences of increasing responsibilization in the context of online therapy services. We find that responsibilization ushers in a new subjectivity of “empowered consumer” that conflicts with the previously internalized subjectivity of consumer-patient. We detail how our informants engage in various practices of reflexivity to reconcile this conflict.

Sensory Attributes in Brand Extensions: The Role of Attribute Centrality in Evaluations of Fit

Samuel B Sekar, University of South Florida, USA

This paper examines the role of sensory attributes in consumer evaluation of brand extensions. Specifically we investigate how a consumer evaluates brand extensions on touch and smell-related attributes across multiple possible extensions. 1224 / Working Papers Differential Correlates of Pathological and Impulsive Buying

Benjamin G. Serfas, University of Duisburg-Essen, Germany Oliver B. Büttner, University of Duisburg-Essen, Germany Matthias Brand, University of Duisburg-Essen, Germany Astrid Müller, Hannover Medical School, Germany Patrick Trotzke, University of Duisburg-Essen, Germany

The present research differentiates pathological from impulsive buying both on a measurement level and on a conceptual level. Firstly we establish discriminant validity between constructs. Secondly we suggest a model explaining the behavioral manifestation of pathological and impulsive buying. Thirdly we analyze the impact on brand loyalty.

From Satisfaction of Consumers to Satisfaction With Consumers: Reverse Rating in the Sharing Economy

Jie(Doreen) Shen, University of Illinois at Urbana-Champaign, USA Aric Rindfleisch, University of Illinois at Urbana-Champaign, USA

The current research examines consumers’ perceptions of the reverse rating (i.e. consumer are rated by providers). It also investigates the impacts of platform positioning on consumers’ evaluation of a sharing platform that employs reverse rating and their future app usage intention as well as consumers’ star rating to providers.

The Interplay of Variety and Processing Strategy on Calorie Estimates

Liang Shen, University of Cincinnati, USA Fengyan Cai, Shanghai Jiao Tong University, China Robert Wyer Jr., University of Cincinnati, USA

Given the inconsistent findings about the relationship between food variety and calorie estimates we propose the effect of variety on calorie estimation and its downstream consequences depend in part on how those items are processed. Three studies confirm the proposed interactive effect and its underlying mechanism.

How Process Ambiguity in Rewards Influences Subsequent Donations: The Role of Mental Accounting

Haijiao Shi, Tsinghua University, China Rong Chen, Tsinghua University, China Xiaobing Xu, Hainan University, China

We find that consumers are more likely to donate rewards earned through an ambiguous process (vs. certain process) because they place these rewards in an independent account whereas rewards obtained from a certain process were booked as deductions from payment accounts. We also test two boundary conditions for the effect.

Arbiters of Controversy: The New Role of Brands in Contemporary Society

Justin Sieow, University of Connecticut, USA Robin A. Coulter, University of Connecticut, USA

Our research investigates the new role of brands as arbiters of social issues. Grounded in institutional theory we identity the various agents present within the brand moral landscape. We highlight the complexity present in the marketplace and explore the dynamic between brands consumers and third-party marketplace agents. Advances in Consumer Research (Volume 48) / 1225 The Counterintuitive Effect of Scarcity on Food Waste

Bonnie Simpson, Western University, Canada Rhiannon MacDonnell Mesler, University of Lethbridge, Canada Katherine White, University of British Columbia, Canada

Across three studies we examine scarcity’s effect on food waste behavior. Study 1 (a field study) examines food waste behavior in control versus scarcity-induced conditions. Study 2 explores scarcity’s effect on food-related behavior in lab. Study 3 tests temporal distance as a moderator of food waste within consumers experiencing scarcity.

The Interplay Between Reminders of Resource Scarcity and Quantity-Quality Trade-offs

Tanya Singh, Concordia University, Canada Laura Goodyear, University of Toronto, Canada Beatriz Pereira, Iowa State University, USA Caroline Roux, Concordia University, Canada

We examine how experiencing scarcity impacts quality-quantity trade-off decisions. Across four studies we find that people experiencing scarcity prefer higher quantity when trade-offs involves larger product quantities. Conversely people experiencing scarcity prefer higher quality when trade-offs involve smaller product quantities. We propose a theoretical framework to explain this counterintuitive result.

Social Influences on the Process of Remaining a Vegetarian

Lucie Sirieix, Montpellier SupAgro, France Gilles Séré de lanauze, Montpellier University, France Margot Dyen, University of West Brittany, France

If the socio-demographic profiles and motivations of vegetarians are well known (Ruby 2012) few studies have looked at the vegetarians’ trajectories (Cherry 2015). The present article redresses this deficiency by examining how social influences can play a role in the ability to maintain a vegetarian diet.

Choosing Randomly Displayed Products For Joint Consumption

Jiaqi (Flora) Song, University of Liverpool, UK Dongjin He, Hong Kong Polytechic University, China Yuwei Jiang, Hong Kong Polytechic University, China

This research investigates the impact of the consumption context on consumers’ product display preference. Two studies demonstrate that consumers have higher purchase intention for products with a randomized display in joint compared to solo consumption contexts an effect driven by the heightened need to belong in joint consumptions.

Environmental Friendliness and Product Design

Jiaqi Flora Song, University of Liverpool, UK Boyoun Grace Chae, Hong Kong Polytechic University, China Yuwei Jiang, Hong Kong Polytechic University, China

Two studies demonstrate that consumers judge products with a glossy rather than matte exterior design as less environmentally friendly. This effect is driven by the belief that glossier products are more processed and less natural and can lead to downstream consequences in regard to disposal behavior.

Mindful Brand: Development and Validation of a Mindful Brand Scale (MBS)

Ellen Campos Sousa, Florida International University, USA Jayati Sinha, Florida International University, USA

This research brings a novelty contribution to the marketing and mindfulness literature proposing a conceptualization and a measurement for a mindful brand from a consumer perspective. A set of characteristics associated with mindful brand was identified and a mindful brand scale (MBS) was developed and validated. 1226 / Working Papers Owning or Sharing? How Feeling Financially Constrained Decreases Participation in Access-Based Services

Emanuela Stagno, BI Norwegian Business School, Norway Klemens Knoferle, BI Norwegian Business School, Norway Luk Warlop, BI Norwegian Business School, Norway

Feeling financially constrained influences consumer attention and behaviors. We propose that this feeling also affects the propensity to engage in alternative consumption modes; namely access-based consumption. Two experiments test whether financially constrained consumers are less willing to participate in access-based services. The preliminary findings seem to not support our hypothesis.

The Role of Consumer Self-identities and Individual Tendencies in Food Waste

Violeta Stancu, Aarhus University, Denmark Liisa Lähteenmäki, Aarhus University, Denmark

There are several well-established negative consequences of food waste for the environment and society. Consumers’ environmental or frugal self-identities are associated with lower self-reported food waste. On the other hand people’s impulsive buying tendency and disgust sensitivity are important factors that associate positively with consumer food waste.

The Role of Multiple Self-Identities in Behaviours Across the Consumption-cycle in Food and Apparel

Catalin Mihai Stancu, Aarhus University, Denmark Alice Grønhøj, Aarhus University, Denmark Liisa Lähteenmäki, Aarhus University, Denmark

Consumers have several self-identities which impact sustainable behaviours. Frugal self-identity was found to have stronger associations with specific categories of behaviours in the consumption-cycle. While sustainable self-identity positively predicted behaviours across the consumption-cycle frugal thrifty or trendy self-identities had both positive and negative relations to behaviours outlining potential clashes.

Embracing and Rejecting “Sparking Joy”—Understanding Japanese Consumer Responses to Marie Kondo’s Original Book

Sumire Stanislawski, Tokyo International University, Japan Shuji Ohira, Chiba University of Commerce, Japan Kosuke Mizukoshi, Tokyo Metropolitan University, Japan

By examining responses to Marie Kondo’s original book consumer reactions to a taste regime counter to mainstream consumerism are explored. Those who embraced and rejected her message were analyzed using text mining. Some internalized new meanings surrounding consumption while others rejected her legitimacy as an influencer of taste.

The Darkside of Online Social Networks: Measuring the Negative Effects of Social Influence in Online Social Networks

Sabrina Stoeckli, University of Bern, Switzerland Fabian Bartsch, IÉSEG School of Management, Paris, France Henry Shen, IÉSEG School of Management, Paris, France

This research conceptualizes and develops a scale of Susceptibility to Social Influence in the context of Online Social Networks such as Facebook or Instagram. Three studies find support for the conceptualization and for a valid and reliable scale. Next steps for scale development and its future application are discussed. Advances in Consumer Research (Volume 48) / 1227 Table to Farm: Persuasive Messaging Influences Consumers’ Willingness to Compost

Nicolette Sullivan, London School of Economics, UK Rick Larrick, Duke University, USA Gavan Fitzsimons, Duke University, USA

Despite consumers’ concern for the environment one measure that could have a profound impact diverting waste from landfills household composting has had low adoption rates. This study investigates the effectiveness of persuasive messages on willingness to compost amongst different segments of consumers.

To Err Is Human, to Correct Is Algorithmic

Chengyao Sun, Washington University, USA Cynthia Cryder, Washington University, USA

Across three studies and nine scenarios we observe that people believe that algorithmic errors can be corrected however they do not believe that human errors can be similarly corrected. More specifically people trust a previously erring algorithm after correction but lose confidence in an erring human even after correction.

The Effect of Tragic Artist Biography on Consumer Evaluation of Artworks

Yeonjin Sung, University of Michigan, USA Seojin Stacey Lee, Seoul National University, South Korea Kiwan Park, Seoul National University, South Korea

Our research proposes that providing a tragic biography of an artist enhances the perceived quality of the artwork by allowing people to appreciate the depth of expression in the artwork (“tragic artist effect”). We also show that this effect occurs independently of alternative processes—empathy effort and eudaimonic fulfillment.

Choice Overload, Does it Really Exist?: A Neurophysiological Assessment of Consumers’ Responses to “the Paradox of Choice”

Anshu Suri, HEC Montreal, Canada Sylvain Sénécal, HEC Montreal, Canada Pierre-Majorique Léger, HEC Montreal, Canada

We contribute to the debate of choice overload by concluding that the psychophysiological measures collected through an eye-tracker and face-reader indicate that consumers not only experience an increased level in cognitive load when selecting from a large set of options but also experience negative emotions in the decision-making process.

What do You Mean by Love? Understanding Brand Love Experiences

Bruno Sutil, Sao Paulo School of Business Administration FGV-EAESP, Brazil Delane Botelho, EAESP-FGV, Brazil Diógenes Bido, Machenzie, Brazil

Just as there are types of interpersonal love there might be types or experiences of brand love. We add to the literature by developing a multidimensional differentiation scale based on the reasons people love a brand to typify their experiences. Four factors emerged: values identity emotional attachment and functionality. 1228 / Working Papers Packaged or Unpackaged? How Displaying a Product With Vs. Without Packaging Influences Product Evaluations

Courtney Szocs, Louisiana State University, USA Sara Williamson, SUNY Old Westbury, USA Adam Mills, Loyola University New Orleans, USA

Managers display products with and without packaging in stores advertisements and online seemingly arbitrarily. Would displaying a product encapsulated in a package as opposed to unpackaged influence consumers’ product evaluations? If so why? Three experimental studies address these research questions and contribute to the literature on packaging and display effects.

You Are Who You Follow: The Impact Of Fitness Influencers On Food Consumption And Identity

Amandeep Takhar, De Montfort University, Leicester, UK Thomas Magede, De Montfort University, UK Jennifer Walker, De Montfort University, UK

This paper builds on prior consumer research relating to food and identity construction. By looking at how online fitness influencers impact the food consumption and identity of young millennials in the U.K. it is evident that participants were influenced by online fitness influencers in negotiating their self and collective identity.

To Cleanse or to Enjoy? The Effect of Affectively Tagged Windfall Time on Consumption Choice

Arash Talebi, EDHEC Business School, France Easa Sahabeh, University of South-Eastern Norway, Norway

We investigate how consumption changes when people gain a windfall block of time with affective tags (positive vs. negative). We find an asymmetrical effect: negative windfall time encourages hedonic (vs. utilitarian) activities supporting a mood regulation account. However positive windfall time does not seem to influence the consumption pattern.

Eco-friendly Versus Polyester Gucci Handbags: The Effect of Matching Green Claims and Temporal Frame on Product Evaluation of Self-enhancement Brands

Nabanita Talukdar, Hult International Business School, USA Shubin Lance Yu, Peking University, China Esterina Nervino, Universita’ degli Studi di Modena e Reggio Emilia, Italy

We examine how message-orientations of green claims using temporal construal affects consumers’ evaluation towards self-enhancement brands’ sustainable products. Through a survey experiment we confirm that a high construal level and high temporal distance green message improved the product evaluation through higher processing fluency for high green orientation individuals.

Indulgent Consumption Signals Interpersonal Warmth

Qing Tang, Nanyang Technological University, Singapore Kuangjie Zhang, Nanyang Technological University, Singapore Xun (Irene) Huang, Nanyang Technological University, Singapore

Three studies demonstrated that showcasing indulgent consumption can have a positive signaling function on one’s perceived warmth. Therefore consumers also choose indulgent food items when they have the motivation to appear warm in front of others. These findings add to the literature of indulgent consumption and consumer signaling behavior. Advances in Consumer Research (Volume 48) / 1229 Consumption Opportunity and The Consideration of Consumption-Opportunity Cost

Tao Tao, Hong Kong Baptist University, China Ran Li, Chinese University of Hong Kong, China

For the first time the authors propose the concept of consumption opportunity and identify it as a scarce exclusive and valuable resource to a consumer. We further demonstrated that when making purchase decisions the consideration of the cost of consumption opportunity could induce consumers to choose the higher-quality smaller-amount option.

I Revenge For Others I Do Not Know: The Role Of Intercultural Competence In Moderating Customer Revenge To Service Failure Occurred To Adjacent Customer

Mohamed Sobhy Temerak, Kent Business School, UK Ruby Zhang, University of Reading, UK Cristiana Lages, University of Reading, UK

We examine the service failure from the observers’ perspective and how it influences their patronage and revenge intentions. Results from a 2x2 experiment revealed observer’s interpretation of service failure was largely influenced by their intercultural competence. They are more sensitive to discrimination and courageous to retaliate with higher intercultural intelligence.

Recycled Plastic on the Upswing: The Effect of Packaging Material on Perceived Healthiness and Purchase Intention

Joyce De Temmerman, Ghent University, Belgium Nico Heuvinck, IESEG School of Management, France Hendrik Slabbinck, Ghent University, Belgium Iris Vermeir, Ghent University, Belgium

This research explores the impact of recycled plastic on perceived healthiness and purchase intention. We find that consumers tend to perceive products in recycled (vs. regular) plastic packaging as healthier and thereby are more likely to purchase products in recycled plastic packaging.

The Impact of Logo Typeface Obliqueness on Product Efficacy Perception

Lefa Teng, Jiangnan University, China Jie Wang, The Hong Kong Polytechnic University, China Yuwei Jiang, Hong Kong Polytechic University, China

This research shows that the obliqueness of logo typefaces influences consumers’ efficacy judgment of products offered by the company. Compared with non-oblique typeface logos oblique typeface logos generate the perception of quicker product effect. This effect is attenuated when the typeface is heavy.

Singled Out For Public Praise: How Scarcity Boosts Demand For Public Recognition

Carina Thuerridl, University of Amsterdam, The Netherlands Bernadette Kamleitner, Vienna University of Economics and Business, Austria

Public recognition is more valuable but less popular than private recognition in crowdfunding and charitable giving. We show that limiting availability of recognition increases preferences for public but not for private recognition. Singling consumers out for public raise may thus be a simple cost-effective strategy to increase monetary contributions. 1230 / Working Papers Measuring the Dissemination of a Subculture Through Consumption . Investigating the Geek Culture Case .

Alexandre Tiercelin, Université de Reims, France Marion Garnier, Grenoble Ecole de Management - Univ Grenoble Alpes ComUE, France Arnaud Delannoy, Ecole de Management de Normandie, Laboratoire Métis, France

This research proposes to understand better the dissemination of a subculture within mass culture by measuring belonging and consumption in the geek culture case. Results confirm subcultural traits and practices and progressive dissemination showing the quantifiable reality of a subculture and the possibility of measuring its spreading.

Social Atmospheres: Priming, Aligning and Defining Shared Experiences of Place

Hill Tim, University of Bath, UK Robin Canniford, University of Melbourne, Australia

Atmospheres are the moods associated with places. Current marketing research explains that firms create atmospheres by controlling material features of servicescapes. Through an ethnographic investigation of English Premier League soccer however we show that atmospheres are also social experiences that define the mood of place through consumers’ embodied performances.

When Does Actual Versus Ideal Self-Congruence Matter? Moderating Effects of Construal-Level on Brand Attitudes

Taku Togawa, Sophia University, Japan Kazuyo Ando, Chiba University of Commerce, Japan

Drawing on construal level theory this study found that consumers with low-level construal tended to focus on actual self-congruence (ASC) with a brand and prefer brands with high ASC levels. Contrarily consumers with high-level construal focused on ideal self-congruence (ISC) and indicated positive attitudes toward brands with high ISC.

If I Think I Can I Can . Can’t I? How Social Mobility Beliefs Influence Your Financial Behavior

Patricia Torres, Florida International University, USA Alexandra Aguirre-Rodriguez, Florida International University, USA

This research explores how social mobility expectations influence one’s financial behavior intentions. Across three studies this research reveals that social mobility expectations impacts behavior intentions; financial self-efficacy mediates whereas financial literacy moderates this relationship. Findings are discussed considering their implications for research on expectancy financial goal pursuit and decision making.

