Pride and Regulatory Behavior: the Influence of Appraisal Information and Self-Regulatory Goals

Pride and Regulatory Behavior: The Influence of Appraisal Information and Self-Regulatory Goals

ANTHONY SALERNO JULIANO LARAN CHRIS JANISZEWSKI

Prior research has demonstrated that while people experience pride after an accomplishment, its experience can both encourage and discourage subsequent regulatory behavior. The current research shows that this seemingly contradictory inﬂuence can be accounted for by considering the information about the event that generated the pride (i.e., appraisal information) and whether or not a self-regulatory goal is active. In the absence of a self-regulatory goal, appraisal information concerning the locus of control (e.g., “I contributed to this achievement; it was not luck”) can be used to make positive inferences about one’s self-concept that reinforce further self-regulation. In the presence of a self-regulatory goal, appraisal information concerning personal agency (e.g., “This was my achievement, not someone else’s achievement”) can be used to make positive inferences about one’s progress toward self-regulation that, in turn, licenses indulgence. Thus appraisal information and the presence of self-regulatory goals is critical to how a self-conscious emotion like pride inﬂuences subsequent behavior. Implications and potential extensions in the areas of self-conscious emotions, goals, and self-regulation are discussed.

Keywords: pride, self-conscious emotions, appraisal information, goals, self- regulation

ride is a positive self-conscious emotion that results Tracy and Robins 2007). Self-conscious emotions (e.g., P from personal achievements or the achievements of pride, embarrassment, empathy, envy, jealousy, guilt, close others (Huang, Dong, and Mukhopadhyay 2014; shame) differ from basic emotions in that their elicitation is accompanied by complex cognitions about the self (Han, Anthony Salerno ([email protected]) is assistant professor of Duhachek, and Agrawal 2014; Tangney and Tracy 2012). marketing, Carl H. Lindner College of Business, University of Cincinnati, These complex cognitions are self-evaluative and self- Cincinnati, OH, 45221-0145. Juliano Laran ([email protected]) is profes- reflective (Campos et al. 1994; Lazarus 1991). Moreover, sor of marketing, University of Miami, Coral Gables, FL, 33124-6524. these complex cognitions can influence a host of subse- Chris Janiszewski ([email protected]) is the Russell quent behaviors including product choice (Huang et al. Berrie Eminent Scholar Chair and professor of marketing, Warrington College of Business Administration, University of Florida, Gainesville, 2014), creative expression (Damian and Robins 2012), and FL 32611-7155. This article is based on the second essay of the first au- self-regulation (Wilcox, Kramer, and Sen 2011). thor’s doctoral dissertation. The authors thank Keith Wilcox, Claudia Self-conscious emotions influence a variety of behav- Townsend, and Keri Kettle for their helpful comments on previous drafts iors, but in inconsistent ways. For example, Williams and of this article. The authors also thank the editor, associate editor, and three DeSteno (2008) found that pride increased regulatory be- anonymous reviewers for their guidance. havior, whereas Winterich and Haws (2011) found that Mary Frances Luce and Vicki Morwitz served as editors, and Andrea pride reduced regulatory behavior. Envy, an emotion eli- Morales served as associate editor for this article. cited from making unfavorable social comparisons with others, can increase (Van de Ven, Zeelenberg, and Pieters Advance Access publication July 11, 2015 2011) and decrease (Crusius and Mussweiler 2012) task

VC The Author 2015. Published by Oxford University Press on behalf of Journal of Consumer Research, Inc. All rights reserved. For permissions, please e-mail: [email protected] Vol. 42 2015 DOI: 10.1093/jcr/ucv037

499 500 JOURNAL OF CONSUMER RESEARCH persistence. Shame, an emotion elicited from a negative novelty (Roseman, Wiest, and Swartz 1994; Smith and self-evaluation, can increase (De Hooge, Breugelmans, and Ellsworth 1985; So et al. 2015). For instance, a negative Zeelenberg 2008) and decrease (Tangney and Fischer outcome that includes an appraisal of certainty generates 1995) prosocial behaviors. Guilt, an emotion elicited from anger. However, if a negative outcome includes an ap- the negative self-reflection of a transgression, can increase praisal of uncertainty, fear is experienced (Lerner and (Goldsmith, Cho, and Dhar 2012) and decrease (Zemack- Keltner 2000). Similarly, a positive outcome that includes Rugar, Bettman, and Fitzsimons 2007) hedonic consump- an appraisal of novelty (familiarity) will elicit surprise tion. Thus there is ambiguity about how self-conscious (calm) (Ellsworth and Smith 1988). Thus appraisal theory emotions exert an influence on subsequent behaviors. emphasizes the importance of the emotion-eliciting situa- This research offers a more nuanced understanding of tion (henceforth appraisal information) for predicting how self-conscious emotions influence behavior. We the felt emotion (henceforth emotional experience; see investigate this issue by assessing how the experience of figure 1, first two boxes, solid line). authentic pride (i.e., pride that evokes a feeling of accomplishment from successful behavior) influences regulatory behavior. We contend that pride can be used to make infer- The Emotion-as-Input Perspective ences about one’s self-concept (i.e., “Who I am”) or self- Although appraisal theory is useful for understanding regulatory goal progress (i.e., “How much I have done”). the source of a felt emotion, the perspective is less infor- Self-concept inferences are a default response, whereas mative for understanding the behavioral consequences of goal progress inferences occur when a self-regulatory goal an emotion. Addressing this possibility, the emotion-as- is active (Laran and Janiszewski 2009). Importantly, the input perspective looks at how contextual factors (e.g., a extent of these inferences should depend on the nature of person’s goal state) impact the accessibility of inferences the appraisal information that led to the pride. For example, associated with an emotion, so as to influence behavior pride that is elicited from an internal locus of control (e.g., (Martin and Davies 1998). For instance, Martin et al. “I am responsible”) should be more likely to generate a (1993) looked at how happiness versus sadness influenced self-concept inference (e.g., “I am disciplined”) than pride the desire to gather the information needed to form an im- elicited from an external locus of control (e.g., “I got pression. They found that the effect of each emotion de- lucky”). Similarly, pride that is a consequence of personal pended on active goals. When a thoroughness goal was agency (“I accomplished that”) should be more likely to salient, happiness (sadness) led people to infer that they generate a goal progress inference than pride that is a had been sufficiently (insufficiently) thorough and to ex- consequence of other agency (e.g., “My mom accom- pend less (more) effort. However, when an enjoyment goal plished that”). was salient, happiness (sadness) led people to infer that This article is organized as follows. We first provide an they were having more (less) fun and to expend more (less) overview of appraisal-tendency theory. Next we discuss effort. Thus contextual factors can determine the inferences the emotion-as-input perspective and its claim that the in- people make from an emotion, which guide subsequent be- fluence of emotions on behavior depends on contextual haviors (see figure 1, solid lines moving right from emo- factors. Then we use the functionalist perspective of emotional experience). tions to predict when pride will result in either self-concept or goal progress inferences that, in turn, will determine regulatory behavior. In four studies, we find that the type of in- Functionalist Perspective of Emotion ference that results from pride depends on the appraisal The functionalist perspective builds on the appraisal-ten- information that elicited the pride and whether a self- dency and emotion-as-input perspectives. The functionalist regulatory goal is active or inactive. Finally, we consider perspective assumes that emotions influence behavior as the broader implications of our framework for studying the a function of a person’s ongoing interaction with the effects of self-conscious emotions on behavior. environment (Crusius and Lange 2014; Frijda 2005). Functionalists assume emotions exert an influence on subsequent behavior owing to two properties. First, emotions THEORETICAL BACKGROUND are composed of neural programs that facilitate adaptive Appraisal-Tendency Theory responses to the environment (DeSteno, Valdesolo, and Bartlett 2006; Griskevicius, Shiota, and Nowlis 2010). An Appraisal-tendency theory is a foundational framework emotion includes (1) antecedent appraisal information con- in the emotions literature (Ellsworth and Smith 1988; cerning the source of the emotion, (2) the emotional expe- Lerner and Keltner 2000; Roseman 1991). Classic ap- rience, (3) schemata for interpreting the meaning of the praisal theories classify discrete emotions based on differ- emotion (i.e., emotion-based inferences), and (4) response ences in appraisal dimensions such as pleasantness, schemata that dictate behavior given the emotion (Campos arousal, certainty, effort, locus of control, agency, and et al. 1994; Griskevicius et al. 2009; Witherington and SALERNO ET AL. 501

FIGURE 1

CONCEPTUAL MODEL OF THE INFLUENCE OF PRIDE ON BEHAVIOR

Contextual Factors (e.g., goal state)

Emotion- Appraisal Emotional Response Based Behavior Information Experience Schemata Inferences

NOTE.—The figure represents different perspectives on how emotions influence behavior. Appraisal-tendency theory investigates the influence of appraisal information on emotional experience. The emotion-as-input perspective investigates how contextual factors determine the influence of emotional experience on emotion-based inferences and behavior, as determined by response schema. The functionalist perspective assumes contextual factors determine how appraisal information informs emotion-based inferences and dictates the response schema that will make use of these inferences (solid lines). This research extends the functionalist perspective by showing that contextual factors determine how information that accompanies the emotional experience, but does not define the emotional experience, informs emotion-based inferences and subsequently influences behavior (dashed lines).

