ASSOCIATION FOR CONSUMER RESEARCH

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

Motivated and the “Rosy View” in Retrospective Evaluations Priya Raghubir, New York University, USA Robert Latimer, New York University, USA

Three experiments show that (1) unpleasant experiences are evaluated more favorably after being recalled in detail, (2) detailed recall improves overall retrospective evaluations of experiences, but not evaluations of the individual aspects recalled, and (3) detailed recall with a social motive encourages selective retrieval of positive aspects of an experience.

[to cite]: Priya Raghubir and Robert Latimer (2013) ,"Motivated Recall and the “Rosy View” in Retrospective Evaluations", in NA - Advances in Consumer Research Volume 41, eds. Simona Botti and Aparna Labroo, Duluth, MN : Association for Consumer Research.

[url]: http://www.acrwebsite.org/volumes/1015481/volumes/v41/NA-41

[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/. Motivated Recall and the “Rosy View” in Retrospective Evaluations Robert Latimer, New York University, USA Priya Raghubir, New York University, USA

ABSTRACT The Fading Affect illustrates the power of social retellings, Academics and stand up comics alike have noted the sometimes motivated retrieval provides a mechanism through which bias can extraordinary divergence of retrospective evaluations from an actual enter recollections, and perspective taking demonstrates that biased experience. Academics, for their part, have presented ample evidence recall will change later . We contend that when people for experiences ranging from spring break vacations (Kemp, Burt, & recall past experiences in detail they selectively retrieve certain Furneaux, 2008; Wirtz, Kruger, Scollon, & Diener, 2003), trips to components of the experience. When people are later asked to Disneyland (Sutton, 1992), and bike trips across California (Mitchell, evaluate an experience, their evaluations are colored by the earlier Thompson, Peterson, & Cronk, 1997). For these experiences and selective retrieval. Our desire to be seen by our peers as wise, others, the bulk of research suggests that changes in evaluations are interesting people who do fun things combined with the frequency of directional; people have a “rosy view” of their past (Mitchell et al., social detailed recall ensures that the color is usually “rosy”. 1997; Wirtz & Kruger, 2002). There are exceptions, however. Wirtz Motivated retrieval, perspective taking, and the Fading et al. (2003) reported an increase in negative affect from online to Affect Bias happily converge to provide an account for changes in retrospective reports, although overall evaluations did improve. retrospective evaluations through detailed recall. The Fading Affect Anecdotally, people also exaggerate negative aspects of experiences, Bias provides direction for how memories will change over time, sacrificing accuracy for entertainment value. Any theory attempting perspective taking provides a mechanism through which memories to explain these shifting evaluations of vacations must account for can change, and motivations provide the perspective, suggest the how the fish you caught grows larger over time, the local food more moderating role of ambiguity, and a reason for the typical “rosy view” savory, and the already staggeringly incompetent mountain guide phenomenon. We present three experiments. In Experiment 1, we even less adroit. demonstrate that detailed recall can improve retrospective evaluations. There are several lines of research how memories and their In Experiment 2, we reverse the effect and suggest a role for selective associated affect change over time. Evidence from attitudes research retrieval of pleasant or unpleasant components, rather than a change suggests that evaluations could become polarized (e.g. Tesser, 1975; in the construal of those components. Finally, Experiment 3 tests the Srull & Wyer, 1981) or diluted (e.g. Linville, 1982; Millar & Tesser, motivational account against competing interpretations by changing 1985; Wilson & Schooler, 1991) over