Product Consecration and Market Expansion: The Case of Vitamin D

Lez Ecima Trujillo Torres, University of Illinois at Chicago, USA Alev Kuruoglu, University of Southern Denmark, Denmark Hazal Celik, University College Dublin, Ireland

Our study investigates the relationship between product consecration and market creation and expansion. Our findings indicate that the global Vitamin D market has expanded from 1996 to 2019 through a process of contested product consecration that involves the cultural and scientific consecration and commercial innovation of Vitamin D over time. Advances in Consumer Research (Volume 48) / 1231 How Customer Identification Can Lead to Dysfunctional Behavior

Huy Quoc Tran, University of South-Eastern Norway, Norway Marit Engeset, University of South-Eastern Norway Luk Warlop, BI Norwegian Business School, Norway

Recent research explores the role of social identity in consumer contexts. In this study we document the negative effect of customer identification on impoliteness. We found that entitlement mediated this relationship and that this effect was conditional on self-control. We also plan to investigate other mechanisms underlying this phenomenon.

The Relationship Between Self-Construal and Neural Processes Associated With Conforming to Consumer-to-Consumer Recommendations

Arina Tveleneva, University of Wisconsin - Madison, USA Xinyi Wang, University of Wisconsin - Madison, USA Matthew Brook O’Donnell, University of Pennsylvania, USA Emily B Falk, University of Pennsylvania, USA Christopher N. Cascio, University of Wisconsin - Madison, USA

The results suggest that those who have a stronger interdependent view of the self compared to a weaker interdependent view of the self are potentially more likely to conform to social feedback when their opinions misalign with others. Interdependent self is also significantly associated with conflict monitoring and mentalizing.

When Does Diversity Help or Backfire? Impact of Cultural Diversity on Brand Perception

Esther Uduehi, University of Pennsylvania, USA Julian K Saint Clair, Loyola Marymount University, USA Mitchell Hamilton, Loyola Marymount University, USA Americus Reed, University of Pennsylvania, USA

This paper is the first to show how diversity as a brand characteristic influences consumer perceptions of the brand’s authenticity. We find that majority-focused brands embracing diversity increases in authenticity while a minority-focused brand embracing diversity may lose authenticity. This loss of authenticity is particularly felt by minority consumers.

Console War (and Peace?): Analysing Cross-Community Interactions in Gaming Discussion Forums on Reddit

Denis Utochkin, Norwegian School of Economics and Business Administration, Norway

In this paper I employ a mix of netnography machine learning and network analysis to examine 12 gaming communities on Reddit. Treating each individual community as inseparable from its immediate context I explore how multiple loyalties and identities are managed by consumers engaging in conversations in a multi-community context.

In Defense of Happy Hedonism: Moving Beyond Materialism and Its Negative Impact on Well-Being

Rita Coelho do Vale, Catolica Lisbon School of Business and Economics, Portugal Rik Pieters, Tilburg University, The Netherlands

This research studies happy hedonism analyzing its relationship with consumer well-being. It explores the idea that pursuing pleasure through the mere consumption of experiences and objects might positively contribute to consumer well-being. Three laboratory studies were conducted to explore the happy hedonism phenomenon and to distinguishing it from consumer materialism. 1232 / Working Papers The Meaning and Transformation of Taste Vocabulary

Rohan Venkatraman, University of Melbourne, Australia Anna E Hartman, University of Melbourne, Australia Ilkka Ojansivu, University of Melbourne, Australia Erica Coslor, University of Melbourne, Australia

We conceptualize the vocabulary used in tasting notes as part of taste regimes. Through qualitative content analysis of whiskey tasting descriptions we found that consumers transformed institutionalised vocabularies to convey variable meanings extending our understanding of adaptation of the language of a taste regime.

“Luxurious and Responsible? Consumer Perceptions of Corporate Social Responsibility Efforts by Luxury Versus Mass-market Brands”

Marlene Vock, Amsterdam Business School, University of Amsterdam, The Netherlands

This research contributes to the inconclusive debate in the field of sustainable luxury on whether the adoption of Corporate Social Re- sponsibility practices is beneficial for luxury brands or backfires a dilemma known as the ‘CSR-Luxury paradox’. Different types of CSR practices might help to explain inconsistencies in previous findings.

The Effectiveness Of Online Recommendations For Private Labels Versus National Brands

Barbara Kobuszewski Volles, Ghent University, Belgium Anneleen Van Kerckhove, Ghent University, Belgium Maggie Geuens, Ghent University, Belgium

This research reveals that consumers when recommended with products online are more inclined to switch to private labels compared to national brands. We present two studies demonstrating this effect and propose that it occurs because recommendations serve more as a signal of quality for private labels than for national brands.

Magically Relieved or Taking Control? The Influence of Messages Aimed at Debt Distressed Consumers

Rory Waisman, University of Alberta, Canada Mohammed El Hazzouri, Mount Royal University, Canada Kelley Main, University of Manitoba, Canada Gerald Häubl, University of Alberta, Canada

Marketing messages used by the debt-remedy industry differentially impact consumers’ financial behaviors depending on the nature of the appeal and consumers’ level of debt. Appeals that focus on debt-relief can backfire whereas messages about taking control of debt have distinct marketing and public-policy advantages.

What a Cute Lemon! The Effect of Whimsical Cuteness on Willingness-To-Pay For Imperfect Produce

Darlene Walsh, Concordia University, Canada Camille Darriet, Concordia University, Canada

The goal of our research is to find a way to encourage consumers to purchase imperfect produce. As a first-step in achieving this goal we demonstrate that priming whimsical cuteness (i.e. adding “googly eyes” to imperfect produce) can increase consumers’ willing-to-pay for imperfect produce. Future research ideas are also discussed. Advances in Consumer Research (Volume 48) / 1233 A Discount Added is an Obstacle Created: The Impact of Multiple Discounts on Consumer Decisions

Xin Wang, Nanjing University, China Chunqu Xiao, Nanjing University, China Hong Zhu, Nanjing University, China

Based on numerical cognition we conducted four experiments to explore the effect of multiple discounts and found that consumers preferred a single discount to multiple discounts. Cognitive fluency and perceived sincerity sequentially mediated the effect. Consumers concern for sales moderated the effect.

Don’t Reveal It Yet! How Goal Disclosure on Social Media Impedes Goal Revision and Commitment

Liangyan Wang, Shanghai Jiao Tong University, China Zhining Yu, Shanghai Jiaotong University, China

People commonly disclose their personal goals publicly on social media platforms based on the conventional belief that disclosure increases goal commitment and improves performance. Results from two experimental studies suggest that this conventional belief only applies to easy goals but not for difficult goals.

Healthy-Angular, Unhealthy-Circular: The Effect of Shape on Consumers Health Choice

Jie Wang, Renmin University of China, China Jing Jiang, Renmin University of China, China Xiaodan Zhang, Peking University, China

This paper explored the matching effect between food healthiness and shape. We demonstrated that angular (vs. circular) shape increases consumers’ preference for healthy options which is because a “healthy-angular unhealthy-circular” match could enhance consumers ease of processing and subsequently influences consumers choice.

Love or Luck? The Effect of Self-Referencing Narratives on Used Product Consumption Behaviors

Yan Wang, Renmin, University of China, China Jing Jiang, Renmin, University of China, China

This research investigates the impact of selling narratives created by sellers on buyers’ purchase behaviors in the context of second-hand market. Across three experiments we find that self-focused (love-oriented) narratives drive low purchase intention but high willingness-to-pay whereas product-focused (luck-oriented) narrative generates both high purchase intention and willingness-to-pay.

Moderate Product-gift Complementarity Eliminates Devaluation Effect in Free-Gift-With-Purchase Offers

Yusu Wang, Texas A&M University, USA Nicholas J. Olson, Texas A&M University, USA

While past literature demonstrated a devaluation effect of the focal product in free-gift-with-purchase offers this research provides evidence supporting our argument that a moderate (vs. high) complementarity between the focal product and the free gift counters this devaluation tendency. Underlying mechanism is also accessed. 1234 / Working Papers Red in the Environment Increases Preference For Competent Brands

Yansu Wang, Renmin University of China, China Jun Pang, Renmin University of China, China Lingyun Qiu, Peking University, China

Across four studies this research demonstrates that exposure to red cues in the environment increases consumer preference for competent but not warm brands. This effect is mediated by competitive orientation and is attenuated when consumers purchase products for others than for themselves.

Robot Anthropomorphism in the Workplace and Employees’ Job Insecurity

Xue Wang, University of Hong Kong, China Sara Kim, University of Hong Kong, China Minki Kim, Korea Advanced Institute of Science and Technology, South Korea

In four studies we show that robot anthropomorphism in the workplace enhances employees’ job insecurity due to an increase in social comparison. The findings contribute to the literature on robot anthropomorphism and job insecurity and provides practical implications to organizations in terms of how to design robots.

Signs Telling Me to Do What Others Do: The Effect of Exposure to Public Rules on Conformity

Xiaoran Wang, Renmin University of China, China Jing Jiang, Renmin University of China, China Yan Wang, Renmin University of China, China

This research takes a novel perspective by proposing an exposure effect of public rules on conformity. Across three experiments we find that in public consumption context exposure to public rules reminds consumers of the presence of others which increases their public self-awareness and hence the preference for majority-endorsed products.

The Dark Side of Precision: Numerical Information and Consumers’ Unhealthy Choices

Jie Wang, Renmin University of China, China Jing Jiang, Renmin University of China, China Xiadan Zhang, Renmin University of China, China

The research explores how numerical precision affects consumers’ health behaviors and how illusory control functions as an underlying mechanism. Across three studies we find that when the numerical information is expressed in a precise (vs. imprecise) format it tends to induce consumers’ illusory control which subsequently increases risky health behaviors.

Visual Simplicity Signals Constructive Brand Authenticity

Yan Wang, Renmin University of China, China Jing Jiang, Renmin University of China, China Xiushuang Gong, Jiangnan University, China

Constructive brand authenticity is subjective and can be contextually determined. This research examines the consumer lay belief about the relationship between aesthetics and constructive brand authenticity. Across three experiments we find that consumers perceive a brand as more authentic when it is with visually simple (vs. complex) designs or displays. Advances in Consumer Research (Volume 48) / 1235 The Case-Gender Congruity Effect: Consumer Choices For Upper Versus Lowercase Brand Names

Na Wen, California State University Northridge, USA

A series of experiments find that the letter case of brand names (upper- vs. lower-case) affects consumer choices but that these effects depend on congruity between brand case and product gender. This research advances understanding of how seemingly subtle brand name characteristics affect gender perceptions and consumer behavior.

The Closer, The Worst: The Impact Of Social Distance to a Dissociative Group And Consumer Choices

Na Wen, California State University Northridge, USA

Two experiments test the idea that consumers are less likely to buy a product if it is associated with a near (vs. distant) dissociative group—which is mediated by construal-level mindsets and moderated by group conformity. This work advances understanding of how different types of dissociative groups affect consumer behavior.

You Are More Patient When You Experienced More: The Impact of Memory Markers on Consumers’ Decisions in the Future

Yingting Wen, ESSEC Business School, France Ayse Onculer, ESSEC Business School, France

This research investigates how retrospective time progression can influence consumers’ decisions in the upcoming future. By using different numbers of memory markers two studies show that people experience different retrospective time progression over the same duration and they have different levels of patience on waiting and intertemporal choice.

The Value of Personal Information - Consumers’ Valuations and Preferences For Personal Data And Privacy

Claudia Wenzel, University of Zurich, Switzerland René Algesheimer, University of Zurich, Switzerland

Data is the oil of our age. Despite its importance research is lacking consistent data value estimations. By combining current valuation methods this research offers a better estimate of the value consumers attribute to their data and by drawing on reference price theory minimiz- es the gap between different valuation methods.

The Unit-Size Effect Revised: A Win-Win Solution For Reducing Both Waist and Waste

Amber Werkman, University of Groningen, The Netherlands Jenny van Doorn, University of Groningen, The Netherlands Koert Van Ittersum, University of Groningen, The Netherlands

People consume less when food is presented in multiple smaller units compared to a few larger units however food waste will increase accordingly. We propose that this so-called unit-size effect is even more impactful during food selection when people select quantities themselves based on unportioned rather than portioned food.

When Do You Want This? How Delivery-scheduling Impacts Motivation Among Online Shoppers

Sara Williamson, SUNY Old Westbury, USA Courtney Szocs, Louisiana State University, USA

Online purchases require consumers to complete multiple sub-tasks such as product selection delivery-scheduling and payment. In this project we investigate how the sequence of these sub tasks can influence motivation in the purchase context. Specifically we investigate how scheduling delivery before making product choices can exhibit motivating effects. 1236 / Working Papers Are Co-branding Products More Attractive? Understanding the Role of Consumer Innovativeness and Self-construal

Sihan Wu, Tsinghua University, China Maggie Wenjing Liu, Tsinghua University, China

In this study we intend to develop and test a model considering that the consumers prefer co-branding product via consumer innovativeness. We further examined the moderation role of self-construal such that positive effect is stronger when individual is in a predominantly interdependent.

Effects of Masculinity and Femininity Incongruence on Consumer Responses to Brands

Xiu Wu, Concordia University, Canada Bianca Grohmann, Concordia University, Canada

This research examines the effect of incongruence between consumers’ gender identity and brand gender. Two studies find a significant main and interaction effect of masculinity and femininity incongruence on consumer-brand boundary—the perceived psychological distance between a consumer and a brand—and affective cognitive and relational brand responses.

Self-Other Differences in the Perceived Efficacy of Self-Enhancement Products

Kaiyang Wu, University of Wisconsin - Madison, USA Evan Polman, University of Wisconsin - Madison, USA

We tested if judgments of self-enhancement products’ efficacy differ according to whether a product user is the self or someone else. In two pre-registered studies we found that people believe that self-enhancement products will be more efficacious for others than for themselves and that usership moderates this self-other difference.

Maximizing What? The Effect of Maximizing Mindset on the Evaluation of Product Bundles

Lan Xia, Bentley University, USA Nada Nasr Bechwati, Bentley University, USA

Do maximizers value product bundles higher than satisficers? We propose that it depends on whether maximizers try to maximize the value of the individual products or that of the entire bundle. The focus of the maximization is influenced by the presence of a discount. Results from three studies are presented.

Robots Don’t Discriminate: Financial Resources, Age, and Preference of AI in Service

Lan Xia, Bentley University, USA

Do consumers prefer robots or humans in service contexts? Our studies show that controlling for self-efficacy and AI performance preferences for AI vary by customer characteristics such that older consumers of lower (vs. higher) financial status tend to prefer robots more due to the expected service they obtain.

Penny Wise and Pound Foolish- The Interaction Between Ease of Comparison and Cognitive Effort And Its’ Impact On Preference And Choice Change

Na Xiao, Laurentian University, Canada

It is proposed that the interaction between ease of comparison and cognitive effort determines the importance of attributes. It is demonstrated that when cognitive effort to process information about choice-options is high (vs. low) consumers will give weight to ease-to-compare attributes in comparison therefore preference and choice change. Advances in Consumer Research (Volume 48) / 1237 The Effect of Handwritten Typefaces on Consumer Evaluation

Chunqu Xiao, Nanjing University, China Xin Wang, Nanjing University, China Hejin Fang, Nanjing University, China Haoyuan Wang, New York Statement University at Buffalo, USA Hong Zhu, Nanjing University, China

Handwritten typefaces strengthen the impact of online reviews. Across three studies the current research found that handwritten typefaces led higher evaluation on the perceived sincerity of reviewers thus benefitted positively reviewed products. This effect diminished when the reviews were negative or consumers had greater domain knowledge.

How Can I Help When I Feel Rejected? The Interplay of Social Exclusion, Social Identity, and Prosocial Behavior

Xuan Xie, University of California Irvine, USA Loraine Lau-Gesk, University of California Irvine, USA Amber Gaffney, Humboldt State University, USA

This research explores prosocial behavior through social exclusion and social identity lens. Results suggested that when feeling rejected prototypical members displayed lower willingness-to-help whereas peripheral members have higher willingness-to-help. When feeling ignored such effect was attenuated for both peripherals and prototypicals. Prosocial behavior alleviates some negative consequences of social exclusion.

Let’s Get Serious About Play: A Hedonic Vs . Eudaimonic Investigation of Goal-Based Mixed Affective Experiences

Xuan Xie, University of California Irvine, USA Sayantani Mukherjee, Central Washington University, USA Loraine Lau-Gesk, University of California Irvine, USA

This research investigates factors of goal-based mixed play experiences. Perceived effort corresponds positively to consumer judgment and perceived fun of the mixed (vs. pure positive) play experiences. Outcome-oriented thinking encourages a eudaimonic lens unaltered by cognitive load while process-oriented thinking shifts to a hedonic lens moderated by cognitive load.

Performance Diversity and Team Performance Expectations

Vivian (Jieru) Xie, Virginia Tech, USA Rajesh Bagchi, Virginia Tech, USA

We investigate the effects of performance diversity on expectations of team performance. We find that perceivers expect less diverse teams to perform better. We demonstrate the underlying process and delineate one approach via which these expectations could be reversed.

The Impact of Corporate Environmental Transgressions on Consumer Negative Reactions: The Role of Political Ideology

Chunyan Xie, Western Norway University of Applied Sciences, Norway Richard P. Bagozzi, University of Michigan, USA

We experimentally test a psychological mechanism underlying consumer negative reactions toward the firm as a function of perception of corporate environmental transgressions. Results showed that social disgust and attitude mediates the relationship between perceived corporate environmental transgressions and consumer reactions. Further political ideology moderates effects of transgressions on social disgust. 1238 / Working Papers The Impact of Financial Perception on Agent Preference of Intelligent Assistant

Chunya Xie, Renmin University of China, China En-Chung Chang, Renmin University of China, China

Four experiments indicate that financial affluent reduces participants’ preference for AI provider of intelligent assistant. Financial affluent leads to a high sense of entitlement which reduces AI preference. Assistant type moderates the effect. When the intelligent assistant is not instructional the effect of financial perception on agent preference reverses.