Crichton 2007). Response schemata use appraisal informa- option) when it was not. Wilcox et al. concluded that in the tion, emotional experience, and emotion-based inferences presence of a regulatory goal, pride was taken as evidence to determine which behaviors are appropriate. Second, of accomplishment that licensed indulgence. In the absence emotions are dynamic systems that can be used to either of a relevant goal, pride made people self-aware, which en- initiate or modify goal-directed behavior (Frijda 2005; couraged the choice of the virtuous option. Keltner and Gross 1999). The accessibility of schemata that dictate the behavioral responses to an emotion depends Functionalist Perspective, Pride, and on contextual factors in the environment, such as a person’s goal state (Tooby and Cosmides 2008). Thus the Regulatory Behavior manner in which an emotion informs behavior is not only a The results of Huang et al. (2014) and Wilcox et al. function of the appraisal information that elicited the emo- (2011) illustrate how appraisal information can be used to tion, but also the context in which the emotion will influ- draw inferences that influence subsequent behaviors. Yet ence behavior (see figure 1, solid lines). their evidence aligns with the functionalist through an The functionalist perspective explains (1) how two dif- appraisal theory lens perspective, in that only appraisal in- ferent emotions can alter behavior, given the same re- formation that changes the emotional experience from one sponse schema, or (2) how two different response discrete emotion to another influences subsequent behavior schemata can alter behavior, given the same emotion. (see solid lines in figure 1). An alternative perspective, Huang et al. (2014) provide an example of the former. appraisal theory through a functionalist lens, assumes that Participants were asked to think about a situation in which incidental appraisal information (i.e., appraisal information they felt proud because of “who they are” or “what they that does not change the emotional experience) can also in- did.” Being proud of one’s self oriented people toward per- fluence subsequent behavior (see dashed lines in figure 1). sonal traits (i.e., hubristic pride), while being proud of That is, different sets of appraisal information can generate one’s actions oriented people toward accomplishment (i.e., the same emotion yet result in different emotion-based in- authentic pride). Consequently, participants who were ferences that, in turn, inform different response schema proud of who they were (what they did) preferred unique when guiding behavior. This perspective challenges the as- (common) products. Wilcox et al. (2011, study 3) provide sumption that only the appraisal information that defines an example of the latter. Participants were asked to write an emotion can inform emotion-based inferences, response about an accomplishment that made them proud. Then a schema, and behavior. health goal was primed, and participants were given a The appraisal theory through a functionalist lens per- choice between (a) French fries or a salad (goal-relevant spective can inform our understanding of the regulatory choice) or (b) a $25 gift certificate for entertainment or consequences of pride. Pride is a self-conscious emotion school supplies (goal-irrelevant choice). Participants pre- that results from personal achievements, or the achieve- ferred French fries (indulgent option) when the choice was ments of close others, that lead to a desirable outcome goal relevant, but the school supply certificate (virtuous (Roseman, Spindel, and Jose 1990; Roseman et al. 1994). 502 JOURNAL OF CONSUMER RESEARCH

Pride appraisal information can vary with respect to effort H2b: In the presence of a self-regulatory goal, an experience (i.e., high vs. low), locus of control (i.e., internal vs. exter- of pride containing internal locus (or external locus) of con- nal), and agency (i.e., personal vs. other), although vari- trol appraisal information will encourage inferences of goal ance along these dimensions often has little influence on progress, inform the goal-based response schema (i.e., the intensity of a pride experience. To illustrate, consider a license), and decrease regulatory behavior. student who is proud of earning an “A.” The “A” could be It should also be possible to manipulate appraisal infor- the result of little effort or extensive effort. Regardless of mation so that the consequences are only observed when the effort, the grade should lead to a feeling of pride (Tracy the goal-based response schema is available. To illustrate, and Robins 2007). Yet we argue that changes in the ap- consider appraisal information that varies in agency, which praisal information (little vs. extensive effort) could influ- is defined as the extent to which an event occurred for the ence regulatory behavior. For example, in the absence of a self (e.g., “I achieved something”) or for someone else self-regulatory goal (a contextual factor), high-effort pride (e.g., “My mom achieved something”) (Russell 1982). should encourage emotion-based inferences about self- Personal agency should be more likely to influence emo- discipline. These inferences should inform the default re- tion-based inferences about goal progress than other sponse schema and encourage regulatory behavior. In the agency. This is because self-conscious emotions generate presence of a self-regulatory goal (a contextual factor), self-evaluative thoughts that inform one’s sense of self or high-effort pride should encourage emotion-based infer- goals (Tangney and Tracy 2012). While the accomplish- ences about goal progress. These inferences should inform ments of a close other may lead to inferences about one’s the goal-based response schema and license indulgence. sense of self (i.e., who I am), the accomplishments of a H1a: In the absence of a self-regulatory goal, an experience close other should not lead to inferences about one’s goal of pride containing high effort (but not low effort) appraisal progress (i.e., what I have done). Hence only personal information will encourage inferences of self-discipline, in- agency should influence the goal-based response schema form the default response schema, and increase regulatory and increase indulgent behavior. Either type of agency behavior. should influence emotion-based inferences about self-dis- H1b: In the presence of a self-regulatory goal, an experience cipline, owing to inferences of shared value systems among of pride containing high effort (but not low effort) appraisal close others (Pinkus et al. 2008; Tracy and Robins 2004). information will encourage inferences of goal progress, in- H3a: In the absence of a self-regulatory goal, an experience form the goal-based response schema (i.e., license), and de- of pride containing personal agency (or other agency) crease regulatory behavior. appraisal information will encourage inferences of self- Consistent with the functionalist perspective, appraisal discipline, inform the default response schema, and increase information need not influence all response schemata in regulatory behavior. the same way. It should be possible to manipulate appraisal H3b: In the presence of a self-regulatory goal, an experience information so that the consequences of this manipulation of pride containing personal agency (but not other agency) are only observed when a certain response schema is avail- appraisal information will encourage inferences of goal able. With respect to the default response schema, consider progress, inform the goal-based response schema (i.e., appraisal information that varies in locus of control, de- license), and decrease regulatory behavior. fined as the extent to which a personal event is determined In summary, the influence of pride on regulatory by internal (e.g., “I had a great day”) or external (e.g., “My behavior depends on the appraisal information associated competitor had a bad day”) forces (Sheldon, Ryan, and with the pride, which will have a distinct impact when self- Reis 1996). An internal locus of control should be more regulatory goals are or are not active. We test these hypoth- likely to influence emotion-based inferences about self-dis- eses in four studies. cipline than an external locus of control; hence only an internal locus of control should increase regulatory behavior given a default response schema. Either source of pride STUDY 1 should encourage emotion-based inferences about accom- Study 1 investigated the influence of pride on regulatory plishment and goal progress (the effort was the same) and, behavior when a goal requiring self-regulation was active consequently, have a similar influence when a goal-based or not (hypotheses 1a and 1b). A self-regulatory goal was response schema is active (i.e., result in licensing and in- primed using a lexical decision task, which is a reliable creased indulgence). goal priming method (Bargh and Chartrand 2000; Laran H2a: In the absence of a self-regulatory goal, an experience and Salerno 2013). Pride was induced through the use of a of pride containing internal locus (but not external locus) of writing task, which is a common method of inducing self- control appraisal information will encourage inferences of conscious emotions (Griskevicius et al. 2010; Van de self-discipline, inform the default response schema, and in- Ven et al. 2011) including pride (Huang et al. 2014; crease regulatory behavior. Wilcox et al. 2011). Self-regulation was assessed through a SALERNO ET AL. 503 task in which participants made a number of choices in- deal of) effort. For this and all subsequent studies, pretests volving virtues and vices, which is a common way to mea- were used to verify that the manipulations successfully al- sure regulatory behavior (Fedorikhin and Patrick 2010; tered the critical pride appraisal information while main- Fishbach and Dhar 2005). taining similar levels of pride across conditions (see online Pride was induced by recollecting an accomplishment appendix A for specific details and results). (e.g., an “A” on a test) that required high or low effort. We The third study was purportedly related to learning about predicted that when no self-regulatory goal was active, par- people’s everyday decisions, but it was actually a measure ticipants experiencing high-effort pride would be more of regulatory behavior. Participants made 15 total deci- likely to identify self-regulation as an opportunity to rein- sions, such that five decisions were said to be made in the force the sense that one has self-discipline, and increase morning, five in the afternoon, and five in the evening (see regulatory behavior. In contrast, when a self-regulatory online appendix B for experimental manipulations and a goal was active, we anticipated that participants experienc- list of the dependent measures). The order of the decisions ing high-effort pride would perceive a greater sense of within each part of the day was randomly presented. For progress toward the goal, and decrease regulatory behavior. each decision, there were two randomly ordered alterna- When pride was elicited from a situation that required low tives. An example of a morning decision was, “What will effort, we did not expect that experience would be used to you do? Get an early start or sleep in?” An example of an make a self-concept or goal progress inference. afternoon decision was, “It is toward the end of the afternoon; what do you decide to do? Exercise or relax?” An example of an evening decision was, “You are eating a de- Method licious dinner and there is a half-serving left. What would Participants and Design. A total of 182 undergraduate you do? Save it for tomorrow or eat it now?” Therefore, in students (52% female, Mage ¼ 20.30) participated in ex- each decision, one of the options was a more virtuous be- change for course credit. Participants were randomly as- havior than the other. After making their choices, partici- signed to one of six conditions in a 2 (self-regulatory goal: pants answered a few demographic questions, were inactive vs. active) 3 (emotion: no-emotion, low effort debriefed, and thanked for their time. pride, high-effort pride) between-subjects design.