time as a result of many factors participants’ motivations during detailed recall. ranging from the complexity of the topic to the need for justification. A substantial body of work also deals with how people feel before EXPERIMENT 1 or after an experience and retrospective evaluations may be based Firms habitually ask customers to recall their consumption on current feelings (Ross, McFarland, & Decourville, 1989). Wilson experience by administering surveys. We asked participants to and Gilbert’s (2003, 2008) AREA (Attend, React, Explain, and provide a description of a complex experience: their last trip by Adapt) contends that people react emotionally only to self-relevant airplane. Participants also provided ratings on pleasantness and events that they do not understand. Taylor’s (1991) mobilization- satisfaction for the best and worst moments of the experience, as minimization account suggests that people will have strong initial well as the overall experience. We anticipated that ratings for the reactions to negative experiences, but that their impact will fade with components of the experience would follow the pattern of the Fading time. Research derived from Pennebaker’s (1997) clinical work has Affect Bias. Participants who spend longer retelling the story of their demonstrated that writing about a traumatic experience will reduce trip should rate the (presumably unpleasant) worst moment of their negative feelings upon later recall of the experience. Walker and trip as more pleasant, but the (presumably pleasant) best moment of colleagues (e.g. Skowronski & Thompson, 2003; Walker, Vogl, & their trip should receive similar ratings from both groups. The Fading Thompson, 1997) have extended this finding to recall of everyday Affect Bias literature does not provide predictions for the impact pleasant and unpleasant experiences and demonstrated an asymmetry of detailed recall on complex experiences with both pleasant and they labeled the . Negative affect associated with unpleasant moments, but if the worst moments improve with detailed unpleasant experiences fades faster than positive affect associated recall and the best moments remain constant with recall, the overall with positive experiences. Further, the authors suggest that the experience should become retrospectively more pleasant. Fading Affect Bias is driven by social recall – talking about an experience with peers (Skowronski et al., 2004). Just as the feelings METHOD evoked by memories can change, so too can the actual contents Participants. Twenty-four students enrolled in an introduction recalled. Cognitively focused psychologists have demonstrated that to marketing course completed the experiment for partial course depending on how a story is told, the details of that story may change credit. when recalled later (Dudokovic, Marsh, & Tversky, 2004; Huston Materials & Procedure. All participants were asked to provide & Marsh, 2005; Marsh & Tversky, 2000). When participants write the basic information: the month and year when they last took a trip about a scenario from one perspective, their memories will reflect by airplane, as well as the departure and destination airports. Eleven that perspective’s bias. Finally, research on motivated retrieval in self of the participants were then asked to “describe very briefly the last deception suggests that details recalled can change depending on the time you took a trip in an airplane. Begin with your arrival at the desires of the person reminiscing (e.g. Kunda,1987, 1990; Klein & starting airport and end with your departure from the destination Kunda, 1992). airport. You have 2 minutes to write your description of the trip, try In this paper we focus on the interplay between three of the to use all of the time.” The remaining thirteen participants received above lines of research to clarify the role of detailed recall in creating identical instructions, with the exceptions that they were asked to retrospective evaluations that diverge from online experience. “describe in as much detail as possible” and were given 15 minutes