Why People Are Crazy For Chewing? An Exploratory Study of the Effect of Chewing on Mental Construal

jialiang xu, University of Manitoba, Canada Fang Wan, University of Manitoba, Canada Norbert Schwarz, University of Southern California, USA Wenqing Wu, University of Manitoba, Canada

Oral-sensory information different from other sensory information can be received by different ways (chewing/ swallowing/ drinking). This perspective has not studied by previous research. Our work examines that chewing (vs swallowing) activates people’s concrete mental construal which may help consumers focus on the current moment and appreciate experience of eating.

Need to Evaluate as a Predictor of Sharing and Seeking Online Recommendations

Mengran Xu, Ohio State University, USA Rebecca Walker Reczek, Ohio State University, USA Richard Petty, Ohio State University, USA

In this research we characterize who is likely to engage in sharing and seeking online word of mouth as a function of their need to evaluate an individual difference measure consisting of two motives the need to express attitudes and the need to acquire information in order to have attitudes.

Swear by My Family: Explore Consumer Attitudes Toward Brands Named After People

Fei (Katie) Xu, Hong Kong Polytechic University, China Yuwei Jiang, Hong Kong Polytechic University, China

Consumers are more interested to purchase from brands which are named after people’s last names than from brands containing people’s first names. This effect is found to be mediated by brand trust and moderated by consumers’ domain-specific knowledge.

No-Modification Allowed! Saying No to a Consumer’s Modification Request Can Increase Product Evaluation

Amanda Pruski Yamim, Grenoble Ecole de Management, France Adilson Borges, NEOMA Business School, France

Through three studies we show that the presence of non-modification policies can increase tastiness evaluation and willingness to pay because they make consumers think the chef put more effort into the food creation. The positive effect of non-compliance is contingent on consumers’ initial beliefs that the chef has high expertise. Advances in Consumer Research (Volume 48) / 1239 Before or After? The Order Effect of Donors’ Personally Identifiable Information on Donation Amount

Li Yan, University of Technology Sydney, Australia Xue Wang, University of Hong Kong, China Minjung Koo, Sungkyunkwan University, South Korea Hean Tat Keh, Monash University, Australia

Across five studies we found that individuals donate a higher amount if their personally identifiable information is requested before they indicate the donation amount. This finding makes important theoretical contributions and offer practical implications for charitable organizations by identifying a simple “nudge” that encourages donors to give higher amounts.

Why Should I Improve Myself? Perceived Economic Mobility and Consumer’s Interest in Self-Improvement Products

Chun-Ming Yang, Ming Chuan University, Taiwan

Across three studies we provide convergent evidence suggesting that the constraints perceived by consumers increased when they had a low perceived economic mobility (PEM) which in turn influenced their interests in self-improvement products. This study contributes to the literature of PEM by extending its downstream consequences.

Contestations of Entrepreneurship and Marketing in Creative Maker Tribes

Amy Yau, Cardiff University, UK

The ethnographic study explores the narratives of becoming and being a creative maker entrepreneur and how the makers imbue themselves as brands to start and develop their business. It explores how makers create offline and online commercial communities and tribes with other makers and consumers in the mass-market dominated marketplace.

Serif For the Conservatives: The Effects of Typeface on The Persuasiveness of Political Campaign

Jiwon Yi, Hongik University, South Korea Nara Youn, Hongik University, South Korea

This research investigated the effect of congruence between serif (vs. sans-serif) typeface and the conservative (vs. liberal) political ideology on the campaign effectiveness. Through two experiments we provided a support for the positive effect of serif typefaces on political campaign evaluation of the conservatives and unveiled underlying psychological process.

Immediate Incentives in Prosocial Contexts

Y. Rin Yoon, Cornell University, USA Kaitlin Woolley, Cornell University, USA

Prior research provides mixed evidence for the effect of financial incentives on prosocial behavior. Three studies (N = 982) fill this gap. We find that immediate (vs. delayed) incentives increase prosocial behavior when prosocial motivation is low but that this effect attenuates when prosocial motivation is high.

Ethnic Identity and Patience in Intertemporal Monetary Decisions

Jiaqi Yu, University of Chicago, USA Oleg Urminsky, University of Chicago, USA Stephanie Chen, London Business School, UK

Ethnic identity importance predicts consumers’ patience in making intertemporal monetary decisions. Two studies based on evidence from Asians and Caucasians in the U.S. and the U.K. conclude that people who consider their ethnic identity more important value the present over the future and make more impatient time discounting choices. 1240 / Working Papers Multiple Social Identities Increase Torelence of Unethical Behaviors

Irina Y. Yu, Chinese University of Hong Kong, China Li Huang, Hofstra University, USA Morgan X. Yang, The Hang Seng University of Hong Kong, China

Consumers possess various social identities and generally show an identity congurent behavior when an identity is momentarily salient. What will happen when multiple identities are salient? We find that the activation of multiple social identities enhances consumers’ moral relativism belief and increases their tolerance of unethical/immoral behaviors.

Refunded Money is More Likely to be Spent

Tianjiao Yu, Washington University in St Louis, USA Cynthia Cryder, Washington University, USA Robyn LeBoeuf, Washington University, USA

We test how consumers’ willingness to spend depends on the transaction history of the money they have available. Specifically we find that money refunded from previous purchases is more likely to be spent than non-refunded money. This research extends prior findings about mental accounting and fungibility.

Sustainability as the Guilt-Relief Pill For Luxury Purchases

Shubin Lance Yu, Peking University, China Nabanita Talukdar, Hult International Business School, USA Sangeeta Trott, ITM-SIA Business School, India

Sustainability claims are not widely used by luxury brands because sustainability is considered incompatible with luxury. The results of two experimental studies suggest that for luxury products with a high price a sustainability claim can reduce the anticipated guilt of purchases which in turn leads to higher purchase intention.

Is First the Worst? Understanding Consumers’ Tendency to Postpone the Better Outcome

Ozge Yucel-Aybat, Pennsylvania State University Harrisburg, USA Marina Carnevale, Fordham University, USA

This research investigates the tendency of consumers to postpone their preferred item to the end. Two studies demonstrate that long-term (vs. short-term) oriented consumers are more likely to save their favorite product for last because they value improvements over time more.

Boosting the Resale Value: Effects of Expired Warranty and Repair Conditions of Secondhand Products

Ulku Yuksel, University of Sydney, Australia Nguyen T Thai, University of Wollongong, Australia Vince Mitchell, University of Sydney, Australia

Secondhand retailers rarely communicate negative information such as manufacturer warranties being expired or secondhand products being repaired. We show that consumers associate secondhand products with expired warranties with higher (lower) resale values than those with no warranties when these products have been repaired (never been repaired) and explain the process. Advances in Consumer Research (Volume 48) / 1241 Repetition Or No Repetition? Effects Of Repetitive Numbers On Bundling Promotion

Jintao Zhang, Drexel University, USA Hongjun Ye, Drexel University, USA Siddharth Bhatt, Pennsylvania State University, Harrisburg, USA Haeyoung Jeong, Drexel University, USA Rajneesh Suri, Drexel University, USA

Consumers are constantly making decisions on numbers along with their journey of consumption. Our study examines the effect of repetitive numbers on bundling price (i.e. 2 for $2.22 vs. 2 for $$2.12) that consumers have higher intentions to purchase bundling products with repetitive number prices.

Save Time For Better Life! The Effect Of Time Saving On Quality Perception

Consumers form their perceptions of quality with different signals. This study examines a novel signal of quality – time – as consumers would associate time saved when shopping with the perceptions of quality of the products the more time consumers save the better quality of the products they perceive.

Food Label Halos and Consumer Perceptions of Food Healthfulness

Yijun Zhao, Renmin University of China, China

This paper demonstrates that label halos bias dieters’ estimates about the healthfulness of vice foods containing such labels. Dieters tend to draw a directional conclusion about food healthfulness to mediate the guilt associated with consuming indulgent food. Providing corrective information to dieters can mitigate these effects.

The Impact of Social Crowding on Cross Sensory Products—— Based on the Brand Alliance Perspective

Yijun Zhao, Renmin University of China, China

This research defines the cross sensory product (CSP) which refers to the product with different sensory attributes provided by two or more co-brands in brand alliance and suggest that feeling crowded makes consumers need more arousal which in turn makes them prefer cross sensory product.

Time, Endowment Effect, and Happiness

Wenxue Zheng, Cornell University, USA Ashley V. Whillans, Harvard Business School, USA

This project demonstrates an endowment effect of time such that people expect higher compensation for time losses than for identical time gains. We also find initial evidence for a link between this endowment effect and happiness.

Where Does My Time Go?! The Trap of Segregated Resources

Wenxue Zheng, Cornell University, USA David R. Just, Cornell University, USA Jura Liaukonyte, Cornell University, USA

This project showed that time follows reversed arithmetic rules of mental accounting. Through hypothetical questionnaires that measured WTP WTA willingness to wait level of irritation we find that aggregated (vs. segregated) time was preferred and participants have a higher WTA and WTP for the aggregated time losses and time gains. 1242 / Working Papers How Crowdedness Enhances Consumers’ Healthy Lifestyle

Wanyi Zheng, University of Hong Kong, China He (Michael) Jia, University of Hong Kong, China Echo Wen Wan, University of Hong Kong, China

Social crowdedness is ubiquitous and has important implications for marketers. This research demonstrates that in a crowded environment consumers are more likely to engage in health-improving consumption which can boost their immune system and thus make themselves less vulnerable to contagious diseases that become more salient under crowdedness.

A Helping Hand: How Online and In-store Shopping Environments Change Information Search For Haptic Products

Lingrui Zhou, Duke University, USA Nicolette Sullivan, London School of Economics, UK Gavan Fitzsimons, Duke University, USA

What information do consumers search for when shopping for haptic products? The current research uses both behavioral data and computer mouse tracking analysis to examine how consumers search for and weigh various attributes across different shopping environments (online in-store or a combination of both).

AI is Better When I’m Sure: The Influence of Certainty Level of Consumer Needs on Their Willingness to Use AI Customer Service

Yimin Zhu, Sun Yat-sen University, China Yingyue Liu, Sun Yat-sen University, China Xiang Fang, Oklahoma State University, USA

This paper explores the effect of certainty level of consumer needs on their willingness to use AI customer service as well as its moderating effect and mediating effect. The findings make important theoretical contributions to AI research in the service context and provide useful guidelines for companies to use AI.

Device, Fast and Slow: How Devices Influence Consumer Decisions

Shuqi Zhu, University of Warwick, UK Sarah Wei, University of Warwick, UK John M. Rudd, University of Warwick, UK Yansong Hu, University of Warwick, UK

How the devices (mobile versus stationary) consumers use influence how they approach a purchase decision? We argue that consumers are less likely to employ deliberative processing when they use mobile devices compared to when they use stationary devices. Evidence from three studies supports this theorizing.

Can Handwritten-appearing Typefaces Boost the Effectiveness of Social Norms? An Experimental Investigation on Sustainable Consumer Behavior

Vita Eva Maria Zimmermann-Janssen, Heinrich Heine University Düsseldorf, Germany Carina Sophia Hütte, Heinrich Heine University Düsseldorf, Germany Julia Kluger, Heinrich Heine University Düsseldorf, Germany

Descriptive normative information (DNI) about a majority’s sustainable behavior proved to be especially persuasive when others are present. As handwriting can evoke human presence perceptions we tested experimentally whether DNI’s influence on sustainable behavior can be increased if presented in a handwritten-appearing (vs. machine-written) typeface under factual social absence. Advances in Consumer Research (Volume 48) / 1243 The More, the Worse – When the Communication of High Majority Sizes Decreases the Impact of Descriptive Normative Information on Consumers’ Willingness to Donate

Vita Eva Maria Zimmermann-Janssen, Heinrich Heine University Düsseldorf, Germany

The ability of descriptive normative information (DNI) about majorities’ charitable behavior to motivate donations is currently discussed in reaction to diverging findings. Drawing on altruism models and impact philanthropy we propose to reconsider crowding-out predictions and provide experimental evidence for DNI’s diminishing persuasiveness at high (vs. low) majority sizes communicated.

Time of Day and Construal Level Interact to Influence Engagement With Information

Ozum Zor, Rutgers University, USA Kihyun Hannah Kim, Rutgers University, USA Ashwani Monga, Rutgers University, USA

Social media messages that match with a high-level construal engage consumers more in the morning whereas those that match a low-level construal engage consumers more in the evening. Empirical evidence comes from an experiment and a Twitter dataset. These findings have implications for both theory and practice. 1244 / Working Papers Film Festival 2020

“Coworkers Lifestyle” Máté Bencze, Corvinus University of Budapest, Hungary Outside In: When Social Dynamics Infiltrate the Extraordinary Experience of the Fair Baptiste Cleret, University of Rouen, France Boris Collet, University of Rouen Normandy, France Cerise Thorel, University of Rouen Normandy, France Stanislavskij’s Reflexive Videography to Explore the Representational World of the Researcher Michele Corengia, Università della Svizzera italiana, Switzerland Airbnb Butterfly Effect: How Consumption Transformations Reverberate on Social and Cityscapes Marlon Dalmoro, University of Vale do Taquari, Brazil Diego Costa Pinto, Nova IMS - Information Management School, Portugal Celso Zanini, Nova IMS - Information Management School, Portugal The Disruptive ‘Other’? Exploring Human-Animal Relations in Tourism Minni Haanpää, University of Lapland, Finland Tarja Salmela, University of Lapland, Finland José-Carlos García-Rosell, University of Lapland, Finland Mikko Äijälä, University of Lapland, Finland Too Big To Care: Uncovering the Ugly Truth of Brand Relationships in the Video Game Industry Jacob Lee Hiler, Ohio University, USA Elana Harnish, Ohio University, USA Pac-Man & the Pack Mentality: Nostalgia & Socialization Factors in Gaming Choice Avery Kerns, Ohio University, USA Jacob Lee Hiler, Ohio University, USA Impact of Religious Differences on Marketplace Configurations Ameya Mittal, Indian Institute of Technology Bombay, India Akshaya Vijayalakshmi, Indian Institute of Management Ahmedabad, India Nitisha Tomar, University of Wisconsin - Madison, USA Ankur Kapoor, Indian Institute of Management Udaipur, India Sustainup Dr. Iris Mohr, St. John’s University, USA Dignity as a Gift: an Immersion in the Gastronomic Experience at Refettorio Paris Ophelie Elise Mugel, Ferrandi Paris, France Winter Texans’ Motivations to Temporarily Relocate: Exploring Costs of Living Ross Murray, The University of Texas Rio Grande Valley, USA Suwakitti Amornpan, The University of Texas Rio Grande Valley, USA Michael Minor, The University of Texas Rio Grande Valley, USA

Advances in Consumer Research 1245 Volume 48, ©2020 1246 / Film Festival Lost in an Emodity: Self-Reflexive Ethnographer Portraits Joonas Rokka, Emlyon Business School, France Maira Lopes, Emlyon Business School, France Vincent Dewaguet, Emlyon Business School, France Anissa Pomies, Emlyon Business School, France Lydia Ottlewski, University of St. Gallen, Switzerland Come Come ! Thomas Stenger, University of Poitiers, France Olivier Coussi, University of Poitiers, France Knowledge Forums

Forum Where Consumer Behavior Meets Language: Applying Linguistic Methods to Consumer Research

Chairs: Ann Kronrod, University of Massachusetts Lowell, USA Grant Packard, York University, Canada Sarah G Moore, University of Alberta, Canada Jonah Berger, University of Pennsylvania, USA

Participants: Jeffrey Inman, University of Pittsburgh, USA Robert Meyer, University of Pennsylvania, USA L. J. Shrum, HEC Paris, France Ashlee Humphreys, Northwestern University - Medill, USA Nicholas Lurie, University of Connecticut, USA Andrea Webb Luangrath, University of Iowa, USA Jeff Lee, American University, USA Matthew D Rocklage, University of Massachusetts, USA Charles Zhang, University of California Riverside, USA Ruth Pogacar, University of Calgary, Canada Ivan Gordeliy, Georgetown University, USA Shiri Melumad, University of Pennsylvania, USA Amir Grinstein, Northeastern University, USA Francisco Villarroal Ordenes, University of Massachusetts, USA Guang-Xin Xie, University of Massachusetts Boston, USA Zoey Chen, University of Miami, USA

What can we discover by applying linguistic methods to consumer research? This Knowledge Forum introduces and discusses the multifaceted opportunities of applying linguistic research methods to explore language in the marketplace and employs a hands-on workshop approach to sprout novel research ideas that use multiple methods of linguistic inquiry.

Forum Forum on the Research and Interpersonal Culture of ACR

Chairs: Tiffany White, University of Illinois at Urbana-Champaign, USA

Presenters: Joel Huber, Duke University, USA Darren Dahl, University of British Columbia, Canada Anirban Mukhopadhyay, Hong Kong University of Science and Technology, China Laura Peracchio, University of Wisconsin - Milwaukee, USA Sharon Shavitt, University of Illinois at Urbana-Champaign, USA

The goal of the forum is to explore the factors that have made ACR culturally successful but ultimately vulnerable. Tiffany White will moderate six scholars and an active audience to discuss ACR’s unique research culture and the impact on that culture of publication norms successful collaborations global heterogeneity and the demands from university promotion systems.

Chairs: Gizem Yalcin, Erasmus University Rotterdam, The Netherlands William Fritz, Oxford University, UK

Discussants: Ana Valenzuela, Baruch College, USA Andrew T. Stephen, Oxford University, UK Bernd Schmitt, Columbia University, USA Christian Hildebrand, University of Geneva, Switzerland Donna L. Hoffman, George Washington University, USA Klaus Wertenbroch, INSEAD, Singapore Markus Giesler, York University, Canada Rebecca Walker Reczek, Ohio State University, USA Rhonda Hadi, Oxford University, UK Robert Meyer, University of Pennsylvania, USA Russell W. Belk, York University, Canada Stefano Puntoni, Erasmus University Rotterdam, The Netherlands Szu-chi Huang, Stanford University, USA Thomas Novak, George Washington University, USA

This roundtable brings together a diverse group of senior academics to offer guidance to early-career researchers who are interested in studying technology in consumer research. Discussants will talk about theoretical methodological empirical and substantive (e.g. publication process career management) challenges in this research area.