Procedure. Participants were told that there were three Results studies in the session, unrelated and pooled together for Regulatory Behavior. Regulatory behavior was com- convenience. The first study was ostensibly intended to de- puted by summing the total number of virtuous choices. termine people’s verbal capabilities, but truthfully it was a An analysis of variance (ANOVA) on the choice score re- lexical decision task that served as our goal manipulation. vealed an interaction between the goal activation and emo- Participants were told to focus their attention on a fixation tion factors (F(2, 176) ¼ 8.16, p < .01; see figure 2). In the point (X) on the computer screen that would be replaced inactive goal condition, participants made more virtuous by a letter string. They were told to press “nine” on the keyboard if the letter string was a word and “one” if it was not, and to respond as quickly as possible. After five prac- FIGURE 2 tice trials, seven target trials varied by condition. STUDY 1 RESULTS Participants in the active self-regulatory goal condition were shown words associated with being regulated (regu- 10 late, health, persistence, willpower, strive, control, virtue); 9 participants in the inactive self-regulatory goal condition 8.24 8.44 were shown neutral words (computer, flower, refrigerator, 8 7.89 notebook, picture, silk, next). Each participant saw the target words twice, along with five filler words, and 20 non- 7 6.60 6.89 6.53 words. The trial order was random. 6 The second study manipulated emotion. In the no-

emotion condition, participants were asked to write about a Virtuous Choices 5 typical day. In the pride conditions, participants were told that the purpose of the study was to learn more about life 4 events that made them feel proud. All participants were 3 told to think about a situation in which they were proud of Inactive Self-Regulatory Active Self-Regulatory receiving an “A” letter grade in an important class. In the Goal Goal low (high) effort pride condition, they were told to think about a time in which the “A” came with very little (a great No Emotion Pride: Low Effort Pride: High Effort 504 JOURNAL OF CONSUMER RESEARCH choices in the high-effort pride (M ¼ 8.24) than in the low Method effort pride (M ¼ 6.89; F(1, 176) ¼ 3.97, p < .05) and no- Participants and Design. Participants were 312 under- emotion (M ¼ 6.60; F(1, 176) ¼ 8.28, p < .01) conditions, graduate students (47% female, M ¼ 20.04) who partic- which did not differ from each other (F < 1). In the active age ipated in exchange for course credit. Participants were goal condition, participants made fewer virtuous choices in randomly assigned to one of six conditions in a 2 (self- the high-effort pride (M ¼ 6.53) than in the low effort pride regulatory goal: inactive vs. active) 3 (emotion: no- (M ¼ 8.44; F(1, 176) ¼ 7.83, p < .01) and no-emotion emotion, earned-pride, fortuitous-pride) between-subjects (M ¼ 7.89; F(1, 176) ¼ 5.06, p < .05) conditions, which did design. not differ from each other (F < 1). Making a self-regulatory goal active increased the number of virtuous choices in the no-emotion (F(1, 176) ¼ 4.58, p < .05) and low ef- Procedure. Participants were told that the session con- fort pride (F(1, 176) ¼ 5.26, p < .05) conditions, but it de- sisted of two unrelated studies. The first study purportedly creased the number of virtuous choices in the high-effort investigated cognitive processes associated with unscram- pride condition (F(1, 176) ¼ 7.40, p < .01). bling sentences, but in truth it served as the goal priming task (Srull and Wyer 1979). Participants were presented with 10 sets of five words and asked to form sentences by Discussion unscrambling each set of words (see online appendix C for Study 1 demonstrates that the influence of pride depends experimental manipulations and dependent measures). In on appraisal information and self-regulatory goal activa- the active self-regulatory goal condition, each sentence tion. High-effort pride increased (decreased) regulatory be- contained a word consistent with a health goal (healthy, havior when a self-regulatory goal was not (was) active, fat, weight, fit, slim, shape, discipline, willpower, exercise, but low effort pride did not have an influence on the workout). In the inactive self-regulatory goal condition, amount of regulatory behavior. These effects occurred these words were replaced with neutral words. After even though a pretest confirmed that low- and high-effort unscrambling the 10 sentences, participants were told that pride generated an equivalent level of pride. This implies they were finished with the first study. that the appraisal information that generated the pride in- The second study was a writing task that manipulated formed inferences about the self-concept or goal progress. emotion. The no-emotion condition was identical to study In studies 2 and 3, we provide direct process evidence for 1 (i.e., participants were asked to write about a typical how appraisal information is incorporated into inferences day). In the pride conditions, the instructions were similar about the self-concept or goal progress. to study 1. Participants were told that the researchers wanted to learn more about what makes people feel proud. In the earned-pride condition, participants were asked to STUDY 2 write about an accomplishment that they felt solely responsible for. In the fortuitous-pride condition, participants In study 2, we wanted to provide further evidence that were asked to write about an accomplishment that occurred the influence of pride on regulatory behavior was a conse- as a result of a lucky break. Examples of participant re- quence of the appraisal information that generated the sponses included getting a job as a result of a contact put- pride. Specifically, when appraisal information varied the ting in a good word, getting the opportunity to start on a locus of control (e.g., internal, “I was responsible” vs. ex- sports team as a result of an injury to another player, and ternal, “I got lucky”), pride should differentially influence winning a music competition because of an unexpected fa- inferences about one’s self-concept but uniformly influ- vorable evaluation from the judges. ence inferences about goal progress. That is, in the absence After writing their response, participants were told they of a self-regulatory goal, pride should encourage regulatory would be receiving a snack as appreciation for their partici- behavior only when the locus of control is internal (hypoth- pation. Participants were asked to make a choice between a esis 2a). In the presence of a self-regulatory goal, pride granola bar (healthy option) and M&M’s (indulgent should encourage indulgent behavior regardless of the lo- option). The snack each participant chose served as our cus of control (hypothesis 2b). measure of regulatory behavior. Participants also re- In study 2, we made a health goal active and measured sponded to potential mediating questions using 9-point regulatory behavior using a choice between a granola bar scales. Inferences about self-discipline being part of one’s and M&Ms. In addition, evidence for the hypothesized me- self-concept were measured by taking the average of three diating processes was assessed by asking participants how items (sample item: “Being a highly regimented person is much self-discipline they perceived themselves to have (an just part of who I am.” a ¼ .82). Inferences about self-regu- inference about the self-concept) and how much progress latory goal progress were measured by taking the average they had made toward eating in a healthy way (an infer- of three items (sample item: “How much progress do you ence about self-regulatory goal progress). currently feel you have made toward healthy eating?” SALERNO ET AL. 505 a ¼ .95). After responding to demographic questions, par- (M ¼ 5.59), and no-emotion conditions (M ¼ 5.61; F < 1). ticipants were debriefed and received their food choice. In addition, within the earned-pride condition, perceptions of self-discipline were higher in the inactive (as compared Results to active) goal condition (F(1, 306) ¼ 13.07, p < .01). An ANOVA on the perception of goal progress revealed Regulatory Behavior. A binary logistic regression re- an interaction between the goal activation and emotion fac- vealed an interaction between the goal activation and emo- tors (F(2, 306) ¼ 4.38, p ¼ .01). In the active goal condi- 2 tion factors (v (1) ¼ 4.98, p < .05; see figure 3). In the tion, perceptions of goal progress were higher in the inactive goal condition, participants were more likely to se- earned-pride (M ¼ 6.73; F(1, 306) ¼ 12.21, p < .01) and lect the granola bar in the earned-pride (76.9%) than in the fortuitous-pride (M ¼ 6.46; F(1, 306) ¼ 7.56, p < .01) con- 2 fortuitous-pride (39.2%; v (1) ¼ 15.05, p < .01) and no- ditions than in the no-emotion condition (M ¼ 5.30). In the 2 emotion (46.4%; v (1) ¼ 10.55, p < .01) conditions, which inactive goal condition, perceptions of goal progress were 2 did not vary (v (1) ¼ .57, not significant [NS]). In the ac- equivalent in the earned-pride (M ¼ 5.82), fortuitous-pride tive goal condition, participants were less likely to select (M ¼ 5.16), and no-emotion conditions (M ¼ 5.64; 2 the granola bar in the earned-pride (42.9%; v (1) ¼ 9.53, F(2, 306) ¼ 1.34, p > .25). In addition, within the earned- 2 p < .01) and fortuitous-pride (36.7%; v (1) ¼ 13.61, pride (F(1, 306) ¼ 4.66, p < .05) and fortuitous-pride p < .01) conditions than in the no-emotion condition (F(1, 306) ¼ 9.46, p < .01) conditions, perceptions of goal (72.7%). The earned-pride and fortuitous-pride conditions progress were higher in the active (as compared to inac- 2 did not vary (v (1) ¼ .38, NS). Looked at differently, par- tive) goal condition. ticipants in the active (as compared to inactive) goal condi- We conducted moderated mediation analyses to test the tion were more likely to select the granola bar in the no- prediction that, when no self-regulatory goal was active, 2 emotion condition (v (1) ¼ 7.96, p < .01), less likely to se- perceptions of self-discipline would mediate the effects of lect the granola bar in the earned-pride condition pride on self-regulation only in the earned-pride condition. 2 (v (1) ¼ 12.24, p < .01), and as likely to select the granola When a self-regulatory goal was active, perceptions of 2 bar in the fortuitous-pride condition (v (1) ¼ .07, NS). goal progress would mediate the effects of pride on self- regulation in both pride conditions. We tested this predic- Mediation Analyses. An ANOVA on the self-concept tion using PROCESS model 8 (Hayes 2013). Emotion was measure revealed an interaction between the goal activation used as the predictor variable, self-regulatory goal as the and emotion factors (F(2, 306) ¼ 3.17, p < .05). In the inac- moderating variable, perceptions of self-discipline and tive goal condition, perceptions of self-discipline were goal progress as the mediators, and food choice as the de- higher in the earned-pride (M ¼ 6.65) than in the fortuitous- pendent variable. Since our predictor variable was categor- pride (M ¼ 5.71; F(1, 306) ¼ 9.22, p < .01) and no-emotion ical with three levels, we constructed two dummy variables (M ¼ 5.86; F(1, 306) ¼ 5.51, p < .05) conditions. In the in order to discern the individual effects of the earned-pride active goal condition, perceptions of self-discipline did not versus fortuitous-pride condition relative to the no-emotion differ across the earned-pride (M ¼ 5.46), fortuitous-pride condition (refer to Hayes and Preacher 2014). In this and all subsequent mediation analyses, the 95% confidence in- FIGURE 3 tervals (CIs) used to generate each indirect effect were conducted at 5000 bootstrap samples. STUDY 2 RESULTS When no self-regulatory goal was active, we initially 90 compared the earned-pride to the no-emotion condition. The path from the emotion factor to regulatory behavior 80 76.9 72.7 through perceptions of self-discipline (indirect effect) was 70 significant (95% CI, .01–.48), while the path through perceptions of goal progress (95% CI, .63 to .37) was not 60 significant. Then we compared the fortuitous-pride to the 50 46.4 42.9 no-emotion condition. The paths from the emotion factor to 39.2 regulatory behavior through perceptions of self-discipline 40 36.7 (95% CI, .06 to .10) and goal progress (95% CI, .04 Granola Bar (%) Granola 30 to .45) were not significant. 20 When a self-regulatory goal was active, we compared the earned-pride to the no-emotion condition and the fortu- 10 itous-pride to the no-emotion condition. With respect to Inactive Self-Regulatory Active Self-Regulatory Goal Goal earned pride, the path from the emotion factor to regulatory behavior through perceptions of goal progress was signifi- No Emotion Pride: Earned Pride: Fortuitous cant (95% CI, 1.58 to .27), while the path through 506 JOURNAL OF CONSUMER RESEARCH perceptions of self-discipline (95% CI, .25 to .10) was other’s accomplishment has a similar influence as one’s not. With respect to fortuitous-pride, the path from the own accomplishment in terms of positively affirming that emotion factor to regulatory behavior through perceptions one is disciplined and capable of achievement (Pinkus of goal progress was significant (95% CI, .68 to .01), et al. 2008). Therefore, either source of pride should signal while the path through perceptions of self-discipline (95% that one is a disciplined individual and reinforce regulatory CI, .05 to .08) was not. behavior (hypothesis 3a).

Discussion Method Study 2 found that when pride was a consequence of an Participants and Design. Participants were 368 un- earned accomplishment, its experience led to inferences dergraduate students (60% female, Mage ¼ 20.45) who par- about one’s self-concept, provided no self-regulatory goal ticipated in exchange for course credit. Participants were was active. When pride was experienced as a consequence randomly assigned to one of six conditions in a 2 of a fortuitous accomplishment, its experience did not al- (self-regulatory goal: inactive vs. active) 3 (emotion: no- low for inferences about one’s self-concept. This result emotion, self-pride, close-other pride) between-subjects suggests that pride influences subsequent regulatory behav- design. ior in accordance with the information that prompted its Procedure. Participants were told that the session con- generation. Study 2 also found that earned pride and fortu- sisted of two unrelated studies. The first study primed the itous pride encouraged similar perceptions of goal progress self-regulatory goal of health using a method adapted from when a self-regulatory goal was active. Consequently, the Fishbach and Labroo (2007). Participants were told that the feeling of pride licensed participants to indulge. The earn- researchers were interested in undergraduate students’ ac- ed/fortuitous pride had similar consequences for self-regu- tivities. Participants in the inactive (active) self-regulatory lation because each type of pride provided a sense of goal condition listed sources they could use to receive the accomplishment, even though the pride was from a differ- news (be healthy) (see online appendix D for the experi- ent source. This finding is particularly important because it mental manipulations and dependent measures). suggests that the information that generates an inference The second study consisted of a modified version of the about the self-concept is not the same as the information writing task from study 2. Participants in the no-emotion con- that informs an inference about goal progress. If the infor- dition were asked to write about a typical day. Participants in mation was the same, the pride manipulation would have the self-pride (close-other pride) condition described a per- behaved consistently in the active and inactive goal sonal (parent’s) accomplishment that made them feel proud. conditions. The remainder of the procedure was the same as study 2 (i.e., a choice between two snacks, mediation measures, demo- STUDY 3 graphic measures), except for the following. First, the snack choice participants made was between a bag of baby carrots Study 2 altered the source of pride to show how ap- (healthy option) and a pack of Oreos (indulgent option). praisal information could differentially influence infer- Second, single-item measures were used for the critical medi- ences about the self-concept but uniformly influence ators. Self-discipline was measured using “Being a highly inferences about goal progress. Study 3 tested the comple- regimented person is just part of who I am.” (1 ¼ Strongly mentary result: how appraisal information can differen- disagree, 9 ¼ Strongly agree). Goal progress was measured tially influence inferences about goal progress but using “How much progress do you currently feel you have uniformly influence inferences about the self-concept. We made toward healthy eating?” (1 ¼ Very little, 9 ¼ Agreat explored this possibility by looking at the agency (personal deal). These changes are a reflection of study 3 being run ear- vs. other) appraisal information pertaining to the pride ex- lier in the research program. perience. We hypothesized that pride about one’s own behavior would be used as evidence of goal progress, but Results pride about a close other’s behavior would not (hypothesis 3b). People experiencing pride as a result of the achieve- Regulatory Behavior. A binary logistic regression re- ment of a close other should not use that success as evi- vealed an interaction between the goal activation and dence of their personal self-regulatory goal pursuit. emotion factors (v2(1) ¼ 3.90, p ¼ .05; see figure 4). In We also hypothesized that pride about a close other’s be- the inactive goal condition, participants were more likely havior would be relevant to regulatory behavior that de- to select baby carrots in the self-pride (73.2%; pends on the self-concept. Pride can be elicited from the v2(1) ¼ 11.63, p < .01) and close-other pride (67.2%; achievements of close others because the other’s accom- v2(1) ¼ 7.74, p < .01) conditions as compared to the no- plishment is taken as evidence of meeting one’s ideal self- emotion condition (43.2%). The self-pride (73.2%) and standards (Tracy and Robins 2004). As a result, a close close-other pride (67.2%) conditions did not vary SALERNO ET AL. 507

FIGURE 4 (M ¼ 6.62) than in the close-other pride (M ¼ 5.75; F(1, 362) ¼ 4.79, p < .05) and no-emotion (M ¼ 5.81; STUDY 3 RESULTS F(1, 362) ¼ 4.55, p < .05) conditions. In the inactive goal condition, perceptions of goal progress were equivalent in 90 the self-pride (M ¼ 5.73), close-other pride (6.18), and no- 80 emotion (M ¼ 6.04; F < 1) conditions. In addition, within 73.2 the self-pride condition, perceptions of goal progress were 70 67.2 64.9 62.9 higher in the active (as compared to inactive) goal condi- 60 tion (F(1, 362) ¼ 5.24, p < .05). 50 We conducted moderated mediation analyses (PROCESS 43.2 model 8; Hayes 2013) to test whether perceptions of self- 40 31.0 discipline would mediate the effects of pride on self- Baby Carrots (%) Carrots Baby 30 regulation in both pride conditions when no self-regulatory goal was active, but perceptions of goal progress would me- 20 diate the effects of pride on self-regulation only in the self- 10 pride condition when a self-regulatory goal was active. Inactive Self-Regulatory Active Self-Regulatory When no self-regulatory goal was active, the path from the Goal Goal self-pride contrast (self-pride vs. no emotion) to regulatory No Emotion Pride: Self Pride: Close Other behavior through perceptions of self-discipline was significant (95% CI, .02–.42), while the path through perceptions of goal progress (95% CI, .06 to .31) was not significant. The path from the close-other pride contrast (close-other (v2(1) .50, NS). In the active goal condition, partici- ¼ pride vs. no emotion) to regulatory behavior through percep- pants were less likely to select baby carrots in the self- tions of self-discipline was significant (95% CI, .03–.44), pride (31.0%) than in the close-other pride (64.9%; while the path through perceptions of goal progress (95% v2(1) ¼ 13.22, p < .01) and no-emotion (62.9%; CI, .06 to .20) was not significant. When a self-regulatory v2(1) 12.20, p < .01) conditions. The no-emotion condi- ¼ goal was active, the path from the self-pride contrast (self- tion (62.9%) and close-other pride (64.9%) conditions did pride vs. no emotion) to regulatory behavior through percep- not vary (v2(1) ¼ .05, NS). Looked at differently, partici- tions of goal progress was significant (95% CI, .45 to pants in the active (as compared to inactive) goal condi- .01), while the path through perceptions of self-discipline tion were more likely to select baby carrots in the no- (95% CI, .30 to .05) was not. The paths from the close- emotion condition (v2(1) 5.23, p < .05), less likely to ¼ other pride contrast (close-other-pride vs. no emotion) to select baby carrots in the self-pride condition regulatory behavior through perceptions of goal progress (v2(1) ¼ 20.30, p < .01), and equally likely to select baby (95% CI, .20 to .10) and self-discipline (95% CI, .10 to carrots in the other-pride condition (v2(1) .07, NS). ¼ .21) were not significant.