Forum Political “Consumption”: Motivations, Decision-Making and Implications

Chairs: Anastasiya Pocheptsova Ghosh, University of Arizona, USA Jennifer Savary, University of Arizona, USA

Discussants: Adam Duhachek, University of Illinois at Chicago, USA Cait Lamberton, University of Pennsylvania, USA Clayton R Critcher, University of California Berkeley, USA David Crockett, University of South Carolina, USA Erick M. Mas, Vanderbilt University, USA Gita Johar, Columbia University, USA Melanie Wallendorf, University of Arizona, USA Monika Lisjak, Arizona State University, USA Nailya Ordabayeva, Boston College, USA Neeru Paharia, Georgetown University, USA Nooshin Warren, University of Arizona, USA Ronald Paul Hill, American University, USA Sharon Shavitt, University of Illinois at Urbana-Champaign, USA Uzma Khan, University of Miami, USA

Consumer engagement in politics can be direct (e.g. political donations campaign merchandise purchases) or indirect (e.g. politicized consumption policy advocacy). This forum brings together thought leaders from CP CCT and Strategy to identify these behaviors integrate extant knowledge and envision a path forward for this promising but underexplored research domain. Advances in Consumer Research (Volume 48) / 1249 Forum Creating Boundary-Breaking, Marketing-Relevant Consumer Research

Chair: Simona Botti, London Business School, UK

Participants: Fleura Bardhi, City University of London, UK Eileen Fischer, York University, Canada Donna L. Hoffman, George Washington University, USA Robert Kozinets, University of Southern California, USA Donald Lehmann, Columbia University, USA John Lynch, University of Colorado, USA Deborah MacInnis, University of Southern California, USA Christine Moorman, Duke University, USA Page Moreau, University of Wisconsin - Madison, USA Vicki G. Morwitz, Columbia University, USA Connie Pechmann, University of California Irvine, USA Oded Netzer, Columbia University, USA

Consumer research often fails to have broad impact. We propose a framework that identifies potential reasons (implicit boundaries) and solutions (boundary-breaking opportunities); we illustrate this framework with an example; we discuss additional reasons solutions and examples in small groups moderated by the organizers; we summarize these discussions in plenary.

Forum Consumer Health and Medical Decision Making: 2020 and Beyond

Chairs: Dipankar Chakravarti, Virginia Tech, USA Haiyang Yang, Johns Hopkins University, USA Meng Zhu, Johns Hopkins University, USA

Presenters: Angela Y. Lee, Northwestern University, USA Anirban Mukhopadhyay, Hong Kong University of Science and Technology, China Brent McFerran, Simon Fraser University, Canada Carey K. Morewedge, Boston University, USA Dawn Iacobucci, Vanderbilt University, USA Jeff Inman, University of Pittsburgh, USA Leonard Lee, National University of Singapore, Singapore Peggy Liu, University of Pittsburgh, USA Pierre Chandon, INSEAD, France Priya Raghubir, New York University, USA Punam Keller, Dartmouth College, USA Ronald Paul Hill, American University, USA Stacy Wood, North Carolina State University, USA Vikas Mittal, Rice University, USA

The global health landscape has been changing rapidly. This has created the need for a new integrated perspective on consumer-relevant health topics through the lens of psychology marketing and economics. This roundtable will discuss emerging topics methodological approaches and future directions on consumer health and medical decision-making. 1250 / Forum Summaries Forum Shedding Light on the Invisibles: Extending Consumer Theories, Methods, and Insights to Include Financially Vulnerable Consumers

Chairs: Gergana Y. Nenkov, Boston College, USA Linda Court Salisbury, Boston College, USA

Participants: Simon Blanchard, Georgetown University, USA Fred Feinberg, University of Michigan, USA Ronald Paul Hill, American University, USA Nina Mazar, Boston University, USA Martin Mende, Florida State University, USA Nailya Ordabayeva, Boston College, USA Julie L. Ozanne, University of Melbourne, Australia Maura Scott, Florida State University, USA Abigail Sussman, University of Chicago, USA Remi Trudel, Boston University, USA Nancy Wong, University of Wisconsin - Madison, USA Min Zhao, Boston College, USA

This interactive knowledge forum focuses on financially vulnerable consumer segments frequently excluded from financial decision- making research (e.g. low-income consumers unbanked or underbanked credit invisibles). Participants will discuss how consumer researchers can enhance theories models and participant samples to be more inclusive of financially vulnerable consumers and improve their financial well-being.

Forum Bringing Consumer Culture Theoretics to Policy Debate

Chairs: Eric Arnould, Aalto University, Finland Søren Tollestrup Askegaard, University of Southern Denmark, Denmark Guliz Ger, Bilkent University, Turkey

This roundtable aims at providing a platform for discussing strategies for knowledge dissemination pertaining to consumer research in particular consumer culture theoretics and transformative consumer research for better policy-making concerning governance of consumption-related issues in relation to social and environmental sustainability.

Forum M-Turk is dying . Don’t die with it . An interactive workshop on collecting data from more difficult and interesting sources.

Chair: Broderick Lee Turner, Virginia Tech, USA

Should we all stop collecting data on Amazon’s Mechanical Turk? It is rife with problems including non-naivety bot-panics fraudulent responses and bored workers. This knowledge forum provides some fun exercises to help researchers old and new wean themselves from their M-Turk habit and consider some alternatives data collection methods. Advances in Consumer Research (Volume 48) / 1251 Forum Shining a Light on Maladaptive Consumption: The Journey so Far and the Roads Less Traveled

Chairs: Meryl P. Gardner, University of Delaware, USA Caroline Roux, Concordia University, Canada Shailendra Pratap Jain, University of Washington, USA

Participants: Paul M Connell, Stony Brook University, USA Brennan Davis, California Polytechnic State University, USA Benét DeBerry-Spence, University of Illinois at Chicago, USA Pamela Scholder Ellen, Georgia State University, USA Paula Fitzgerald, West Virginia University, USA Marcia Flicker, Fordham University, USA Michael Kamins, Claremont Graduate University, USA Peggy Liu, University of Pittsburgh, USA Ingrid Martin, California State University Long Beach, USA Marlys Mason, Oklahoma State University, USA Hieu Nguyen, California State University Long Beach, USA Carly Pacanowski, University of Delaware, USA Janis Pappalardo, USA Federal Trade Commission, USA Martin Reimann, University of Arizona, USA Ainslie Schultz, Providence College, USA Christine Skubisz, University of Delaware, USA Jerome Williams, Rutgers University, USA

This interactive session will bring together researchers representing various paradigms methods and approaches to studying maladaptive consumption and ACR attendees interested in investigating the topic. We will assess the main questions studied in the literature in order to identify issues that have received less attention and develop cross-paradigmatic collaborations.

Forum Field Experiments: A Practical Tutorial

Chairs: Rajesh Chandy, London Business School, UK Stephen Anderson, Stanford University, USA Praveen Kumar Kopalle, Dartmouth College, USA

This tutorial will address practical aspects of planning and executing field experiments. We will use examples of successful and failed research projects to illustrate a toolkit for field experiments. We will then engage in a facilitated discussion of ways to apply the toolkit to participants’ research questions. 1252 / Forum Summaries Author Index

Auschaitrakul, Sumitra ...... 1182, 1204 Bhatt, Siddharth ...... 1241, 1241 A Ayala, Edgardo ...... 256 Bhattacharjee, Amit ...... 902, 902 Aaker, Jennifer ...... 1221 Aydinli, Aylin ...... 422 Bhoumik, Kshitij ...... 1185 Abe, Makoto ...... 11 Aydinoglu, Nilufer Z...... 266 Bido, Diógenes ...... 1227 Abi, Gulen Sarial ...... 749 Bigne, Enrique ...... 1201 Abilova, Almira ...... 1181 Bilgin, Baler ...... 216 Aboelenien, Aya . . . . 177, 179, 1181, 1181 B Biller, Miriam ...... 218 Abreu, Luis ...... 181 Babin, Barry J...... 469 Biraglia, Alessandro . . . . 1048, 1182, 1195 Acar, Oguz A...... 183 Badot, Olivier ...... 1027 Bischof, Severin Friedrich . . . . . 220, 648 Achar, Chethana ...... 936 Bae, Hankyul ...... 1182 Biswas, Dipayan ...... 1185 Adaval, Rashmi ...... 1216 Baek, Elisa C...... 891 Bitterly, T. Bradford ...... 930 Adler, Susanne ...... 186 Baek, Eunsoo ...... 1032, 1183 Bittner, Samantha ...... 222 Affinito, Sal ...... 968 Bagchi, Rajesh . . . 486, 576, 661, 788, 1237 Blair, Sean ...... 913, 1000 Agarwal, James ...... 1191 Bagga, Charan K ...... 202 Blanchard, Simon ...... 1250 Agrawal, Nidhi ...... 269, 698, 936 Bagozzi, Richard P...... 1237 Bodur, H. Onur ...... 1194 Aguirre-Rodriguez, Alexandra . . . . . 1230 Bakpayev, Marat ...... 836, 836 Boegershausen, Johannes . . . . .1052, 1086 Äijälä, Mikko ...... 1245 Ballantyne, Anne Gammelgaard . . . . . 134 Boettger, Tim M...... 220, 461 Airoldi, Massimo ...... 635 Banerjee, Pramit ...... 432 Bogard, Jon ...... 943 Aka, Ada ...... 825, 1139 Barasch, Alixandra ...... 902, 994, 1223 Bogard, Jonathan E...... 223 Akaka, Melissa Archpru . . . . . 1069, 1183 Bardhi, Fleura ...... 1249 Boistel, Philippe ...... 488 Akamatsu, Naoki ...... 1181 Barnea, Uri ...... 1091 Boller, Daniel ...... 1185 Akdevelioglu, Duygu ...... 874 Barnes, Aaron J...... 203, 1183 Bolton, Lisa E...... 792 Akponah, Precious O...... 16 Barnhart, Michelle ...... 1106 Bone, Sterling ...... 324 Albuquerque, Paulo ...... 891 Barros, Lucia S. G...... 207 Bonezzi, Andrea ...... 408, 1011, 1165 Algesheimer, René . . . . . 250, 1186, 1235 Bartsch, Fabian ...... 1226 Bonsu, Samuel K...... 227 Alhanouti, Roua ...... 1181 Basil, Debra Z ...... 1189 Borges, Adilson ...... 1238 Alinaghian, Leila ...... 608 Basílio, Roberta Gabriela ...... 205 Botelho, Delane . .207, 1183, 1185, 1210, 1210, Alman, Sophie ...... 949 Baskentli, Sara ...... 1183 1222, 1227 Alves, Catia Carvalheiro ...... 1182 Batista, Juliana M...... 207, 1183 Botti, Simona ...... 1249 Alves, Hans ...... 335 Baudet, Alex Paul ...... 177 Bourgeon-Renault, Dominique . . . . . 1220 Amine, Abdelmajid ...... 81 Baumgartner, Hans ...... 374, 381 Bouwer, Anna ...... 1185 Amir, Ofra ...... 1195 Baxter, Weston ...... 778 Boyd, Eric ...... 293 Amir, On ...... 1155 Bechwati, Nada Nasr ...... 1236 Bracher, Martina ...... 1223 Amornpan, Suwakitti ...... 1245 Becker, Joshua ...... 842 Bradford, Tonya ...... 1069 Anderson, Kelley Cours ...... 188 Becker, Larissa Carine Braz ...... 1183 Bradford, Tonya Williams ...... 1118 Anderson, Stephen ...... 1251 Beisecker, Christopher ...... 402 Brakus, J. Josko ...... 1182 Ando, Kazuyo ...... 1230 Bekk, Magdalena ...... 406 Brakus, J. Joṣko ...... 1048 Andrade, Eduardo B...... 606, 760 Belaid, Ons ...... 1184 Brand, Matthias ...... 1224 Andriuzzi, Andria ...... 190 Belk, Russ ...... 639 Brei, Vinicius Andrade ...... 580 Angle, Justin ...... 954 Belk, Russell ...... 836, 987 Brendl, Miguel ...... 535 Anin, Camille ...... 21, 192 Belk, Russell W...... 173, 209, 506, 1248 Brick, Danielle ...... 847 Anlamlier, Eda ...... 194 Benartzi, Shlomo ...... 1064, 1101 Brick, Danielle J...... 230 Aquino, Karl ...... 101 Bencze, Máté ...... 1245 Briers, Barbara ...... 851 Aragón, Oriana Rachel ...... 1182, 1220 Bendle, Neil ...... 202, 825 Brito, Eliane Zamith ...... 331 Arechar, Antonio A...... 863 Bennett, Delancy ...... 212 Broniarczyk, Susan ...... 1222 Arens, Zachary ...... 434 Bentsen, Kristin ...... 1184 Bruckdorfer, Raphaela Elisabeth . . . . 1186 Argenbright, Keith ...... 1129 Berg, Hanna ...... 1184 Bruggmann, Anja ...... 1223 Argo, Jennifer ...... 1204 Berger, Jonah ...... 930, 981, 1247 Brunk, Katja H...... 47 Ariely, Dan ...... 1194 Berger, Katja ...... 1184 Bruno, Hernán A...... 616 Arnould, Eric ...... 1250 Bergner, Anouk ...... 214 Bryksina, Olya ...... 232, 234 Arnould, Eric J ...... 196 Berman, Jonathan Z...... 1096 Brylla, Daniel Erik ...... 1186 Arora, Sunny ...... 1182 Berman, Jonathan Zev ...... 568, 958 Bublitz, Andrea ...... 1186 Arsel, Zeynep ...... 179, 874 Bermes, Alena ...... 1184, 1185 Bucher, Jan-Hendrik ...... 31 Ashby, Rhys ...... 198 Bernard, Anna ...... 443 Budescu, David ...... 842 Ashkenazi, Amit ...... 1155 Berry, Christopher ...... 1196 Buechel, Eva C...... 236 Asif, Esra ...... 1182 Berry, Emily ...... 1129 Buechner, Bryan ...... 238 Askanius, Tina ...... 753 Beruchashvili, Mariam ...... 1184, 1214 Buechner, Bryan M...... 936 Askegaard, Søren Tollestrup ...... 1250 Besharat, Ali ...... 1160 Bulusu, Padma ...... 314 Atalay, A. Selin ...... 831 Beshears, John ...... 1101 Burghartz, Pia ...... 857 Atanasova, Aleksandrina ...... 200 Bettany, Shona ...... 1221 Burson, Katherine ...... 299 Atasoy, Ozgun ...... 535 Beverland, Michael ...... 368 Busquet, Francesc ...... 831 Atik, Deniz ...... 262 Bhatia, Sudeep ...... 825, 825, 1139 Büttner, Ana Julia ...... 1186 Aufschnaiter, Christiane ...... 26 Büttner, Oliver B...... 1186, 1205, 1224

Byrne, Derek ...... 1220 Chen, Tianqi ...... 280 Dagalp, Ileyha ...... 47 Chen, Wei-Fen ...... 1188, 1188 Dahl, Darren ...... 183, 1006, 1247 Chen, Yifan ...... 1199 Dahl, Darren W...... 1165 C Chen, Yi-Jing ...... 271 Dai, Hengchen ...... 1101 Cabugao, Amanda ...... 1194 Chen, Yunqing ...... 278 Dai, Xianchi ...... 687 Cadario, Romain ...... 1124 Chen, Yu-Shan Athena ...... 1188 Dallolio, Adriana Schneider ...... 331 Cai, Fengyan ...... 1224 Chen, Zoey ...... 1247 Dalmoro, Marlon ...... 1245 Cailluet, Ludovic ...... 610 Cheng, Yin-Hui ...... 1189 Dalton, Amy ...... 1208 Cambier, Fanny ...... 240 Cheng, Yuan ...... 350 Dam, Christian ...... 1191 Campana, Mario ...... 244, 246 Chernev, Alexander ...... 913 Dang, Chu (Ivy) ...... 480 Campbell, Margaret C...... 248 Chernishenko, Jennifer ...... 1189 Danienta, Nadia ...... 295 Candrian, Cindy ...... 250 Chi, Yunjia ...... 1189 Danziger, Shai ...... 1022 Canniff, Miranda ...... 1200 Chien, Yi-Wen ...... 1189 Darke, Peter ...... 1202 Canniford, Robin ...... 1112, 1198, 1230 Chimenti, Paula ...... 66 Darriet, Camille ...... 1232 Cannon, Christopher ...... 252, 254 Cho, Byung Geun ...... 907 Das, Gopal ...... 1017, 1191 Cao, Edita ...... 722 Choi, Incheol ...... 1038 Daskalopoulou, Athanasia ...... 168 Capone, Aline Delmanto ...... 1186 Choi, Jungsil ...... 285 Davidson, Martin ...... 1200 Cappelli, Sonia ...... 1184 Chong, Chooi Ling ...... 38 Davis, Brennan ...... 1251 Cárdenas, Diana M...... 256 Choueiki, Ziad ...... 1189 de Bellis, Emanuel . .393, 651, 851, 857, 1052 Cardoso, Flavia ...... 301, 1186 Chow, Ai Ming ...... 177 De Bondt, Caroline ...... 225 Carnevale, Marina ...... 1240 Christandl, Fabian ...... 287 de Fortuny, Enric Junqué ...... 981 Carroll, William Ryall ...... 362 Christensen, Kate ...... 868 de Hooge, Ilona E...... 348, 720 Carvalho, Lilian ...... 1187 Christenson, Brett ...... 746 de Kerviler, Gwarlann ...... 449 Cascio, Christopher N...... 1231 Chrysochou, Polymeros ...... 1194 De La Rosa, Wendy ...... 1221 Caserotti, Marta ...... 1187 Chu, Charlene K...... 1038 De La Rosa, Wendy J...... 1064 Casey, Katherine ...... 1187 Chuang, Shih-Chieh ...... 1189 de lanauze, Gilles Séré ...... 1225 Casotti, Leticia M...... 126 Chun, Helen ...... 1032 de Luca, Ramona . . . . . 1185, 1210, 1210 Casotti, Leticia Moreira ...... 148 Chung, JungHan ...... 1190 de Maya, Salvador Ruiz ...... 857 Cassinger, Cecilia ...... 1187 Chung, Sorim ...... 289 de Rojas, Isadora Matsuda Sanchez . . . 1027 Castaño, Raquel ...... 256, 1080 Clair, Julian K Saint ...... 1231 de Ruyter, Ko ...... 322 Castelo, Noah ...... 1052, 1165 Clarkson, Joshua J...... 238, 936 De Temmerman, Joyce ...... 1229 Catapano, Rhia ...... 1086 Claus, Bart ...... 291 DeBerry-Spence, Benét . . . 1118, 1191, 1251 Cavusoglu, Lena ...... 262 Clegg, Melanie ...... 1052 DeCelles, Katy ...... 1170 Celik, Hazal ...... 1230 Cleret, Baptiste ...... 1245 Del Bucchia, Céline ...... 1106 Cerio, Eva ...... 264 Clergue, Valentina ...... 1190 Delannoy, Arnaud ...... 1230 Ceylan, Gizem ...... 863, 1017 Cocker, Hayley ...... 1220 Dellaert, Benedict ...... 1196 Ceylan, Melis ...... 266, 1198 Cole, Catherine ...... 1206 Dellande, Stephanie ...... 1200 Chae, Boyoun Grace ...... 1225 Collet, Boris ...... 1245 Demangeot, Catherine ...... 449 Chae, Rebecca ...... 1011 Collins, Hanne ...... 1017 Demirdag, Ipek ...... 297, 868 Chai, Shaowei ...... 1187 Connell, Paul M ...... 1251 Deng, Zhongzhun ...... 1191 Chaker, Fadwa ...... 227 Consiglio, Irene ...... 248, 1182 Derbaix, Maud ...... 1220 Chakravarti, Dipankar . . . . .918, 918, 1249 Conte, Leonardo ...... 1190 Deshpande, Sameer ...... 346 Chan, Cindy ...... 1134 Cordeiro, Rafaela Almeida ...... 205 Desmichel, Perrine ...... 1191 Chan, Elaine ...... 851 Corengia, Michele ...... 1190, 1245 Dewaguet, Vincent ...... 1246 Chan, Eugene ...... 1196, 1197 Cornelissen, Gert ...... 1218 Dewitte, Siegfried ...... 303 Chan, Raymond ...... 1220 Corus, Canan ...... 624 Diamantopoulos, Adamantios ...... 1192 Chandon, Pierre ...... 1124, 1249 Coslor, Erica ...... 1232 Dias, Rodrigo ...... 930 Chandy, Rajesh ...... 1251 Cossu, Martina ...... 1170 Dickinson, Sonia ...... 104 Chaney, Damien ...... 514 Costa-Migeon, Sandrine ...... 305, 316 DiCosola, Blake ...... 1124 Chang, Chia-Chi ...... 1187 Cotte, June ...... 370, 790 Diehl, Kristin ...... 1017 Chang, Chia-Han ...... 271 Coulter, Robin ...... 717 Dietvorst, Berkeley J...... 1091, 1155 Chang, Chun-Tuan ...... 271 Coulter, Robin A...... 1224 Dimofte, Claudiu V...... 190 Chang, En-Chung ...... 1238 Coussi, Olivier ...... 1246 Ding, Yu ...... 863 Chang, Sylvia S...... 269, 974 Cowan, Kirsten ...... 293 Djedidi, Amina ...... 52, 1192 Chang, Yaping ...... 1187, 1206 Critcher, Clayton R ...... 1248 do Vale, Rita Coelho ...... 443, 1231 Chartrand, Tanya ...... 1150, 1150 Crockett, David ...... 1248 Doering, Tim P...... 299 Chartrand, Tanya L...... 1086 Crolic, Cammy ...... 216 Dogra, Unnati ...... 1182 Chatzidakis, Andreas ...... 272 Cruz, Angela Gracia B...... 1186 Dolbec, Pierre-Yann ...... 874 Cheema, Amar ...... 1193 Cryder, Cynthia . . . . 1064, 1198, 1227, 1240 Dolifka, David ...... 1192 Chen, Charlene Y ...... 275 Cui, Nan ...... 790 Dommer, Sara Loughran ...... 963, 1006 Chen, Frances ...... 1017 Czellar, Sandor ...... 86, 90, 1190 Donkers, Bas ...... 1196 Chen, Jia ...... 799, 1191 Donnelly, Grant E...... 913, 968 Chen, Ning ...... 283 Donskoy, Raquel ...... 301 Chen, Rong ...... 1224 D Doré , Bruce P...... 891 Chen, Rui ...... 1209 D’Angelo, Jennifer ...... 1058 Drenten, Jenna ...... 874, 880 Chen, Shiyun ...... 1188 da Costa Tayt-son, Débora Bogéa . . . . .148 du Plessis, Christilene ...... 588 Chen, Siyun ...... 1188 Dach, Ofer Dekel ...... 1190 Du, Jessie ...... 1011 Chen, Stephanie ...... 1239 Dachs, Ofer Dekel ...... 43 Duan, Shen ...... 1192 Advances in Consumer Research (Volume 48) / 1255

Duani, Nofar ...... 902 Feldberg, Alexandra ...... 1170 Gaviria, Pilar Ximena Rojas ...... 1186 Dubois, David ...... 86, 588 Feng, Wenting ...... 1193 Ge, Lin ...... 438, 1194 Duclos, Rod ...... 673, 676 Ferguson, Graham ...... 104 Genevsky, Alexander ...... 891 Duffy, Katherine ...... 246, 683 Fernandes, Daniel ...... 328, 404 Ger, Guliz ...... 1250 Dugue, Tatianna ...... 1200 Fernbach, Philip M...... 1208 Gerhart, Jana-Verena ...... 1195 Duhachek, Adam ...... 936, 1248 Ferraz, Sofia Batista ...... 1212 Gerrath, Maximilian H. E. E...... 1195 Durante, Kristina ...... 1124 Ferreira, Daniela ...... 66 Gershoff, Andrew D...... 299 Durif, Fabien ...... 1196 Ferreira, Marcia Christina ...... 331, 333 Gershon, Rachel ...... 958, 1064, 1096 Durso, Geoffrey R. O...... 994 Ferreira, Maura ...... 1193 Geuens, Maggie . . . . 225, 1189, 1213, 1232 Dyen, Margot ...... 305, 1225 Festjens, Anouk ...... 398, 400 Geyskens, Kelly ...... 667 Figueiredo, Bernardo ...... 331 Ghaffari, Mahsa ...... 1195 Fischer, Arnout R.H...... 348 Ghosh, Anastasiya Pocheptsova .414, 525, 1038, E Fischer, Eileen ...... 1, 1118, 1184, 1249 1248 Eagar, Toni ...... 1192 Fishbach, Ayelet ...... 628, 949 Ghosh, Bikram ...... 1069 Ebrahimi, Mahdi ...... 1192, 1199 Fitzgerald, Paula ...... 1251 Ghoshal, Tanuka ...... 560 Eckhardt, Giana ...... 200, 272 Fitzsimons, Gavan . 847, 930, 1220, 1227, 1242 Gier, Nadine ...... 891 Eckhardt, Giana M...... 482 Fitzsimons, Gavan J...... 1086 Giesler, Markus ...... 1248 Eckles, Dean ...... 863 Fleury, Fernando A...... 1185 Gilad, Zohar ...... 1195 Eckmann, Lisa ...... 308 Flicker, Marcia ...... 1251 Gill, Tripat ...... 1195 Edinger-Schons, Laura-Marie ...... 379 Florack, Arnd ...... 335, 1192 Gillespie, Brian ...... 362 Eelen, Jiska ...... 291 Ford, John B...... 602 Gilovich, Tom ...... 1038 Effron, Daniel A...... 1096 Fotheringham, Darima ...... 337 Gino, Francesca ...... 1017 Egger, Martin ...... 1192 Fox, Craig R...... 943 Giuffredi-Kähr, Andrea ...... 974 Eichinger, Isabel ...... 857 Frank, Darius-Aurel ...... 1194 Giurge, Laura M...... 365 Ein-Gar, Danit ...... 494 Franke, Nikolaus ...... 857 Givi, Julian ...... 367 Ekpo, Akon E...... 1118 Franzen, Léon ...... 1194 Gloukhovtsev, Alexei ...... 1206, 1223 Eksell, Jörgen ...... 1187 Fridman, Ariel ...... 1096 Gneezy, Ayelet ...... 1129 El Hazzouri, Mohammed ...... 1232 Friedmann ...... 339 Godefroit-Winkel, Delphine ...... 1106 El Kihal, Siham ...... 831 Friedrich, Goetz ...... 825 Gohary, Ali ...... 1196 El-Bialy, Rowan ...... 310 Fritz, William ...... 1248 Goksel, Selin ...... 958, 958 Elder, Ryan ...... 471 Fritze, Martin P ...... 1194 Goldsmith, Kelly ...... 1193, 1212 Ellen, Pamela Scholder ...... 1251 Fröberg, Emelie ...... 463 Goldszmidt, Rafael ...... 760 Ellsaesser, Florian ...... 831 Fuchs, Christoph ...... 183, 1052 Golf-Papez, Maja ...... 368 Elmoukhliss, Mehdi ...... 1193 Fuchs, Klaus ...... 1214 Gollnhofer, Johanna ...... 31, 987 Elshiewy, Ossama ...... 312 Fuchs, Matthias ...... 340 Gonçalves, Dilney ...... 814 Emontspool, Julie ...... 57 Fukuda, Reo ...... 1181 Gong, Xiushuang ...... 1234 Emrich, Oliver ...... 1195 Fulmer, Alexander G...... 342, 614 Gonzales, Gabriel E...... 1196 Engeler, Isabelle ...... 740 Fulmer, Alexander Goldklank ...... 1032 González, Pilar Castro ...... 874 Engeset, Marit ...... 1197, 1231 Fulvio-Mason, Olivia ...... 1194 Goode, Miranda R...... 370 Epley, Nicholas ...... 968 Funk, Gabriela ...... 1080 Goodman, Joseph K...... 1038 Epstein, Ziv ...... 863 Fürst, Andreas ...... 522 Goodwin, Geoff ...... 863 Erevelles, Sunil ...... 314 Goodyear, Laura ...... 372, 1225 Errmann, Amy ...... 1193 Goor, Dafna ...... 847, 896, 968, 1006 Eslan, Camille ...... 316 G Gorbaniuk, Oleg ...... 1204 Esteky, Sina ...... 1193 Gaerth, Maximilian ...... 344, 562 Gordeliy, Ivan ...... 1247 Estes, Zachary . . . . . 318, 412, 1170, 1211 Gaertig, Celia ...... 886, 958, 1091 Gorge, Helene ...... 610 Etkin, Jordan ...... 181, 1101 Gaeth, Gary ...... 1188 Gorissen, Karen ...... 374 Euh, Hyun ...... 389, 1197 Gaffney, Amber ...... 1237 Gorji, Paniz ...... 1196 Evangelidis, Ioannis . . . 886, 949, 949, 1022 Gal, David ...... 254, 398 Goss, R Justin ...... 1196 Evans, Sarah ...... 62 Gallage, Samanthika ...... 346 Gottschalk, Sabrina ...... 548 Evers, Ellen R. K...... 902 Gallo, Iñigo ...... 1038 Goukens, Caroline ...... 398, 400, 667 Gamlin, Jessica ...... 230 Gourmelen, Andréa ...... 633 Gao, Cheng ...... 1201 Govind, Rahul ...... 974 F Gao, Fei ...... 350 Grace, Sarah ...... 1196 Gao, Huachao ...... 352 Grant, Annetta ...... 987, 1112 Fajardo, Tatiana M...... 1150 Granulo, Armin ...... 1052 Fajardo, Tatiana Marie ...... 222 Gao, Leilei ...... 278, 816 Gao, Randy Y...... 886 Grappe, Cindy ...... 1196 Falk, Emily B ...... 1231 Graul, Antje ...... 502 Falk, Emily B...... 891 Gao, Xin ...... 348 García-Rosell, José-Carlos ...... 1245 Greenacre, Luke ...... 1197 Falk, Tomas ...... 1206, 1223 Grewal, Dhruv ...... 322, 836 Fan, Linying (Sophie) ...... 320 Gardner, Meryl P...... 1251 Garg, Nitika ...... 974 Grewal, Lauren ...... 1058 Fan, Yafeng ...... 1193 Griffin, Dale ...... 1091 Fang, Hejin ...... 1237 Garnier, Marion ...... 1230 Gaston-Breton, Charlotte ...... 354, 490 Grillo, Tito L. H...... 1032 Fang, Xiang ...... 1242 Grinstein, Amir ...... 836, 896, 1247 Farace, Stefania ...... 322 Gathen, Constanze ...... 72 Gatter, Shirin ...... 357 Grohmann, Bianca ...... 1194, 1236 Faro, David ...... 958 Grønhøj, Alice ...... 600, 1221, 1226 Fehl, Amy Greiner ...... 324 Gaur, Sanjaya Singh ...... 766 Gaustad, Tarje ...... 359 Grønseth, Bjørn Ove ...... 1205 Fei, Lin ...... 326, 1091, 1155 Groth, O. H...... 1074 Feinberg, Fred ...... 1250 Gaviria, Pilar Rojas ...... 1106 1256 / Author Index

Gruen, Adèle ...... 244, 1214 Hedgcock, William ...... 868 Hyodo, Jamie D...... 1058 Guèvremont, Amélie ...... 1196 Heitmann, Mark ...... 831, 1212 Hyun, Hyowon ...... 1199 Guilbeault, Douglas ...... 842 Helme-Guizon, Agnès ...... 192 Hyun, Na Kyong ...... 907 Guillard, Valérie ...... 377 Hemetsberger, Andrea ...... 121 Hӓubl, Gerald ...... 1175 Gullo, Kelley ...... 847 Henkel, Alex ...... 1052 Gunadi, Manissa P...... 1022 Herd, Kelly B...... 614 Güntürkün, Pascal ...... 379 Herrmann, Andreas ...... 393 I Guo, Bingxuan ...... 1197 Hershfield, Hal ...... 1064 Iacobucci, Dawn ...... 1249 Guo, Leilei ...... 1207 Herter, Márcia Maurer ...... 1212 Ibrahim, Nahid ...... 1175 Guo, Yafei ...... 1197 Herziger, Atar ...... 913 Iesue, Joseph ...... 1187 Guo, Zhaoyang ...... 1209 Herzog, Walter ...... 402, 657 Ikonen, Iina ...... 422 Gupta, Samir ...... 1129 Heuvinck, Nico ...... 1229 Inbar, Yoel ...... 902 Gurova, Olga ...... 554 Higgins, Matthew ...... 16 Ince, Elise Chandon ...... 414, 661 Györke, Mariann ...... 324 Hildebrand, Christian . . 214, 831, 1011, 1052, Inman, J. Jeffrey ...... 981 1074, 1185, 1248 Inman, Jeff ...... 1249 Hildebrand, Diogo ...... 404 Inman, Jeffrey ...... 1247 H Hiler, Jacob Lee ...... 1245, 1245 Isaac, Mathew S...... 425, 486, 1017 Ha, Young-Won ...... 1190 Hill, Ronald Paul ...... 1248, 1249, 1250 Isaac, Matthew ...... 851 Haanpää, Minni ...... 1245 Hill, Tim ...... 1198 Ishii, Hiroaki ...... 1199, 1219 Haasova, Simona ...... 335, 1190 Hinze, Tobias ...... 406 Isisag, Anil ...... 1112 Habib, Rishad ...... 101 Hirt, Edward R...... 936 Iyengar, Raghuram ...... 681 Hackley, Chris ...... 1027 Ho, Ming-Shen ...... 786, 936 Izadi, Anoosha ...... 1192, 1199 Hackley, Rungpaka Amy ...... 1027 Hoang, Chi ...... 930 Hadar, Liat ...... 544, 1022 Hoffman, Donna L...... 918, 1248, 1249 Haderdauer, Ann-Kathrin ...... 1195 Hoffmann, Stefan ...... 1205 J Hofmann, Eva ...... 94 Hadi, Rhonda ...... 1074, 1248 Jaakkola, Elina ...... 1183 Hafeez, Ajmal ...... 1197 Hofstetter, Reto ...... 1052, 1080 Hogg, Margaret K...... 1106 Jacob, Jorge Rodrigues ...... 1200 Hagerty, Serena ...... 902 Jacob, Pradeep ...... 427 HaghighiBardineh, SeyedNasir . . . .673, 676 Hogg, Margaret Kathleen ...... 1220 Holmqvist, Jonas ...... 1194 Jagadale, Sujit Raghunathrao ...... 1118 Hair, Michael ...... 570 Jain, Shailendra Pratap . . 738, 851, 974, 1251 Halevy, Nir ...... 847 Holthoewer, Jana ...... 1052 Hong, Jennifer Seokhwa ...... 408, 410 Jang, Eunyoung ...... 432, 434 Halkias, Georgios ...... 381, 1192 Jang, Hyunkyu ...... 430 Hall, Matthew J...... 383, 1134 Hong, Soonkwan ...... 1198 Hong, Ying-yi ...... 1188 Jang, Minkwang ...... 1064 Hall, Matthew James ...... 1058 Jang, Wonseok ...... 769 Hamilton, Mitchell ...... 212, 1231 Hou, Yuansi ...... 1198 Hovy, Dirk ...... 981 Janiszewski, Chris ...... 769, 857, 1022 Hamilton, Rebecca ...... 1144 Jarrier, Elodie ...... 1220 Hammerschmidt, Maik ...... 657 Howard, Chuck ...... 372, 925 Howard, Ray Charles “Chuck” . . . . . 1091 Jasinenko, Anna ...... 287 Hampton, William ...... 1074 Jaufenthaler, Philipp ...... 436 Han, Eunha ...... 1197 Howe, Holly ...... 930, 1150 Howe, Holly S ...... 1217 Jemison, Courtney ...... 1069 Han, Eunjoo ...... 385 Jeon, Seongun ...... 1200 Han, Hyerin ...... 389, 1197 Hoyer, Wayne D...... 974 Hsee, Christopher ...... 1207 Jeong, Haeyoung ...... 1241, 1241 Han, Jerry J...... 907 Jia, He (Michael) ...... 438, 907, 1242 Han, Yegyu ...... 578, 918 Hsee, Christopher K...... 886 Hsu, Jen-Hsien ...... 77 Jiang, Jing . 1193, 1233, 1233, 1234, 1234, 1234 Handelman, Jay ...... 987, 1219 Jiang, Lan ...... 1200 Hang, Haiming ...... 391 Hu, Yansong ...... 1242 Hu, Yu (Jeffrey) ...... 350 Jiang, Yiran ...... 790 Hao, Xianyu ...... 1197 Jiang, Yuwei .320, 500, 1198, 1225, 1225, 1229, Hardisty, David ...... 1091 Huang, Der-Wei ...... 896, 936 Huang, Junhui ...... 1199 1238 Harman, Brian ...... 43 Jin, Fei ...... 440 Harnish, Elana ...... 1245 Huang, Li ...... 1189, 1199, 1240 Huang, Liang ...... 414 Jin, Jia ...... 1210 Hartl, Barbara ...... 94 Jin, Liyin ...... 685 Hartman, Anna E ...... 1232 Huang, Mei ...... 1199 Huang, Szu-chi . . . . 1043, 1129, 1155, 1248 Jin, Yufang ...... 1200 Hartman, Anna E...... 1118 Jinjun, Yu ...... 1200 Hartmann, Benjamin Julien ...... 47 Huang, Xun (Irene) ...... 500, 1228 Huang, Yanliu ...... 1216 Joe, Gladstone ...... 825 Hartmann, Jochen ...... 214, 831 Johar, Gita ...... 1248 Hartmann, Maximilian Alexander . . . . 1184 Huang, Zhihong ...... 1183 Huber, Joel ...... 1247 Johar, Gita V...... 851, 863, 1043, 1043 Hasenmaile-Aspin, Johanna ...... 393 John, Deborah Roedder ...... 631, 1000 Hassan, Louise May ...... 1213, 1213 Huh, Young Eun ...... 418, 851 Humayun, Mariam ...... 987 John, Leslie ...... 1080 Hattat, Ahmet M ...... 570 Johnson, Aaron P...... 1194 Häubl, Gerald . 340, 857, 943, 963, 1165, 1232 Humphreys, Ashlee . . . . . 918, 1017, 1247 Hung, Wei-Ken ...... 1188 Jones, Aziza Caimile ...... 1124 Haumann, Till ...... 379 Jones, Hunter ...... 1201 Hawa, Sarah ...... 1182 Hur, Elina Yewon ...... 1175 Husemann, Katharina ...... 272 Jounaid, Anis ...... 81 Haws, Kelly ...... 644, 1212 Jr., Robert Wyer ...... 1224 Haws, Kelly L...... 994, 1160 Husemann, Katharina C...... 200 Hussein, Mohamed A...... 943, 1129 Juma, Stephen ...... 1201 Hayran, Ceren ...... 1198 Jun, Sunkyu ...... 642 He, Daniel ...... 395 Hussein, Mohamed Ayman ...... 1155 Hütte, Carina Sophia ...... 1242 Jung, Minah ...... 949, 1144 He, Dongjin ...... 1198, 1225 Jung, Minah H...... 886, 1022 He, Yun ...... 1208 Hüttl-Maack, Verena ...... 357, 420 Huynh, Denny ...... 1220 Jung, Suyeon ...... 457 Healey, Patrick Matthew ...... 1198 Jung, Wonsuk ...... 868, 1201 Advances in Consumer Research (Volume 48) / 1257