Mediation Analyses. An ANOVA on the perceptions of self-discipline revealed an interaction between goal activa- Discussion tion and emotion (F(2, 362) ¼ 3.01, p ¼ .05). In the inac- Study 3 provides additional evidence that inferences tive goal condition, perceptions of self-discipline were about the self-concept depend on different appraisal infor- higher in the self-pride (M ¼ 6.41; F(1, 362) ¼ 5.64, mation compared to inferences about goal progress. When p < .05) and close-other-pride (M ¼ 6.59; F(1, 362) ¼ 9.42, a self-regulatory goal was active, pride derived from a p < .01) conditions than in the no-emotion condition personal accomplishment was interpreted as evidence of (M ¼ 5.58). In the active goal condition, perceptions self-regulatory goal progress, whereas pride derived from a of self-discipline were equivalent in the self-pride parent’s accomplishment was not. Self-pride and (M ¼ 5.36), close-other pride (M ¼ 5.79), and no-emotion close-other pride encouraged similar perceptions of self- conditions (M ¼ 5.68; F < 1). In addition, within the self- discipline, provided a self-regulatory goal was not active, pride (F(1, 362) ¼ 8.20, p < .01) and other-pride which consequently encouraged self-regulation. Similar to (F(1, 362) ¼ 4.95, p < .05) conditions, perceptions of self- study 2, the goal activation by agency interaction suggests discipline were higher in the inactive (compared to active) that the appraisal information informing the inference goal condition. about the self-concept is not the same as the information An ANOVA on the perceptions of goal progress also re- that informs the inference about goal progress. If the ap- vealed an interaction between goal activation and emotion praisal information was the same, each pride manipulation (F(2, 362) ¼ 3.43, p < .05). In the active goal condition, would have led to consistent patterns of self-regulation in perceptions of goal progress were higher in the self-pride both goal conditions. 508 JOURNAL OF CONSUMER RESEARCH

STUDY 4 about how people make decisions in general. In actuality, the study was used to reduce participants’ willingness to Studies 1 through 3 have shown that appraisal informa- rely on their feelings when making judgments (see Avnet tion can influence a perception of self-discipline, when a et al. 2012). The instructions began with an overview of self-regulatory goal is not active, or goal progress, when a how people use logical reasoning versus feelings in their self-regulatory goal is active. Yet the fact that the studies decision making. Afterward, participants were asked to manipulated appraisal information while holding the level come up with 10 different situations in which they “trusted of experienced pride constant raises the possibility that their feelings to make a judgment or a decision and it was pride was not a necessary condition (see figure 1, dashed the right thing to do.” Previous research has shown that line moving up from emotional experience). To investigate people find it difficult to generate 10 instances of trusting this issue, we drew from research showing that people’s their feelings, and this subjective lack of success leads trust in their emotions influences their emotional experi- them to become more distrustful of their feelings (Avnet ence (Avnet, Pham, and Stephen 2012; Epstein et al. et al. 2012). Participants in the control condition did not 1996). Specifically, we had participants either perform a complete this task (see online appendix E for experimental task that made them distrust their emotions (low trust in manipulations and dependent measures). emotions condition) or not (trust in emotions, henceforth Participants then completed the same goal priming task the control condition), and observed whether this impacted from study 3, with the exception that the self-regulatory subsequent self-regulation. If the emotional experience of goal was changed to saving money. Participants in the in- pride is not necessary for appraisal information to guide active (active) self-regulatory goal condition listed sources self-regulation, people made to distrust their pride may fo- they could use to receive the news (save money). cus less on the feeling and more heavily on its antecedents The next study consisted of the writing task from study 3. (i.e., appraisal information). If this is the case, people made Participants in the no-emotion condition were asked to write to distrust their pride should still show the previous pattern about a typical day. Participants in the pride condition wrote of increased (decreased) regulatory behavior in the absence about a personal accomplishment that has always made (presence) of a self-regulatory goal. However, if the emo- them feel proud. tional experience of pride is necessary for appraisal infor- In the final study, participants read that the researchers mation to inform self-regulation, people made to distrust were interested in students’ budgeting habits. Participants their pride may deem its feeling and its antecedents as non- were asked to indicate the likelihood of engaging in eight diagnostic to behavior. If this is the case, self-regulation budgeting methods for the upcoming month, derived from should not be guided by pride and should instead vary as a Bayuk, Janiszewski, and LeBoeuf (2010). These items function of self-regulatory goal activation. were then averaged and served as our dependent measure To provide further support for our assertion that the of regulatory behavior. The eight items (a ¼ .73) used emotional experience of pride is critical to influencing 9-point scales (1 ¼ Not at all likely to use this method, self-regulation, we also measured people’s general ten- 9 ¼ Very likely to use this method), and example items in- dency to trust their emotions, using the Trust in Feelings cluded “Cook at home more,” “Decrease how much you Scale (Avnet et al. 2012). Our interest in this scale was for purchase on impulse,” and “Combine errands in one trip to participants who were not manipulated to distrust their save money on gas.” After indicating their saving inten- emotions (i.e., the control condition). Specifically, we tions, participants responded to three items about self- wanted to see whether people’s trust in their emotions in- discipline being part of one’s self-concept (sample item: teracted with the effect of pride on increasing (decreasing) “People would say I have exceptional self-discipline.” self-regulation. We predicted that the previously men- a ¼ .81) and goal progress (sample item: “How much prog- tioned self-concept (goal progress) processes would be ress do you currently feel you have made toward saving driven by people who trust their emotions. money?” a ¼ .91). Next, participants were asked to indicate their current mood (1 ¼ Negative mood, 9 ¼ Positive Method mood). This item was included to address whether the effects of pride on regulatory behavior are specific to pride Participants and Design. Participants were 257 under- or a more generalized positive mood state. We expected graduate students (57% female, M ¼ 20.38) who partici- age that pride would lead to an increase in positive mood but pated in exchange for course credit. Participants were that this increase in mood would not be responsible for any randomly assigned to one of eight conditions in a 2 (trust influence on self-regulation. in emotions: control vs. low) 2 (self-regulatory goal: in- After completing demographic questions and filler items, active vs. active) 2 (emotion: no-emotion vs. pride) be- participants completed the Trust in Feelings Scale (Avnet tween-subjects design. et al. 2012). The scale contains 12 questions assessing how Procedure. In the low trust in emotions condition, par- much people trust different types of information across vari- ticipants began by completing a study that was said to be ous decision contexts. Seven of the 12 items pertained to SALERNO ET AL. 509 how much people trust their feelings across decisions (e.g., Mactive ¼ 6.47; F(1, 249) ¼ 10.55, p < .01; see figure 5b). “When choosing a roommate, to what extent do you rely on, These results are consistent with our conceptualization. that is, believe and trust what your feelings tell you about When people were made to distrust the emotional experi- this roommate?” and “When deciding on a job, to what ex- ence of pride, self-regulation was not guided by emotion tent do you rely on, that is, believe and trust what your feel- but by the accessibility of a self-regulatory goal. ings tell you about this job?”). The remaining items asked Mediation Analysis. An ANOVA on the perceptions of how much people trusted other sources of information self-discipline item revealed a three-way interaction of the across decisions (e.g., To what extent do you rely on, that is, trust in emotions, goal activation, and emotion factors believe and trust what your parents say about renting a new (F(1, 249) ¼ 6.94, p < .01). In the control condition, there apartment?”). Each item used a 7-point scale (1 ¼ Not trust was an interaction between the goal activation and emotion at all, 7 ¼ Trust very much), and the seven feelings-based factors (F(1, 249) ¼ 7.19, p < .01). In the inactive goal con- items (a ¼ .77) were used to form an index of people’s gen- dition, perceptions of self-discipline were higher in the eral propensity to trust their emotions. pride (M ¼ 6.77) than in the no-emotion condition (M ¼ 5.52; F(1, 249) ¼ 9.64, p < .01). In the active goal Results condition, perceptions of self-discipline did not differ between the pride (M ¼ 5.54) and no-emotion conditions Regulatory Behavior. An ANOVA revealed an interac- (M ¼ 5.85; F < 1). In addition, within the pride condition, tion of the trust in emotions, goal activation, and emotion perceptions of self-discipline were higher in the inactive factors (F(1, 249) ¼ 5.88, p ¼ .01). In the control condition, (as compared to active) goal condition (F(1, 249) ¼ 8.11, the analysis revealed an interaction between the goal acti- p < .01). However, as predicted, in the low trust in vation and emotion factors (F(1, 249) ¼ 11.91, p < .01; see emotions condition, the interaction between the goal figure 5a). In the inactive goal condition, participants were activation and emotion factors was no longer significant more willing to save in the pride (M ¼ 6.66) than in the no- (F(1, 249) ¼ 1.17, p > .25). emotion condition (M ¼ 5.85; F(1, 249) ¼ 7.01, p < .01). In An ANOVA on the goal progress perception item re- the active goal condition, participants were less willing to vealed a three-way interaction of the trust in emotions, goal save in the pride (M ¼ 5.71) than in the no-emotion condi- activation, and emotion factors (F(1, 249) ¼ 6.57, p ¼ .01). tion (M ¼ 6.42; F(1, 249) ¼ 5.59, p < .05). Participants in In the control condition, there was an interaction between the active (as compared to inactive) goal condition were the goal activation and emotion factors (F(1, 249) ¼ 7.12, more willing to save in the no-emotion condition (F(1, p < .01). In the active goal condition, perceptions of goal 249) ¼ 3.99, p < .05), but less likely to do so in the pride progress were higher in the pride (M ¼ 6.44) than in the no- condition (F(1, 249) ¼ 8.83, p < .01). emotion condition (M ¼ 5.33; F(1, 249) ¼ 6.75, p ¼ .01). In the low trust in emotions condition, the analysis re- In the inactive goal condition, perceptions of goal progress vealed a nonsignificant interaction (F < 1) and a main ef- were equivalent in the pride (M ¼ 5.43) and no-emotion con- fect of self-regulatory goal activation (Minactive ¼ 5.75, ditions (M ¼ 5.95; F(1, 249) ¼ 1.49, p > .20). In addition,