Just, David R...... 1241 Kim, Yeonjeong ...... 842 Jüttner, Uta ...... 1223 Kim, Youngju ...... 1203 L Kimura, Junko ...... 108 L.Kettle, Keri ...... 963 Kim-Vick, Jihyun ...... 1203 Laamanen, Mikko ...... 482 K King, Dan ...... 1182, 1204 Labrecque, Lauren I ...... 1185, 1211 Labroo, Aparna ...... 1043 Kahn, Barbara E...... 1134 Kinsella, Elaine Louise ...... 749 Labroo, Aparna A ...... 1124 Kaiser, Ulrike ...... 857 Kireyev, Pavel ...... 831 Labroo, Aparna A...... 467, 525, 531 Kakaria, Shobhit ...... 1201 Kirk, Colleen P...... 624 Lages, Cristiana ...... 1229 Kakuda, Nakaya ...... 1200, 1201 Kizgin, Hatice ...... 143 Laghaie, Arash ...... 484 Kallmuenzer, Andreas ...... 618 Kleinlercher, Kristina ...... 461 Lähteenmäki, Liisa ...... 1226, 1226 Kamins, Michael ...... 1251 Klesse, Anne-Kathrin ...... 857 Lai, Ai-Ling ...... 16 Kamiya, Annaysa Salvador Muniz . . . . 1183 Kluger, Julia ...... 1242 Lai, Edward Yuhang ...... 486 Kamleitner, Bernadette ...... 1211, 1229 Knoferle, Klemens ...... 930, 1226 Lamberton, Cait ...... 803, 1248 Kan, Christina ...... 1022, 1204 Knutson, Brian ...... 891 Landgraf, Polina ...... 1205 Kaneko, Mitsuru ...... 11 Ko, Min Jeong ...... 1204 Landwehr, Jan R...... 308 Kang, Christine ...... 1201 Koch, Teresa ...... 335 Lapostolle, Mathilde ...... 659 Kang, In-Hye ...... 445, 1134 Kocher, Bruno ...... 86, 90 Laran, Juliano ...... 769 Kansal, Aparna ...... 447 Kogut, Tehila ...... 494 Laroutis, Dimitri ...... 488 Kapitan, Sommer ...... 1221 Kolanska, Magdalena ...... 1204 Larrick, Rick ...... 1227 Kapoor, Ankur ...... 764, 1245 Kolesova, Svetlana ...... 463 Lasaleta, Jannine ...... 404 Karachi, Hedieh ...... 1202 Koley, Shruti ...... 1204 Lasarov, Wassili ...... 1205 Karampela, Maria ...... 289 Koller, Monika ...... 1204 Lasri, Sarah ...... 1206 Karat, Ella ...... 954 Komarova, Yuliya ...... 1222 Lau-Gesk, Loraine ...... 1237, 1237 Kardas, Michael ...... 968 Konya-Baumbach, Elisa ...... 218 Laurent, Gilles ...... 86 Kardes, Frank ...... 954 Koo, K.B...... 1204 Laurin, Kristin ...... 943 Kassirer, Samantha ...... 958 Koo, Minjung ...... 1239 Lavrusheva, Olga ...... 1206 Katitas, Ayça ...... 618 Kopalle, Praveen Kumar ...... 1251 Leary, R. Bret ...... 1196 Kato, Bianca ...... 1202 Kordrostami, Elika ...... 602, 1221 LeBoeuf, Robyn ...... 1240 Katz, Daniel ...... 925 Koschate-Fischer, Nicole ...... 522, 524 Lee, Andrew ...... 1206 Kay, Aaron C...... 1086 Koskela-Huotari, Kaisa ...... 1215 Lee, Angela Y...... 1249 Kedzior, Richard ...... 880 Kößmeier, Christina ...... 1205 Lee, Chang-Yuan ...... 1101 Keh, Hean Tat ...... 1196, 1239 Kou, Sining ...... 1205 Lee, Christina Kwai Choi ...... 38 Keinan, Anat ...... 280, 847, 1006 Koulaei, Afra ...... 1205, 1205 Lee, Hsiao-Ching ...... 271 Kelleher, Carol ...... 582, 1106 Kousi, Sofia ...... 381 Lee, Hsin-Hsuan Meg ...... 490 Keller, Carolyn Wells ...... 1202 Kozinets, Robert ...... 1249 Lee, Hyun-Woo ...... 1218 Keller, Punam ...... 1249 Krabbe, Mathias Sosnowski ...... 1191 Lee, Jeff ...... 1247 Kenning, Peter ...... 1184, 1185 Krajbich, Ian ...... 693, 1139 Lee, Jeffrey ...... 981 Kermani, Saeid ...... 1202 Kralj, Sonja N...... 465 Lee, Kyoungmi ...... 1204 Kerns, Avery ...... 1245 Kraus, Florian ...... 344, 562 Lee, Leonard ...... 1249 Kerrigan, Finola ...... 797 Krause, Franziska ...... 857 Lee, Saerom ...... 385 Kettle, Keri L...... 963 Krause-Galoni, Rebecca J...... 467 Lee, Seojin Stacey ...... 1227 Khambatta, Poruz ...... 994 Kravets, Olga ...... 1106 Lee, Shinhyoung ...... 1218 Khamitov, Mansur . . . . 825, 974, 974, 1080 Krefeld-Schwalb, Antonia ...... 1139 Lee, Sujin ...... 453 Khan, Amna ...... 1220 Krekels, Goedele ...... 86, 90, 1191 Lee, Sunme ...... 1206 Khan, Jashim ...... 209 Krey, Nina ...... 469 Léger, Pierre-Majorique ...... 1227 Khan, Uzma ...... 913, 1248 Krijnen, Job M. T...... 943 Lehmann, Donald ...... 220, 1185, 1249 Khenfer, Jamel ...... 451 Krishna, Aradhna ...... 471 Lei, Jing ...... 755 Kidwell, Blair ...... 531 Krishnan, Shanker ...... 896 Leliveld, Marijke C...... 492 Kim, B. Kyu ...... 438, 907 Kristianto, Nathasya Pricilia B . . . . . 1219 Levav, Jonathan ...... 949 Kim, Claire Heeryung ...... 1203 Kristofferson, Kirk ...... 337, 690, 1058 Levin, Irwin ...... 1188 Kim, Dongyoup ...... 1199, 1202 Kronrod, Ann ...... 836, 954, 954, 1247 Levine, Emma ...... 958 Kim, Eunice ...... 1203 Krüger, Tinka ...... 1205 Levontin, Liat ...... 494, 1195 Kim, Hyoseok ...... 963 Kuester, Sabine ...... 218 Lewis, Joshua ...... 497 Kim, Junghyun ...... 1203 Kumar, Amit ...... 968 Li, Eric ...... 987 Kim, Junha ...... 453 Kumar, Smriti ...... 474 Li, Ran ...... 1207, 1207, 1229 Kim, Kihyun Hannah ...... 822, 1243 Kumbargeri, Amogh ...... 478, 974 Li, Ruiqin ...... 1207 Kim, Kyeongheui ...... 457, 868 Kunchamboo, Vimala ...... 38 Li, Ruoou ...... 236, 414 Kim, Minki ...... 1234 Kupor, Daniella ...... 943, 1043 Li, Shaobo (Kevin) ...... 500 Kim, Moon-Yong ...... 1202 Kurt, Didem ...... 1000 Li, Stella Yiyan ...... 502 Kim, Myung Joo ...... 1203 Kuruoglu, Alev ...... 1230 Li, Wei ...... 1206 Kim, Nicole ...... 896 Kutsch, Elmar ...... 608 Li, Xi ...... 825 Kim, Pureum ...... 1038 Kwak, Hyokjin ...... 808 Li, Xilin ...... 1207 Kim, Sahoon ...... 1203, 1203 Kwakman, Pascale ...... 1215 Li, Xitong ...... 350 Kim, Sara ...... 1234 Kwan, Canice M. C...... 480 Li, Yang ...... 981 Kim, Soo ...... 847 Kwon, Jaehwan ...... 576 Li, Yi ...... 198 Kim, Tami ...... 1170 Kwon, Junbum ...... 1203 Liang, Jianping ...... 1207 Kim, Yaeeun ...... 459 Kwong, Jessica Y. Y...... 1209 Liang, Shuyu ...... 1207, 1208 Kim, Yenee ...... 455 Liaukonyte, Jura ...... 1241 1258 / Author Index

Lichrou, Maria ...... 1187 MacInnis, Deborah ...... 1032, 1249 Meng, Matthew D...... 535, 1196 Lichtenstein, Donald ...... 1022 Maclaran, Pauline ...... 333, 1106 Menon, Geeta ...... 1129, 1214 Lieberman, Alicea ...... 958, 1129 Madan, Shilpa ...... 851, 1043 Merian, Sybilla ...... 1214 Light, Nicholas ...... 1208, 1208 Maehle, Natalia ...... 1211 Mesler, Rhiannon MacDonnell . . . 1189, 1225 Liljedal, Karina T...... 1184 Mafael, Alexander ...... 548 Messinger, Paul Richard ...... 1058 Lim, Sarah ...... 504, 1155 Magede, Thomas ...... 143, 1228 Meyer, Michael ...... 1155 Lima, Vitor M...... 506 Maheswaran, Durairaj ...... 851 Meyer, Robert ...... 1247, 1248 Lin, Chia Hua ...... 1187 Mai, Christoph ...... 522 Meyer-Waarden, Lars ...... 1219 Lin, Meng-Hsien (Jenny) . . . . . 762, 1215 Mai, Robert ...... 1205 Meynhardt, Timo ...... 287 Lin, Stephanie ...... 1124 Mai, Yixia ...... 518 Meyvis, Tom ...... 400, 1011 Lin, Stephanie C...... 994 Maiberger, Tobias ...... 524 Miao, Jin ...... 395 Lin, Wang ...... 1208 Maimaran, Michal ...... 525 Michał, Wilczewski ...... 1204 Lin, Yu-Wei ...... 799 Main, Kelley ...... 1232 Michel, Géraldine ...... 190 Linares, Claire ...... 1080 Malär, Lucia ...... 974 Micheli, Maria Rita ...... 246 Lindridge, Andrew ...... 1192, 1220 Maleev, Nikita ...... 1185 Mick, David Glen ...... 4 Ling, I-Ling ...... 1129 Malik, Sumit ...... 1074 Milkman, Katherine ...... 1101 Lingg, Linda ...... 1223 Malkoc, Selin ...... 1101 Milkman, Katherine L...... 994 Linzmajer, Marc ...... 648 Malloy-Diniz, Leandro ...... 404 Miller, Elizabeth ...... 474 Lisjak, Monika ...... 337, 1000, 1248 Malter, Maayan ...... 1043 Millet, Kobe ...... 1210 Little, Vicki Janine ...... 38 Mandal, Sudipta ...... 527, 974 Mills, Adam ...... 1228 Liu, Hongyan ...... 1209 Mandel, Naomi ...... 427 Milne, George R...... 474 Liu, Jiajia ...... 1211 Manis, KT ...... 529 Mimoun, Laetitia ...... 1118, 1214 Liu, Jingshi ...... 1208 Manke, Birte ...... 987 Min, Kate E...... 847 Liu, Joyce Jingshi ...... 508 Manoharan, Bhupesh ...... 1211, 1211 Minina, Alisa ...... 537 Liu, Maggie Wenjing ...... 1199, 1236 Mantonakis, Antonia ...... 963 Minor, Michael ...... 1245 Liu, Meng ...... 786 Mao, Huifang ...... 851 Minson, Julia ...... 1017 Liu, Mengmeng ...... 1209 Marckhgott, Eva ...... 1211 Mishra, Nirajana ...... 541, 1214 Liu, Peggy ...... 784, 1249, 1251 Mariadassou, Shwetha ...... 994 Mitchell, Alex ...... 592 Liu, Peggy J...... 847, 994 Markos, Ereni ...... 1211 Mitchell, Vince ...... 1240 Liu, Shian-Ko ...... 1189 Marth, Sarah ...... 94 Mitkidis, Panagiotis ...... 1194 Liu, Wei ...... 1209 Martin, Brett A.S...... 801 Mittal, Ameya ...... 1245 Liu, Wumei ...... 512 Martin, Christian ...... 1190 Mittal, Chiraag ...... 1204 Liu, Xiao ...... 1022 Martin, Diane ...... 1192 Mittal, Vikas ...... 1249 Liu, Yan (Lucy) ...... 1022 Martin, Ingrid ...... 1251 Mizukoshi, Kosuke ...... 1226 Liu, Yingyue ...... 1209, 1242 Martin, Nathanael S...... 1212 Mogilner, Cassie ...... 740 Loewenstein, George ...... 596 Martinez, Luis ...... 1182 Mohan, Bhavya ...... 1170 Lohan, Karan Pratap Singh ...... 1209 Martins, Fábio Miguel Ferrony Varela . . 1212 Mohr, Dr. Iris ...... 1245 Longoni, Chiara ...... 410, 913 Martins, Renata ...... 1212 Moisio, Risto ...... 1214 Lopes, Maira ...... 1246 Mas, Erick M...... 531, 1193, 1212, 1248 Mokarram-Dorri, Sadaf ...... 303 Lopez, Alberto ...... 1080 Masè, Stefania ...... 537 Moldes, Olaya ...... 1214 López, Inés López ...... 857 Mason, Marlys ...... 324, 1251 Moldovan, Sarit ...... 544, 546 López, Rocío Alarcón ...... 857 Mason, Richard ...... 1064 Mollen, Saar ...... 1215 Lou, Yung-Chien ...... 1209 Masuda, Akiko ...... 566 Möller, Jana ...... 548 Lowe, Michael ...... 907, 907, 1000 Materne, Maxine ...... 1212 Monga, Ashwani ...... 822, 1243 Lowe, Michael L...... 1150 Mathur, Pragya ...... 851 Monnier, Arnaud ...... 550, 857 Lowrey, Tina M...... 533, 954 Matos, Geraldo ...... 212 Monnot, Elisa ...... 620 Lu, Shijie ...... 825 Matsumoto, Kohei ...... 1191 Monsurrò, Luigi ...... 552 Luan, Chi-Cheng ...... 1210, 1210 Matsushita, Koji ...... 1213 Moore, Sarah G ...... 1058, 1204, 1247 Luangrath, Andrea Webb . . . . . 868, 1247 Mattila, Anna ...... 896 Moore, Sarah G...... 1017 Luedicke, Marius K...... 918 Matz, Sandra ...... 825 Moorhouse, Michael ...... 370 Lukas, Marcel ...... 372, 925, 1091 Mau, Gunnar ...... 1186 Moorman, Christine ...... 1249 Luna, David ...... 1011, 1074 Mazar, Nina ...... 1250 Moreau, Page ...... 1249 Lunardo, Renaud ...... 514 McCartney, W. Benadict ...... 1006 Moreira, Catia Silva da C...... 126 Luo, Siria Xiyueyao ...... 1210 McCullough, Heath ...... 695 Morewedge, Carey ...... 1101 Luri, Ignacio ...... 1069 McFarland, Richard G...... 455 Morewedge, Carey K...... 1249 Lurie, Nicholas ...... 1205, 1247 McFerran, Brent ...... 1198, 1249 Morgan, Carter ...... 963 Lynch, John ...... 1249 McGowan, Miriam ...... 1213, 1213 Morin-Delerm, Sophie ...... 574 Lynch, John G...... 925 Mead, Nicole L...... 1006 Morozova, Daria ...... 554 Lyu, Dong ...... 1210 Mecit, Alican ...... 533 Morren, Meike H...... 896 Lyu, Jennifer ...... 925 Meersseman, Eva ...... 1213 Morrin, Maureen ...... 1217 Lyu, Wei ...... 773 Mehr, Katie S...... 994 Morris, Joshua I...... 994 Mehta, Ravi ...... 1165 Morris, Michael ...... 851 Meister, Matthew ...... 925 Morwitz, Vicki ...... 1022 M Melumad, Shiri ...... 981, 1247 Morwitz, Vicki G. . . 266, 410, 657, 1011, 1249 Ma, Jingjing ...... 1211 Melzner, Johann ...... 1011, 1165 Mosleh, Mohsen ...... 863 Maalej, Mariem El Euch ...... 604 Memmi, Sarah A...... 181 Mourali, Mehdi ...... 1215, 1216 Macdonald, Emma K...... 582 Mende, Martin ...... 836, 1250 Mourey, James A...... 1011 Maciejovsky, Boris ...... 842 Meng, Lu ...... 1213 Mrkva, Kellen ...... 1221 Advances in Consumer Research (Volume 48) / 1259

Mugel, Ophelie Elise ...... 1245 Olson, Nicholas J...... 1233 Pedersen, Per Egil ...... 1184 Mukherjee, Ashesh ...... 1222 Omar, Suha ...... 77 Peixoto, Fabricia V...... 207 Mukherjee, Sayantani ...... 1237 Onculer, Ayse ...... 112, 724, 726, 1235 Pellandini-Simanyi, Lena ...... 1190 Mukhopadhyay, Anirban .851, 1017, 1208, 1247, Onuklu, Nur Yazgan ...... 1217 Peñaloza, Lisa ...... 582, 1106 1249 Oppewal, Harmen ...... 1197 Pendarvis, Nicholas ...... 212 Mulier, Lana ...... 556 Ordabayeva, Nailya . 896, 936, 1006, 1248, 1250 Pendarvis, Nicholas J...... 918 Müller, Astrid ...... 1224 Ordenes, Francisco Villarroal ...... 1247 Peng-Li, Danni ...... 1220 Müller, Brigitte ...... 90 Ortiz, Maria ...... 1217 Pennycook, Gordon ...... 863 Munz, Kurt P...... 1011 Osborne, Delane ...... 104 Penz, Elfriede ...... 94 Munzel, Andreas ...... 1219 Ostberg, Jacob ...... 680 Peracchio, Laura ...... 1247 Muratore, Isabelle ...... 558 Ostinelli, Massimiliano ...... 1074, 1165 Pereira, Beatriz ...... 1225 Murphy, Stephen ...... 1112, 1198 Otter, Thomas ...... 484 Perkins, Andrew W...... 637 Murray, Ross ...... 1245 Ottlewski, Lydia ...... 1246 Peronard, Jean-Paul ...... 134 Myaeng, Seo Young ...... 418 Otto, Ashley ...... 238 Peschel, Anne Odile ...... 312 Otto, Ashley S...... 936 Pessoa, Luís ...... 506 Otto, Christian ...... 362 Petersen, Francine E...... 283 N Ourahmoune, Nacima ...... 1192 Petr, Christine ...... 1220 Nader, Brishna ...... 1215 Overton, Graham ...... 886, 1096 Petty, Richard ...... 1238 Nanarpuzha, Rajesh . . . . .710, 1027, 1222 Ozanne, Julie L...... 1250 Peverini, Paolo ...... 552 Narayanan, Priya ...... 1215 Ozanne, Marie ...... 896 Pezzuti, Todd ...... 1048 Nariswari, Angeline ...... 1215 Özçaglar-Toulouse, Nil ...... 782 Pfeiffer, Bruce E...... 722 Nasa, Jayant ...... 560 Ozcan, Tim ...... 570 Pham, Michel Tuan ...... 564 Nastasoiu, Alina ...... 202 Ozturk, Cem ...... 1000 Philp, Matthew ...... 586, 1218 Natter, Martin ...... 1214 Picot-Coupey, Karine ...... 469 Nave, Gideon ...... 825, 825 Pieters, Rik ...... 1231 Nayakankuppam, Dhananjay ...... 1206 P Pino, Giovanni ...... 1195 Neef, Patrick Wolfgang ...... 1216 Pacanowski, Carly ...... 1251 Pinto, Diego Costa ...... 1212, 1245 Nelson, Leif D...... 1022 Pachur, Thorsten ...... 1139 Pizzutti, Cristiane ...... 1032 Nenkov, Gergana Y...... 1250 Packard, Grant ...... 981, 1247 Plant, Bernice ...... 1217 Nepomuceno, Marcelo Vinhal ...... 586 Padhaiskaya, Tatsiana ...... 1218 Plassmann, Hilke ...... 891 Nervino, Esterina ...... 1228 Paharia, Neeru . . . . . 280, 344, 1134, 1248 Poehlman, T. Andrew ...... 1220 Netzer, Oded ...... 831, 1249 Pai, Jieun ...... 453 Pogacar, Ruth ...... 954, 1247 Neybert, Emma ...... 1216 Palmeira, Mauricio ...... 868 Pol, Gratiana ...... 1048 Nguyen, Hang ...... 1199 Palo, Teea ...... 1215 Polman, Evan ...... 1236 Nguyen, Hieu ...... 1251 Palumbo, Helena ...... 1218 Polyakova, Alexandra ...... 590 Nguyen, Thanh ...... 874 Pancer, Ethan ...... 1218 Pomerance, Justin ...... 1208 Nian, Tingting ...... 1170 Pandelaere, Mario ...... 1201 Pomies, Anissa ...... 1246 Nindl, Fabian ...... 649 Pandey, Arpita ...... 738 Ponchio, Mateus Canniatti ...... 205 Nishii, Mayuko ...... 1216 Pang, Jun ...... 1234 Poncin, Ingrid ...... 240 Niu, Mengmeng ...... 562 Panteqi, Mivena ...... 1218 Pond, Colton ...... 913 Nøjgaard, Mikkel Ørholm ...... 1216 Paolacci, Gabriele ...... 902 Poole, Maxwell ...... 1218 Nordgren, Loran F...... 981 Pappalardo, Janis ...... 1251 Posten, Ann-Christin ...... 712 Norrgrann, Anu ...... 97 Paquier, Marie-Catherine ...... 574 Pounders, Kathrynn ...... 880 Norton, Michael . . . . 508, 902, 1170, 1195 Parameswaran, M. G...... 1182 Pradhan, Anuja ...... 1220 Norton, Michael I...... 847, 968 Parguel, Béatrice ...... 160 Praxmarer-Carus, Sandra ...... 72 Noseworthy, Theodore J...... 730 Park, Alexander ...... 1064 Preiksaitis, Kimberley ...... 592 Novak, Thomas ...... 918, 1248 Park, Choong W...... 1209 Prelec, Drazen ...... 842 Novakowski, Dallas ...... 1216 Park, Hanyong ...... 576 Price, Linda L ...... 717 Nunan, Daniel ...... 1195 Park, Hyun Young ...... 285 Prinsloo, Emily ...... 596 Nunoo, Jemima ...... 227 Park, Jaewoo ...... 1199, 1219 Prokopec, Sonja ...... 724, 726, 744 Nusrat, Farhana ...... 1216 Park, Jen H...... 1043, 1160 Puntoni, Stefano . 958, 1052, 1155, 1181, 1248 Nyilasy, Gergely ...... 1217 Park, Jihye ...... 1218 Purdie-Greenaway, Valerie ...... 1200 Park, Jooyoung ...... 1219 Pusaksrikit, Theeranuch ...... 1220 Park, Jungkun ...... 1199, 1202 Puzakova, Marina ...... 808, 836 O Park, Kiwan ...... 1227 O’Brien, Ed ...... 968, 1032 Park, Lisa ...... 1144 Q O’Donnell, Matthew Brook ...... 1231 Park, Sang Kyu ...... 578 O’Donnell, Michael ...... 902 Park, Sangchul ...... 1218 Qiu, Lingyun ...... 1234 O’Malley, Lisa ...... 1187 Park, Sohyeon ...... 1219 Quental, Camilla ...... 1106 Oba, Demi ...... 1217 Park, Taehoon ...... 445 Querci, Ilaria ...... 552 Oh, Ga-Eun (Grace) ...... 1217, 1217 Parker, Jeffrey ...... 671 Quintanilla, Claudia ...... 256 Oh, Hyewon ...... 1038 Parmentier, Marie-Agnès ...... 880 Quoidbach, Jordi ...... 1195 Oh, Travis Tae ...... 564 Pasdiora, Maria Alice ...... 580 Ohira, Shuji ...... 566, 1226 Patry-Beaudoin, Gabrielle ...... 1219 Ojansivu, Ilkka ...... 1232 Pavone, Giulia ...... 1219 R Ok, Ekin ...... 101 Pechmann, Connie ...... 1249 Rabino, Rebecca ...... 529, 598 Okutur, Nazli Gurdamar ...... 568 Peck, Joann ...... 868, 868, 1201 Radas, Sonja ...... 842 Olivola, Christopher ...... 825, 842 Pecot, Fabien ...... 574 Ragelienė, Tija ...... 600 1260 / Author Index

Raghubir, Priya ...... 1129, 1249 Sahabeh, Easa ...... 1228 Sepehri, Amir ...... 673, 676 Raghunathan, Raj ...... 560 Sahay, Arvind ...... 1215 Septianto, Felix ...... 1193 Rahmani, Leila ...... 1190 Saintives, Camille ...... 514 Serfas, Benjamin G...... 1224 Rahmani, Vahid ...... 602, 1221 Sajtos, Laszlo ...... 1222 Serin, Nuket ...... 679 Raies, Karine ...... 604 Sakashita, Mototaka ...... 108 Servadio, Luigi ...... 680 Rakshit, Krishanu ...... 447, 1211, 1211 Salisbury, Linda Court ...... 1250 Sevilla, Julio ...... 486, 644, 735 Ramos, Guilherme ...... 606 Salmela, Tarja ...... 1245 Sezer, Ovul ...... 968 Ramos, Jairo ...... 1221 Salvador, Marielle ...... 604 Shaddy, Franklyn ...... 949 Rand, David G...... 863 Samper, Adriana ...... 690 Shah, Anuj K...... 1144 Randers, Louise ...... 1221 Sample, Kevin L...... 644 Shah, Avni ...... 372, 1006 Rank-Christman, Tracy ...... 625, 1183 Samu, Sridhar ...... 1027, 1222 Shaji, Jossin ...... 1027, 1222 Rao, Akshay R...... 389 Sanders, Elisabeth ...... 524 Shamayleh, Ghalia ...... 874 Ravella, Haribabu ...... 918 Sandes, Fabio Shimabukuro ...... 1222 Shankar, Avi ...... 1112 Razmdoost, Kamran ...... 608 Sandrine, Mueller ...... 825 Shanks, Ilana ...... 836 Reczek, Rebecca Walker . . . 913, 1238, 1248 Sangle-Ferriere, Marion ...... 1222 Sharif, Marissa ...... 994 Reed, Americus ...... 963, 1231 Santana, Jannsen ...... 1027 Sharif, Marissa A...... 681 Reeves-Morris, Sophie Alexandra . . . .1221 Sarna, Nikkita ...... 1222 Sharifi, Shahin ...... 198 Regany, Fatima ...... 610 Sarstedt, Marko ...... 186 Shavitt, Sharon . . 203, 1150, 1203, 1247, 1248 Rehman, Varisha ...... 669 Satyavageeswaran, Prakash . .710, 1027, 1222 Shaw, Alex ...... 1086 Reich, Brandon ...... 612 Savani, Krishna ...... 851, 1043 Shaw, Deirdre ...... 683, 782 Reich, Taly ...... 342, 614, 1032, 1150 Savary, Jennifer ...... 1248 Shaw, Steven D...... 891 Reiff, Joseph ...... 1101 Sayin, Eda ...... 1074 Sheinin, Daniel A...... 570 Reimann, Martin ...... 1038, 1080, 1251 Sayman, Serdar ...... 112 Shen, Hao ...... 203, 1194 Reinhard, Julia ...... 616 Scaraboto, Daiane ...... 331, 880 Shen, Henry ...... 1226 Reitsamer, Bernd F...... 618 Schaerer, Michael ...... 588 Shen, Jie(Doreen) ...... 1183, 1224 Reniou, Fanny ...... 620 Schaffner, Dorothea ...... 1223 Shen, Liang ...... 1224 Rentfrow, Jason ...... 825 Schamp, Christina ...... 831, 1184, 1212 Shen, Luxi ...... 326 Richins, Marsha L...... 7 Schau, Hope ...... 1069, 1183 Shen, Ya-Ju ...... 1189 Riehle, Ramona ...... 622 Scheibehenne, Benjamin ...... 1139 Shen, Zhengyu ...... 685 Rifkin, Jacqueline R...... 1134 Scheidegger, Gianluca ...... 648 Shennib, Fuad A...... 949 Rifkin, Laura ...... 624 Scherer, Anne ...... 250, 1186 Shepherd, Steven ...... 1086 Rindfleisch, Aric ...... 295, 1224 Schill, Marie ...... 1106 Shi, Bing ...... 518 Ringberg, Torsten ...... 1074 Schindler, David ...... 524 Shi, Haijiao ...... 1224 Ringler, Christine ...... 625 Schlager, Tobias . . . . . 649, 651, 776, 1170 Shi, Yang ...... 480 Rizvi, Shaheer Ahmed ...... 1058 Schlosser, Ann ...... 653, 1223 Shim, Jisoo ...... 1203 Roberts, Annabelle ...... 628 Schmidt, Kristina ...... 657 Shiu, Edward ...... 1213, 1213 Robinot, Élisabeth ...... 160 Schmitt, Bernd ...... 1052, 1248 Shoham, Meyrav ...... 546 Rocha, Ana Raquel Coelho ...... 148 Schmitt, Julien ...... 659 Shrum, L. J...... 533, 954, 1247 Rocklage, Matthew D ...... 1247 Schneider, Gustavo ...... 661 Shu, Stephen ...... 1064 Rocklage, Matthew D...... 981 Scholz, Christin ...... 891 Shu, Suzanne ...... 297, 868, 868 Rodas, Maria A...... 631 Scholz, Joachim ...... 663 Shu, Suzanne B...... 223, 1038 Rodhain, Angélique ...... 633 Schöps, Jonathan D...... 115 Si, Kao ...... 687 Roese, Neal ...... 1202 Schreier, Martin ...... 183, 857, 857 Sieow, Justin ...... 1224 Rokka, Joonas ...... 1246 Schrift, Rom Y...... 1175 Silva, Lívia Lessa de A...... 126 Romani, Simona ...... 552 Schroeder, Jonathan ...... 874 Silver, Ike ...... 863, 1086, 1096 Romero, Marisabel ...... 1160 Schroeder, Juliana ...... 968 Silverman, Jackie . . . 994, 1091, 1096, 1101 Rose, Randy ...... 695 Schroll, Roland ...... 436 Simard, Caitlin Elizabeth ...... 130 Rosengren, Sara ...... 463 Schultz, Ainslie ...... 1251 Simmons, Joseph P...... 886, 1091 Rosenthal, Benjamin ...... 635, 1027 Schumacher, Anika ...... 667 Simonetti, Aline ...... 1201 Ross, Spencer ...... 1221 Schwarz, Norbert ...... 810, 863, 1238 Simonson, Itamar ...... 949, 949, 1160 Rotman, Jeff D...... 637 Schwarz, Sarah ...... 121 Simpson, Bonnie ...... 1225 Rotman, Jeffrey ...... 1202 Schweitzer, Nicola ...... 851 Singh, Jatinder J...... 974 Roux, Caroline ...... 1181, 1225, 1251 Scopelliti, Irene ...... 418, 596 Singh, Tanya ...... 1225 Roux, Dominique ...... 639 Scornavacca, Eusebio ...... 1187 Sinha, Jayati ...... 679, 1225 Roy, Arani ...... 1217, 1222 Scott, Maura ...... 1250 Sirieix, Lucie ...... 305, 1225 Ruan, Bowen ...... 794, 886 Scott, Maura L...... 836 Skubisz, Christine ...... 1251 Rubaltelli, Enrico ...... 1187 Scott, Rebecca ...... 1118 Slabbinck, Hendrik ...... 556, 1229 Rucker, Derek ...... 320, 467 Scott, Sydney ...... 958, 958, 1198 Slovic, Paul ...... 1187 Rucker, Derek D. . . . . 252, 254, 981, 1000 Sedghi, Tara Madleen ...... 420 Small, Deborah ...... 863 Rudd, John M...... 1242 Segal, Shoshana ...... 1058, 1223 Smaniotto, Cristiano ...... 57, 134 Rudolph, Thomas ...... 220, 461, 648 Seittu, Henriikka ...... 1223 Smaoui, Fatma ...... 173 Russell, Cristel ...... 1069 Sekar, Samuel B ...... 1223 Smeets, Paul ...... 1144 Ryu, Hyerin ...... 642 Sela, Aner ...... 578, 707, 1175 Smith, Evelyn ...... 653 Sellier, Anne-Laure ...... 1080 Smith, Evelyn Olivia ...... 1223 Selvanayagam, Karthik ...... 669 Smith, Jamie ...... 537 S Sénécal, Sylvain ...... 1227 Smith, Leah ...... 695 Sabadie, William ...... 1184 Sengupta, Rumela ...... 671, 1191 Smith, Ned ...... 842 Sadek, Jack ...... 1181 Seo, Yuri ...... 1193 Smith, Rosanna ...... 732, 735 Advances in Consumer Research (Volume 48) / 1261