FIGURE 5

STUDY 4 RESULTS

A - Control B - Low Trust in Emotions 9 9 8 8 7 6.66 7 6.64 6.42 6.31 5.85 5.71 6 6 5.58 5.91 5 5 4 4 Saving Intentions Saving Saving Intentions Saving 3 3 2 2 1 1 Inactive Self-Regulatory Active Self-Regulatory Inactive Self-Regulatory Active Self-Regulatory Goal Goal Goal Goal

No Emotion Pride No Emotion Pride 510 JOURNAL OF CONSUMER RESEARCH within the pride condition, perceptions of goal progress self-regulation (b ¼1.40, p < .01). In the inactive goal were higher in the active (as compared to inactive) goal con- condition, there was a marginally significant interaction dition (F(1, 249) ¼ 4.92, p < .05). However, consistent with between the emotion and trust in emotions factors (b ¼ .65, our conceptualization, in the low trust in emotions condition, p ¼ .07). At 1 standard deviation (SD) below the mean of the interaction between the goal activation and emotion fac- trust in emotions (i.e., low trust in emotions), participants tors was no longer significant (F(1, 249) ¼ 1.02, p > .25). in the no-emotion and pride conditions did not differ in An ANOVA on the mood item only revealed a main ef- their level of self-regulation (b ¼ .16, p > .70). At 1 SD fect of the emotion factor (F(1, 249) ¼ 10.59, p < .01). above the mean of trust in emotions (i.e., high trust in emo- Consistent with research showing that pride is a positive tions), participants exerted higher levels of self-regulation affective state (Tracy and Robins 2007; Williams and in the pride than in the no-emotion condition (b ¼ 1.34, DeSteno 2008), participants in the pride condition felt p < .01). In the active goal condition, there was an interac- more positive (M ¼ 6.59) than participants in the no-emotion between the emotion and trust in emotions factors tion condition (M ¼ 5.86). (b ¼.75, p < .05). At 1 SD below the mean of trust in We predicted that the previously tested causal pathways emotions, participants in the no-emotion and pride condi- would be observed for participants in the control condition tions did not differ in their level of self-regulation but not for participants made to distrust their emotions. We (b ¼.03, p > .70). At 1 SD above the mean of trust in used a moderated mediation analysis using PROCESS emotions, participants exerted lower levels of self- model 12 (Hayes 2013), which allowed us to probe the regulation in the pride than in the no-emotion condition conditional indirect effects of three predictor variables. (b ¼1.38, p < .01). Emotion was used as the predictor variable, self-regulatory goal activation as the first moderating variable, trust in Discussion emotions as the second moderating variable, perceptions of self-discipline, goal progress, and mood as the mediators, Study 4 provides support for the claim that the emotional and saving intentions as the dependent variable. experience of pride is necessary for appraisal information In the control condition, when no self-regulatory goal to influence self-regulation. When people were made to dis- was active, the path from emotion to regulatory behavior trust their emotions, the effect of pride on self-regulation through perceptions of self-discipline was significant (95% was no longer observed. Instead, regulatory behavior was CI, .05–.39), while the paths through perceptions of goal guided by the presence or absence of a self-regulatory progress (95% CI, .02 to .23) and mood (95% CI, .04 goal. This is consistent with our claim that pride must be to .14) were not. However, when a self-regulatory goal was present for the self-concept and goal progress inferences to active, the path from emotion to regulatory behavior be made. In addition, we found that when there was an in- through goal progress perceptions was significant (95% fluence of the self-concept and goal progress inferences CI, .42 to .03), while the paths through perceptions of (i.e., the control condition), these effects were being driven self-discipline (95% CI: .21 to .07) and mood (95% CI, by people who indicated having a high level of trust in .02 to .11) were not. In the low trust in emotions condi- their emotions. These same effects were not found for tion, regardless of whether a self-regulatory goal was inac- those who indicated a low level of trust in their emotions. tive or active, the paths from emotion to regulatory These results demonstrate that if people do not rely on their emotions, the appraisal information that generated the behavior through self-discipline (95% CIinactive, .24 emotion is also perceived to be unreliable in guiding subse- to .03; 95% CIactive, .10 to .20), goal progress (95% quent behavior. CIinactive, .25 to .07; 95% CIactive, .08 to .21), and mood (95% CIinactive, .03 to .15; 95% CIactive, .02 to .16) were not significant. GENERAL DISCUSSION Ancillary Analysis. We conducted additional analyses Prior research has shown that the effect of pride on self- to confirm the importance of experiencing pride in self- regulation is not always easy to predict. In some cases pride regulation. The analysis focused on the control condition, can aid self-regulation (Patrick, Chun, and MacInnis 2009; with the self-regulatory goal activation and emotion factors Wilcox et al. 2011; Williams and DeSteno 2008). In other as between subjects, and trust in emotions as a measured cases, pride can impair self-regulation (Wilcox et al. 2011; variable. We conducted a moderation analysis using Winterich and Haws 2011). Clearly pride, like other self- PROCESS model 3 (Hayes 2013) that allows one to probe conscious emotions, is a complex emotion whose influence the conditional effect of the predictor variable (emotion) on behavior depends on other moderating factors. In our re- on the dependent variable (self-regulation) at different lev- search, we focus on authentic pride and identify two critical els of the categorical (self-regulatory goal activation) moderating factors. We show that the type of appraisal and continuous (trust in emotions) moderators. The information and the salience of a self-regulatory goal analysis yielded a significant three-way interaction on influence when pride aids versus impairs self-regulation. SALERNO ET AL. 511

Across four studies, we show that (1) pride resulting from individual becomes more prone to engage in rumination effort can lead to inferences of a disciplined self-concept or (dwelling on what they did wrong) that can subsequently of goal progress, depending on whether a self-regulatory undermine the ability to self-regulate (Magen and Gross goal is inactive or active, (2) altering appraisal information 2010). Yet when guilt is due to a self-regulatory transgres- regarding the locus of control (perceived agency) alters sion, people may respond positively to this failure by re- inferences about one’s self-concept (goal progress) and grouping and increasing their regulatory behavior influences self-regulation, and (3) the emotional experi- (Zemack-Rugar, Corus, and Brinberg 2012). Thus there are ence of pride must be perceived as diagnostic for inferences many appraisal dimensions that may be relevant to self- from appraisal information to occur and influence self- conscious emotions and self-regulation, and by studying regulation. these dimensions concurrently, researchers can better un- derstand these complex relationships. By considering the joint influence of appraisal informa- Theoretical Contributions tion and a person’s goal state, the current research demon- In a recent review of the current state of emotion re- strates that pride can be used to make an inference about search, So et al. (2015) present a framework for how to le- one’s self-concept or goal progress. Previous research has verage appraisal information in order to advance our shown that the self-concept is a malleable, dynamic system understanding of how emotions influence behavior. The that is sensitive to input from one’s current environment in- authors point out that traditional appraisal-based cluding one’s emotional state (Burke and Stets 2009; Isbell approaches to the study of emotion will typically take a et al. 2013; Wheeler, DeMarree, and Petty 2007). More re- single appraisal dimension (e.g., certainty) and then exam- cent advances have shown evidence of “emotion profiles,” ine how two emotions that vary on this dimension (e.g., where some emotions appear to be particularly diagnostic fear and happiness) influence behavior (e.g., tolerance for of certain aspects of a person’s self-concept (Coleman and risk). However, they point out that considerably less re- Williams 2015). For instance, Coleman and Williams search has been conducted that takes a specific emotion (2013) found that anger/disgust/sadness was perceived to and then examines how varying two or more appraisal di- be highly relevant to people when an athlete/environmen- mensions can lead to different behavioral responses. This talist/volunteer identity was active. Our research contrib- is surprising when you consider that most emotions are utes to this area by showing that pride evokes an emotion composed of combinations of appraisals (Cavanaugh et al. profile of being a disciplined individual. This in turn en- 2007). This observation is especially true in the case of courages identity-consistent self-regulation. It is also im- self-conscious emotions, where appraisal information is portant to note that our self-concept mechanism is distinct cognitively complex and interpreted by the individual from the effects one might expect on self-regulation due to through a self-evaluative lens (Roseman et al. 1994; a boost in self-esteem. While pride does boost one’s self- Tangney and Tracy 2012; Tracy and Robins 2004, 2007). esteem (McFerran, Aquino, and Tracy 2014; Tracy and Our research addresses this gap by examining how the ef- Robins 2007), this boost in self-esteem has no influence on fort, locus of control, and agency appraisal information of people’s subsequent self-regulation (Williams and DeSteno pride influence self-regulation. 2008), suggesting that our self-concept mechanism cannot Future research on self-conscious emotions might exam- be accounted for by enhanced self-esteem. ine other instances in which varying multiple sources of Our findings also contribute to the understanding of how appraisal information can have a differential effect on reg- emotions interact with goal states to influence behavior. ulatory behavior. For instance, guilt can be a consequence Previous research looking at the relationship between goals of appraisal information pertaining to inaction (i.e., guilt and emotions has primarily examined how the attainment about what the person did not do) or action (i.e., guilt about (nonattainment) of a goal can elicit positive (negative) what the person did do; Han et al. 2014). In the context of emotion (Carver and Scheier 1998; Laran 2010). However, self-regulation, it might be the case that when people expe- research examining the possibility that emotions can be rience guilt due to inaction, self-regulation is encouraged used to inform goal pursuit is still largely in its infancy out of a need to act and to make up for what they failed to (Salerno, Laran, and Janiszewski 2014). In this respect, the do previously. Yet when guilt is due to action, self- functionalist perspective is an ideal framework for studying regulation may be discouraged out of the concern that this issue because one of its central propositions is that further action is likely to bring about additional negative emotions can be used to initiate, but also modify, goal- outcomes (Lazarus 1991). Likewise, guilt can also be a directed behavior (Frijda 2005; Keltner and Gross 1999). consequence of appraisal information pertaining to whether the guilt was due to a moral transgression (e.g., cheating on a test) or a self-regulatory transgression (e.g., overeating Limitations and Future Research on the weekend; Goldsmith et al. 2012). It may be the case Our research is not without limitations. Self-regulation that when guilt is due to a moral transgression, the was conceptualized as the amount of restraint that a 512 JOURNAL OF CONSUMER RESEARCH consumer exhibits, but this is not always the case. Indeed, could ultimately be used to aid consumers in the pursuit of research has shown that an equally valid form of self-regu- long-term goals and well-being. lation might involve overriding restraint with self-reward (Kivetz and Simonson 2002). Future research may investigate how pride and other emotions interact with hyperopic DATA COLLECTION INFORMATION versus myopic consumers (Haws and Poynor 2008). For instance, hyperopic consumers normally have a difficult time The first author supervised the collection of data by re- disengaging from self-regulation. It may be the case that search assistants using participants from the University of when these people are made to feel pride, they have an eas- Cincinnati (studies 1 and 2), University of Miami (study ier time allowing for self-reward. Alternatively, it may be 3), and the University of Florida (study 4) between the the case that hyperopic consumers use their feeling of pride spring of 2014 and the spring of 2015. The first author was as a reason to continue exhibiting restraint. Future research primarily responsible for the data analysis with input from could investigate these possibilities. the second and third authors. Data were discussed through- Although our research focused specifically on authen- out the entire research program by all authors. tic pride, we believe our findings may help spark future research into how hubristic pride affects self- REFERENCES regulation. Unlike authentic pride, which results from a specific behavior, hubristic pride results from a positive Avnet, Tamar, Michel T. Pham, and Andrew T. Stephen (2012), “Consumers’ Trust in Feelings as Information,” Journal of attribution about one’s traits (Huang et al. 2014; Tracy Consumer Research, 39 (December), 720–35. et al. 2014; Williams and DeSteno 2008). Hubristic pride Bargh, John A. and Tanya L. Chartrand (2000), “The Mind in the has also been linked to increased luxury consumption Middle: A Practical Guide to Priming and Automaticity and status seeking (McFerran et al. 2014). This might Research,” in Handbook of Research Methods in Social and lead one to conclude that hubristic pride would always Personality Psychology, ed. Harry T. Reis and Charles M. prompt indulgence. However, based on our findings dem- Judd, New York: Cambridge, 253–85. onstrating the importance of appraisal information, we Bayuk, Julia Belyavsky, Chris Janiszewski, and Robyn A. LeBoeuf (2010), “Letting Good Opportunities Pass Us By: would argue that this may not always be the case. For in- Examining the Role of Mind-Set during Goal Pursuit,” stance, one could alter the perceived permanence of trait Journal of Consumer Research, 37 (December), 570–83. appraisal information by making it seem less enduring. Burke, Peter J. and Jan E. Stets (2009), Identity Theory, New This could be done by highlighting how the traits that York: Oxford. elicit hubristic pride do not last forever (e.g., beauty, Campos, Joseph J., Donna L. Mumme, Rosanne Kermoian, and strength); doing so may increase the importance of self- Rosemary G. Campos (1994), “A Functionalist Perspective on the Nature of Emotion,” Monographs of the Society for regulation. Research in Child Development, 59 (February), 284–303. Carver, Charles S. and Michael F. Scheier (1998), On the Self- Conclusion Regulation of Behavior, New York: Cambridge. Cavanaugh, Lisa A., James R. Bettman, Mary Frances Luce, and One of the most important aspects of consumers’ lives is John W. Payne (2007), “Appraising the Appraisal-Tendency their ability to engage in self-regulation. While prior re- Framework,” Journal of Consumer Psychology, 17 (July), search has established that emotions play an influential 169–73. Coleman, Nicole V. and Patti Williams (2013), “Feeling Like My role in self-regulation, this relationship is complicated by Self: Emotion Proﬁles and Social Identity,” Journal of the fact that many emotions can both help and hinder regu- Consumer Research, 40 (August), 203–22. latory behavior in different situations. While the current re- ———. (2015), “Looking for My Self: Identity-Driven Attention search focused on the bidirectional influence of pride on Allocation,” Journal of Consumer Psychology, 25 (July), 504–11. self-regulation, similar demonstrations have been shown Crusius, Jan and Jens Lange (2014), “What Catches the Envious for guilt (Giner-Sorolla 2001; Goldsmith et al. 2012), Eye? Attentional Biases within Malicious and Benign Envy,” shame (De Hooge et al. 2008; Duhachek, Agrawal, and Journal of Experimental Social Psychology, 55 (November), 1–11. Han 2012), and envy (Crusius and Mussweiler 2012; Van Crusius, Jan and Thomas Mussweiler (2012), “When People Want de Ven et al. 2011). In considering the role of both ap- What Others Have: The Impulsive Side of Envious Desire,” praisal information and a person’s goal state, the current Emotion, 12 (February), 142–53. research was able to provide greater insight into when Damian, Rodica I. and Richard W. Robins (2012), “The Link be- pride will aid or inhibit regulatory behavior, and the pro- tween Dispositional Pride and Creative Thinking Depends on cesses by which these influences occur. It is worth explor- Current Mood,” Journal of Research in Personality,46 (December), 765–69. ing whether appraisal information and a person’s goal state De Hooge, Ilona E., Seger M. Breugelmans, and Marcel could provide greater clarity to how other self-conscious Zeelenberg (2008), “Not So Ugly After All: When Shame emotions impact self-regulation. By investigating such Acts as a Commitment Device,” Journal of Personality and possibilities, a systematic framework may emerge that Social Psychology, 95 (October), 933–43. SALERNO ET AL. 513