Smith, Stephanie M...... 693, 1139 Tao, Tao ...... 1207, 1229 Utgård, Jakob ...... 359 So, Jane ...... 698 Tari, Anna ...... 728, 913 Utochkin, Denis ...... 1231 Soares, Raquel Reis ...... 1212 Täuber, Susanne ...... 492 Sobolev, Michael ...... 820 Taylor, Charles R...... 808 Södergren, Jonatan ...... 139 Taylor, Nükhet ...... 730 V Sokoll, Cynthia ...... 701 Tellis, Gerard ...... 1048 Valenzuela, Ana ...... 1011, 1074, 1248 Song, Camilla Eunyoung ...... 703, 707 Temerak, Mohamed Sobhy ...... 1229 Vallström, Niklas ...... 139 Song, Jiaqi (Flora) ...... 1225 Teng, Lefa ...... 1229 Valsesia, Francesca ...... 1058, 1134 Song, Jiaqi Flora ...... 1225 Teow, Jasper ...... 1160, 1165 Van Boven, Leaf ...... 1221 Sørensen, Elin Brandi ...... 354 Thai, Nguyen T ...... 1240 Van den Bergh, Bram ...... 744 Sousa, Ellen Campos ...... 1225 Thau, Stefan J ...... 275 van der Heijden, Kimberley . . . . .398, 400 Spangenberg, Katie ...... 425 Thøgersen, John ...... 1221 van der Laan, Nynke ...... 1215 Spence, Charles ...... 1219 Thomas, Manoj ...... 550 van der Sluis, Helen ...... 690 Spence, Mark T ...... 1191 Thomas, Tabitha S ...... 152, 156 van der Waal, Nadine ...... 1215 Spiller, Stephen A ...... 1192 Thomas, Tandy Chalmers ...... 62 van Doorn, Jenny ...... 836, 1052, 1235 Spotswood, Fiona ...... 1112 Thompson, Craig ...... 1112 van Hest, Iris ...... 1215 Springer, Fanny ...... 1204 Thomsen, Thyra Uth ...... 354 van Horen, Femke ...... 412, 1200, 1210 Spruyt, Adriaan ...... 1189 Thorel, Cerise ...... 1245 van Ittersum, Koert ...... 715 Sridhar, Karthik ...... 1074 Thuerridl, Carina ...... 1229 Van Ittersum, Koert ...... 1235 Srivastava, Arti ...... 710 Tiercelin, Alexandre ...... 1230 Van Kerckhove, Anneleen . . . 225, 755, 1232 Srivastava, Joydeep ...... 459 Tim, Hill ...... 1230 van Loo, Ellen J...... 720 Srna, Shalena ...... 1096, 1101 Timoshenko, Artem ...... 831 van Osselaer, Stijn ...... 857, 1155 Staats, Brad ...... 968 Ting, Tin-yuet ...... 1188 van Osselaer, Stijn M. J...... 248 Stagno, Emanuela ...... 1226 Tipaldi, Petra ...... 1214 van Trijp, Hans ...... 720 Stamatogiannakis, Antonios ...... 1205 Tobias, Ebert ...... 825 VanBergen, Noah ...... 1212 Stancu, Catalin Mihai ...... 1226 Togawa, Taku ...... 1199, 1219, 1230 Vandenbosch, Mark ...... 202 Stancu, Violeta ...... 1226 Tomar, Nitisha ...... 764, 1245 Varga, Marton ...... 891 Stanislawski, Sumire ...... 1226 Ton, Lan Anh N...... 732, 735 Velloso, Luciana ...... 244 Steffel, Mary ...... 1150 Tong, Lester ...... 891 Venkatraman, Rohan ...... 1232 Steinhart, Yael ...... 546 Tonietto, Gabriela ...... 1101 Venkatraman, Vinod ...... 891 Steinmetz, Janina ...... 712, 1124 Torelli, Carlos ...... 1183, 1203, 1203 Veresiu, Ela ...... 880, 918 Stenger, Thomas ...... 1246 Torelli, Carlos J...... 631 Verlegh, Peeter ...... 422, 1200 Stephen, Andrew T...... 1248 Tormala, Zakary ...... 943, 1086 Vermeir, Iris . . . . . 556, 1189, 1213, 1229 Stevens, Lorna ...... 1106 Torres, Lez Ecima Trujillo . . . . 1191, 1230 Véron, Tobias ...... 1223 Stoeckli, Sabrina ...... 1226 Torres, Lez Trujillo ...... 1118, 1118 Verstraeten, Julie ...... 755 Stoner, Jennifer L ...... 1183 Torres, Patricia ...... 1230 Vial, Céline ...... 316 Stoop, Jan ...... 1144 Touré-Tillery, Rima ...... 1124 Vicdan, Handan ...... 758 Storch, Julia ...... 715 Tournesac, Yann ...... 488 Vieites, Yan ...... 760 Strehlau, Suzane ...... 1186 Tran, Chi ...... 612 Vijayalakshmi, Akshaya . . . .762, 764, 1245 Strehlau, Vivian Iara ...... 1186 Tran, Huy Quoc ...... 1231 Villarroel, Francisco ...... 322 Streicher, Mathias Clemens ...... 318 Trendel, Olivier ...... 451 Vincent, Racheal Louis ...... 766 Strizhakova, Yuliya ...... 717 Trespeuch, Leo ...... 160 Visconti, Luca M...... 1190 Su, Sarena ...... 1043 Tripathi, Sanjeev ...... 478, 738 Vocino, Andrea ...... 1202 Su, Steve ...... 851 Trott, Sangeeta ...... 1240 Vock, Marlene ...... 1232 Su, Xiaolei ...... 1201 Trotzke, Patrick ...... 1224 Vohs, Kathleen ...... 1182, 1197 Suarez, Maribel Carvalho ...... 130 Trudel, Remi ...... 728, 778, 913, 1250 Volles, Barbara Kobuszewski ...... 1232 Šubrtová, Kristina ...... 720 Trupia, Maria Giulia ...... 740 von Janda, Sergej ...... 218 Sullivan, Nicolette ...... 1227, 1242 Tsai, Claire ...... 907 Vosgerau, Joachim . 596, 886, 1144, 1170, 1170 Sumter, Sindy Resita ...... 1215 Tseng, Wei-Chih ...... 1209 Voyer, Benjamin G...... 1222 Sun, Chengyao ...... 1227 Tsiros, Michael ...... 806 Sundar, Aparna ...... 598, 722 Tu, Yanping ...... 742, 811, 1175 Sung, Billy ...... 1222 Tuk, Mirjam ...... 744 W Sung, Yeonjin ...... 471, 1227 Tully, Stephanie M...... 1064 Suri, Anshu ...... 1227 Turner, Broderick ...... 1221 Waisman, Rory ...... 1232 Suri, Rajneesh ...... 1241, 1241 Turner, Broderick Lee ...... 746, 1250 Waisman, Rory M...... 943, 1139 Sussman, Abigail ...... 1091, 1250 Tusche, Anita ...... 891 Wakeman, Wiley ...... 508 Sussman, Abigail B...... 925, 925 Tveleneva, Arina ...... 1231 Walker, Jennifer ...... 1228 Sutil, Bruno ...... 1227 Wallendorf, Melanie ...... 1248 Suurmets, Seidi ...... 1074 Walsh, Darlene ...... 1232 Syrjälä, Henna ...... 97 U Walsh, Gianfranco ...... 1186 Szocs, Courtney ...... 1228, 1235 Wan, Echo Wen ...... 512, 1242 Ubel, Peter A...... 1129 Wan, Fang ...... 1199, 1206, 1238 Uduehi, Esther ...... 1231 Wan, Jing ...... 715 T Ülkümen, Gülden ...... 943 Wang, Ze ...... 786 Ulqinaku, Aulona ...... 749 Wang, Di ...... 801 Takhar, Amandeep ...... 143, 1190, 1228 Ulu, Sevincgul ...... 194 Wang, Haizhong ...... 512 Talebi, Arash ...... 724, 726, 1228 Ulver, Sofia ...... 753 Wang, Haoyuan ...... 1237 Talukdar, Nabanita ...... 1228, 1240 Urbina, Susan Danissa Calderon . . . . .1205 Wang, Jiaqian (Jane) ...... 771 Tang, Qing ...... 1228 Urminsky, Oleg ...... 949, 1064, 1239 Wang, Jie ...... 1229, 1233, 1234 1262 / Author Index

Wang, Jingjing ...... 1211 Wu, Ruomeng ...... 786, 1150 Yu, Tianjiao ...... 1240 Wang, Juan ...... 1202 Wu, Shuang ...... 469 Yu, Yiqi ...... 771, 805 Wang, Liangyan ...... 1233 Wu, Sihan ...... 1236 Yu, Zhihao ...... 742 Wang, Lili ...... 773 Wu, Wenqing ...... 1238 Yu, Zhining ...... 1233 Wang, Qian (Janice) ...... 1220 Wu, Xiu ...... 1236 Yuan, Hong ...... 612 Wang, Rebecca Jen-Hui ...... 1017 Wu, Yuechen ...... 1134, 1144 Yuan, Jingting ...... 811 Wang, Tingting ...... 1207, 1208 Wyss, Philipp ...... 1223 Yucel-Aybat, Ozge ...... 1240 Wang, Xiang ...... 769, 1175 Yuen, Vincentia ...... 806 Wang, Xiaoran ...... 1234 Yuksel, Mujde ...... 1211 Wang, Xin ...... 1233, 1237 X Yuksel, Ulku ...... 1240 Wang, Xin (Shane) ...... 825 Xia, Lan ...... 1236, 1236 Wang, Xinyi ...... 1231 Xiao, Chunqu ...... 1233, 1237 Wang, Xue ...... 1188, 1197, 1234, 1239 Xiao, Na ...... 234, 1236 Z Wang, Yajin ...... 1000 Xie, Chunya ...... 1238 Zallot, Camilla ...... 902 Wang, Yan ...... 1233, 1234, 1234 Xie, Chunyan ...... 1237 Zambaldi, Felipe ...... 1212 Wang, Yansu ...... 1234 Xie, Guang-Xin ...... 1200, 1247 Zane, Daniel M...... 1134 Wang, Yijie ...... 1200 Xie, Vincent ...... 954 Zanette, Maria Carolina ...... 168 Wang, Yun-Hui ...... 1210 Xie, Vivian (Jieru) ...... 788, 1237 Zanini, Celso ...... 1245 Wang, Yusu ...... 1233 Xie, Xuan ...... 1237, 1237 Zayer, Linda Tuncay ...... 880 Wang-Ly, Nathan ...... 925 Xu, Fei (Katie) ...... 1238 Zeelenberg, Marcel ...... 1210 Ward, Adrian ...... 1032 Xu, Haiyue (Felix) ...... 792 Zeng, Fue ...... 1189 Ward, Morgan ...... 1000, 1150 xu, jialiang ...... 1238 Zengxiang, Chen ...... 1208 Ward, Morgan K ...... 1006 Xu, Jing ...... 811 Zhang, C. Yiwei ...... 925 Warlop, Luk . . . 291, 930, 1205, 1226, 1231 Xu, Lan ...... 790 Zhang, Charles ...... 1247 Warr, Richard ...... 1194 Xu, Mengran ...... 1238 Zhang, Charles Y...... 810 Warren, Caleb ...... 930, 1048 Xu, Minzhe ...... 742, 794, 1160 Zhang, Jintao ...... 1241, 1241 Warren, Nooshin ...... 1248 Xu, Xiaobing ...... 1224 Zhang, Ke ...... 1198 Watson, Jared ...... 896, 1086, 1134 Xu, Yixiang ...... 868 Zhang, Kuangjie ...... 1228 Watson, Jared Joseph ...... 1058 Xue, Haibo ...... 796 Zhang, Ruby ...... 1229 Weber, Bernd ...... 891 Xue, Sherrie Ying Ying Ying ...... 1124 Zhang, Xiadan ...... 1234 Wegerer, Philipp K...... 164 zhang, Xiaodan ...... 440, 1199 Wei, Sarah ...... 1242 Zhang, Xiaodan ...... 1233 Wei, Xinliang ...... 1188 Y Zhang, Xuan ...... 773 Weihrauch, Andrea ...... 776 Zhang, Ying ...... 805 Weijo, Henri ...... 1223 Yalcin, Gizem . . . . 1155, 1165, 1181, 1248 Yalkin, Cagri ...... 797 Zhang, Yinlong ...... 1197 Weijters, Bert ...... 374 Zhang, Yuli ...... 808 Weingarten, Evan ...... 1155 Yamim, Amanda Pruski ...... 1238 Yan, Dengfeng ...... 352 Zhao, Min ...... 907, 1250 Wen, Na ...... 1235, 1235 Zhao, Xin ...... 796 Wen, Yingting ...... 1118, 1235 Yan, Jun ...... 1187, 1206 Yan, Li ...... 1239 Zhao, Ye ...... 755 Wenzel, Claudia ...... 1235 Zhao, Yijun ...... 1241, 1241 Werkman, Amber ...... 1235 Yang, Adelle X...... 1160, 1165 Yang, Cathy ...... 831 Zheng, Wanyi ...... 1242 Wertenbroch, Klaus ...... 1184, 1248 Zheng, Wenxue ...... 1241, 1241 Wheeler, Christian ...... 1124 Yang, Chun-Ming ...... 1239 Yang, Haiyang ...... 1249 Zheng, Xiaoying ...... 811 Wheeler, S. Christian ...... 994 Zheng, Yanmei ...... 769 Whelan, Brian ...... 314 Yang, Morgan X...... 1193, 1240 Yang, Shiyu ...... 799 Zhong, Jing Yang ...... 391 Whillans, Ashley V...... 651, 1241 Zhou, Li ...... 816 White, Katherine ...... 1225 Yang, Shuhan ...... 874 Yang, Yang ...... 703, 1160 Zhou, Lingrui ...... 930, 1150, 1242 White, Tiffany ...... 1247 Zhou, Xiaozhou (Zoe) ...... 814 Whitley, Sarah ...... 541, 778 Yang, Zhiyong ...... 1215 Yanit, Mehmet ...... 1206 Zhu, Donghong ...... 1191 Whitson, Jennifer ...... 453 Zhu, Guowei ...... 816 Wiart, Lucie ...... 782 Yao, Jun ...... 198, 801 Yap, Andy J ...... 275 Zhu, Hong ...... 1233, 1237 Widney, Jennifer ...... 370 Zhu, John Jianjun ...... 502 Wiener, Hillary ...... 1150 Yau, Amy ...... 1239 Ye, Hongjun ...... 1241, 1241 Zhu, Meng ...... 1144, 1170, 1249 Wieseke, Jan ...... 379 Zhu, Shuqi ...... 1242 Wieser, Verena E...... 622 Yeomans, Mike ...... 1017 Yi, Jiwon ...... 1239 Zhu, Yimin ...... 1209, 1242 Wilcox, Keith ...... 1200 Zhu, Ying ...... 1211 Williams, Jerome ...... 1251 Yi, John ...... 930 Yi, Youjae ...... 1218 Zimmermann, Laura ...... 820 Williams, Lawrence E ...... 1006, 1208 Zimmermann-Janssen, Vita Eva Maria . . 1242, Williamson, Sara ...... 1228, 1235 Yin, Eden ...... 1048 Yin, Siyuan ...... 803 1243 Wilson, Hugh N...... 582 Zor, Ozum ...... 822, 1243 Winterich, Karen Page . . 385, 792, 963, 1197 Yin, Yidan ...... 509 Yoo, Hyesung ...... 389 Zou, Zhimin ...... 1209 Wirtz, Jochen ...... 1194 Zouaoui, Hela ...... 173 Woermann, Niklas ...... 31, 622, 1216 Yoon, Carolyn ...... 891, 1011 Yoon, Nari ...... 1080 Zwebner, Yonat ...... 1096, 1175, 1223 Wong, Nancy ...... 1250 Zwick, Ruth ...... 544 Wood, Stacy ...... 1249 Yoon, Y. Rin ...... 1239 Woolley, Kaitlin 365, 504, 784, 994, 1175, 1239 Youn, Nara ...... 1239 Wu, Alisa ...... 1043 Yu, Irina Y...... 1193, 1240 Wu, Kaiyang ...... 1236 Yu, Jiaqi ...... 1239 Yu, Shubin Lance ...... 1228, 1240 ASSOCIATION FOR CONSUMER RESEARCH

$119.00 ISBN 978-0-915552-81-8 59900>

9 780915 552818