DeSteno, David, Piercarlo Valdesolo, and Monica Y. Bartlett Determine Contrasting Effects of Incidental Pride on (2006), “Jealousy and the Threatened Self: Getting to the Uniqueness Seeking,” Journal of Consumer Research,41 Heart of the Green-Eyed Monster,” Journal of Personality (October), 697–712. and Social Psychology, 91 (October), 626–41. Isbell, Linda M., Joseph McCabe, Kathleen C. Burns, and Elicia Duhachek, Adam, Nidhi Agrawal, and Dahee Han (2012), “Guilt C. Lair (2013), “Who Am I? The Influence of Affect on the versus Shame: Coping, Fluency, and Framing in the Working Self-Concept,” Cognition and Emotion,27 Effectiveness of Responsible Drinking Messages,” Journal of (September), 1073–90. Marketing Research, 49 (December), 928–41. Keltner, Dacher and James J. Gross (1999), “Functional Accounts of Ellsworth, Phoebe C. and Craig A. Smith (1988), “Shades of Joy: Emotions,” Cognition and Emotion, 13 (September), 467–80. Patterns of Appraisal Differentiating Pleasant Emotions,” Kivetz, Ran and Itamar Simonson (2002), “Self Control for the Cognition and Emotion, 2 (December), 301–31. Righteous: Toward a Theory of Precommitment to Indulgence,” Epstein, Seymour, Rosemary Pacini, Veronika Denes-Raj, and Journal of Consumer Research, 29 (September), 199–217. Harriet Heier (1996), “Individual Differences in Intuitive- Laran, Juliano (2010), “The Influence of Information Processing Experiential and Analytical-Rational Thinking Styles,” Goal Pursuit on Post-Decision Affect and Behavioral Journal of Personality and Social Psychology, 71 (August), Intentions,” Journal of Personality and Social Psychology, 390–405. 98 (January), 16–28. Fedorikhin, Alexander and Vanessa M. Patrick (2010), “Positive Laran, Juliano and Chris Janiszewski (2009), “Behavioral Mood and Resistance to Temptation: The Interfering Consistency and Inconsistency in the Resolution of Goal Influence of Elevated Arousal,” Journal of Consumer Conflict,” Journal of Consumer Research, 35 (April), 967–84. Research, 37 (December), 698–711. Laran, Juliano and Anthony Salerno (2013), “Life-History Fishbach, Ayelet and Ravi Dhar (2005), “Goals as Excuses or Strategy, Food Choice, and Caloric Consumption,” Guides: The Liberating Effect of Perceived Goal Progress on Psychological Science, 24 (February), 167–73. Choice,” Journal of Consumer Research, 32 (December), Lazarus, Richard S. (1991), Emotion and Adaptation, New York: 370–77. Oxford. Fishbach, Ayelet and Aparna A. Labroo (2007), “Be Better or Be Lerner, Jennifer S. and Dacher Keltner (2000), “Beyond Valence: Merry: How Mood Affects Self-Control,” Journal of Toward a Model of Emotion-Specific Influences on Personality and Social Psychology, 93 (August), 158–73. Judgment and Choice,” Cognition and Emotion, 14 (July), Frijda, Nico H. (2005), “Emotion Experience,” Cognition & 473–93. Emotion, 19 (June), 473–97. Magen, Eran and James J. Gross (2010), “Getting Our Act Giner-Sorolla, Roger (2001), “Guilty Pleasures and Grim Together: Toward a General Model of Self-Control,” in Self Necessities: Affective Attitudes in Dilemmas of Self- Control in Society, Mind, and Brain, ed. Ran R. Hassin, Control,” Journal of Personality and Social Psychology,80 Kevin N. Ochsner, and Yaacov Trope, New York: Oxford, (February), 206–21. 335–53. Goldsmith, Kelly, Eunice Kim Cho, and Ravi Dhar (2012), “When Martin, Leonard L. and Blossom Davies (1998), “Beyond Guilt Begets Pleasure: The Positive Effect of a Negative Hedonism and Associationism: A Configural View of the Emotion,” Journal of Marketing Research, 49 (December), Role of Affect in Evaluation, Processing, and Self- 872–81. Regulation,” Motivation and Emotion, 22 (March), 33–51. Griskevicius, Vladas, Noah J. Goldstein, Chad R. Mortensen, Jill Martin, Leonard L., David W. Ward, John W. Achee, and Robert M. Sundie, Robert B. Cialdini, and Douglas T. Kenrick S. Wyer Jr. (1993), “Mood as Input: People Have to Interpret (2009), “Fear and Loving in Las Vegas: Evolution, Emotion, the Motivational Implications of Their Moods,” Journal of and Persuasion,” Journal of Marketing Research, 46 (June), Personality and Social Psychology, 64 (March), 317–26. 385–95. McFerran, Brent, Karl Aquino, and Jessica L. Tracy (2014), Griskevicius, Vladas, Michelle N. Shiota, and Stephen M. Nowlis “Evidence for Two Facets of Pride in Consumption: Findings (2010), “The Many Shades of Rose-Colored Glasses: An from Luxury Brands,” Journal of Consumer Psychology,24 Evolutionary Approach to the Influence of Different Positive (October), 455–71. Emotions,” Journal of Consumer Research, 37 (August), Patrick, Vanessa M., HaeEun Chun, and Deborah J. MacInnis 238–50. (2009), “Affective Forecasting and Self-Control: Why Han, DaHee, Adam Duhachek, and Nidhi Agrawal (2014), Anticipating Pride Wins over Anticipating Shame in a Self- “Emotions Shape Decisions through Construal Level: The Regulation Context,” Journal of Consumer Psychology,19 Case of Guilt and Shame,” Journal of Consumer Research, (July), 537–45. 41 (December), 1047–64. Pinkus, Rebecca T., Penelope Lockwood, Ulrich Schimmack, and Haws, Kelly L. and Cait Poynor (2008), “Seize the Day! Marc A. Fournier (2008), “For Better and for Worse: Encouraging Indulgence for the Hyperopic Consumer,” Everyday Social Comparisons between Romantic Partners,” Journal of Consumer Research, 35 (December), 680–91. Journal of Personality and Social Psychology,95 Hayes, Andrew F. (2013), An Introduction to Mediation, (November) 1180–1201. Moderation, and Conditional Process Analysis: A Roseman, Ira J. (1991), “Appraisal Determinants of Discrete Regression-Based Approach, New York, NY: Guilford. Emotions,” Cognition and Emotion, 5 (3), 161–200. Hayes, Andrew F. and Kristopher J. Preacher (2014), “Statistical Roseman, Ira J., Martin S. Spindel, and Paul E. Jose (1990), Mediation Analysis with a Multicategorical Independent “Appraisals of Emotion-Eliciting Events: Testing a Theory Variable,” British Journal of Mathematical and Statistical of Discrete Emotions,” Journal of Personality and Social Psychology, 67 (November), 451–70. Psychology, 59 (November), 899–915. Huang, Xun, Ping Dong, and Anirban Mukhopadhyay (2014), Roseman, Ira J., Cynthia Wiest, and Tamara S. Swartz (1994), “Proud to Belong or Proudly Different? Lay Theories “Phenomenology, Behaviors, and Goals Differentiate 514 JOURNAL OF CONSUMER RESEARCH

Discrete Emotions,” Journal of Personality and Social ———. (2007), “The Psychological Structure of Pride: A Tale of Psychology, 67 (August), 206–21. Two Facets,” Journal of Personality and Social Psychology, Russell, Dan (1982), “The Causal Dimension Scale: A Measure of 92 (March), 506–25. How Individuals Perceive Causes,” Journal of Personality Tracy, Jessica L., Aaron C. Weidman, Joey T. Cheng, and Jason P. and Social Psychology, 42 (June), 1137–45. Martens (2014), “Pride: The Fundamental Emotion of Salerno, Anthony, Juliano Laran, and Chris Janiszewski (2014), Success, Power, and Status,” in Handbook of Positive “Hedonic Eating Goals and Emotion: When Sadness Emotions, Vol. 1, ed. Michele M. Tugade, Michelle N. Decreases the Desire to Indulge,” Journal of Consumer Shiota, and Leslie D. Kirby, New York: Guilford, 294–310. Research, 41 (June), 135–51. Van de Ven, Niels, Marcel Zeelenberg, and Rik Pieters (2011), Sheldon, Kennon M., Richard Ryan, and Harry T. Reis (1996), “Why Envy Outperforms Admiration,” Personality and “What Makes for a Good Day? Competence and Autonomy Social Psychology Bulletin, 37 (June), 784–95. in the Day and in the Person,” Personality and Social Wheeler, Christian S., Kenneth G. DeMarree, and Richard E. Psychology Bulletin, 22 (December), 1270–79. Petty (2007), “Understanding the Role of the Self in Prime to Smith, Craig A. and Phoebe C. Ellsworth (1985), “Patterns of Behavior Effects: The Active-Self Account,” Personality and Cognitive Appraisal in Emotion,” Journal of Personality and Social Psychology Review, 11 (August), 234–61. Social Psychology, 48 (April), 813–38. Wilcox, Keith, Thomas Kramer, and Sankar Sen (2011), So, Jane, Chethana Achar, DaHee Han, Nidhi Agrawal, Adam “Indulgence or Self-Control: A Dual Process Model of the Duhachek, and Durairaj Maheswaran (2015), “The Effect of Incidental Pride on Indulgent Choice,” Journal of Psychology of Appraisal: Speciﬁc Emotions and Decision- Consumer Research, 38 (June), 151–63. Making,” Journal of Consumer Psychology, 25 (July), 359–71. Williams, Lisa A. and David DeSteno (2008), “Pride and Srull, Thomas K. and Robert S. Wyer (1979), “The Role of Perseverance: The Motivational Role of Pride,” Journal of Category Accessibility in the Interpretation of Information Personality and Social Psychology, 94 (June), 1007–17. about Persons: Some Determinants and Implications,” Journal Winterich, Karen Page and Kelly L. Haws (2011), “Helpful of Personality and Social Psychology, 37 (October), 1660–72. Hopefulness: The Effect of Future Positive Emotions on Tangney, June P. and Kurt W. Fischer (1995), Self-Conscious Consumption,” Journal of Consumer Research, 38 (October), Emotions: The Psychology of Shame, Guilt, Embarrassment, 505–24. and Pride, New York: Guilford. Witherington, David C. and Jennifer A. Crichton (2007), Tangney, June P. and Jessica L. Tracy (2012), “Self-Conscious “Frameworks for Understanding Emotions and Their Emotions,” in Handbook of Self and Identity, ed. Mark R. Development: Functionalist and Dynamic Systems Leary and June P. Tangney, New York: Guilford, 446–78. Approaches,” Emotion, 7 (August), 628–37. Tooby, John and Leda Cosmides (2008), “The Evolutionary Zemack-Rugar Yael, Canan Corus, and David Brinberg (2012), Psychology of the Emotions and Their Relationship to “The ‘Response-To-Failure’ Scale: Predicting Behavior Internal Regulatory Variables,” in Handbook of Emotions, Following Initial Self-Control Failure,” Journal of Marketing Vol. 3, ed. Michael Lewis, Jeannette M. Haviland-Jones, and Research, 49 (December), 996–1014. Lisa F. Barrett, New York: Guilford, 114–37. Zemack-Rugar Yael, James R. Bettman, and Gavan J. Fitzsimons Tracy, Jessica L. and Richard W. Robins (2004), “Putting the Self (2007), “The Effects of Nonconsciously Priming Emotion into Self-Conscious Emotions: A Theoretical Model,” Concepts on Behavior,” Journal of Personality and Social Psychological Inquiry, 15 (2), 103–25. Psychology, 93 (December), 927–39.