Advances in Consumer Research 388 Volume 41, ©2013 Advances in Consumer Research (Volume 41) / 389 to provide their descriptions. After completing their descriptions, all were asked to list an unpleasant aspect (F(1, 74) = 93.0, p < .0005). participants were asked to rate the pleasantness of the trip on a -3 The valence of the aspect listed also had a marginally significant to 3 (very unpleasant to very pleasant) scale and their satisfaction effect on participants’ evaluation of the overall flight, but only in the with the trip on a 1 to 7 (not at all satisfied to completely satisfied) detailed recall condition. Participants who recalled a pleasant aspect scale. Participants provided ratings on both of these scales for the in detail (M = 5.2) evaluated the overall flight more positively than overall experience, the best moment of the experience, and the worst participants who recalled a negative aspect in detail (M = 4.4) (F(1, moment of the experience. 74) = 4.1, p < .1). When participants wrote about a product they had recently bought, the difference between recalling the pleasant RESULTS & DISCUSSION (M = 4.8) and unpleasant (M = 4.7) aspect was not significant, p > The ratings were submitted to a 2 (Minutes writing, between .2. There were no significant effects of condition on either scale, all subjects) X 3 (Component, within subjects) mixed ANOVA. ps > .2. Experiment 2 demonstrated that rehearsing components of Participants were randomly assigned to conditions, so we can an experience will slant the overall evaluation towards the valence assume that differences between groups in the recalled pleasantness of those components. Perhaps more importantly, Experiment 2 of and satisfaction with the flights were caused by the manipulations. showed that the valences of specific aspects of an experience are Spending more time describing the past airplane trip made the at least somewhat resistant to change during detailed recall. This trip retrospectively more pleasant, F(1,22) = 5.175, p = .033, and could be because of participants’ low motivation to see unpleasant marginally more satisfying, F(1,22) = 3.211, p = .087. The longer components of their flight – such as cramped seats – in a more descriptions had their intended effect; Delta would be well advised pleasant light, conflicting motivations to take the rosy view of the to get me to describe my trip if they want to keep my business. Not aspect or demonstrate how well they were following the instructions surprisingly, the overall experience, best moment, and worst moment by choosing a truly unpleasant aspect, or the lack of ambiguity in the differed in their retrospective pleasantness, F(1,22) = 39.10, p < aspects chosen. .0005, and rated satisfaction, F(1,22) = 34.43, p < .0005. However the interaction between the length of the description and moment did EXPERIMENT 3 not reach significance, p > .1. Experiments 1 and 2 relied on participants’ innate motivation to take the rosy view of an experience. In Experiment 3, we provide EXPERIMENT 2 participants with explicit motives during recall in order to provide Experiment 1 demonstrated that detailed recall can result in another demonstration of how recall can change retrospective improved retrospective evaluations, but the source of the effect is evaluations and examine some alternatives to the motivational uncertain. Participants who extensively recalled the flight may have account. In particular, Gilbert and Wilson’s (2003, 2008) AREA selectively retrieved a greater number of pleasant components of model of affective adaptation suggests that recall with a focus on their experience, changed their construal of the components they explanation should reduce the intensity of all affect associated with did retrieve, or retrieved components from a larger experience (i.e. an experience, while the motivational account predicts a strong their whole trip rather than just the flight). In order to test the effects reduction in negative affect for unpleasant experiences, particularly of detailed recall in a more constrained setting, we restricted recall when social pressure is salient. to a single component of a flight experience, either pleasant or unpleasant. METHOD Participants. Four hundred and five participants were recruited METHOD using Amazon’s Mechanical Turk service. Participants were Participants. Seventy eight undergraduate marketing students compensated $0.25 for an average of 14 minutes participation participated for partial credit in their introductory course. in a survey advertised as taking between 15 and 20 minutes. 137 Materials & Procedure. As in Experiment 1, participants were participants were excluded from the analysis for failing to read the first asked to describe some basic information about the last flight body text of a test question or for failing to describe an experience. they had taken, including their destination, airline, and the recency Excluding these participants did not change the direction of the of the flight. Participants were then asked to list a single aspect – results. either pleasant or unpleasant depending on condition – of their most Materials & Procedure. Participants were asked to list a single recent flight. On the next page of the questionnaire, participants were experience from their lives that happened in that last three months, asked to provide a detailed description of either the aspect they had lasted between one hour and 24 hours, that they were present at (e.g. just listed or a product they had purchased recently (control). After not a news story), and that was either overall positive, negative, or finishing the description, participants completed several unrelated mixed in valence. Participants in the detailed recall conditions were studies for approximately 40 minutes. Finally, they rated the then asked to describe this experience and told that their descriptions pleasantness (1 very unpleasant – 7 very pleasant), enjoyableness (1 would be evaluated by a judge later. The fictional judge’s criteria not at all enjoyable – 7 very enjoyable), and their satisfaction (1 very varied by condition. He would evaluate participants for how “fun and dissatisfied – 7 very satisfied) with both the aspect of the flight they interesting you would be as a friend”, how “successful at explaining had listed and the flight overall. things you would be as a teacher”, or how “reliable you would be as a witness”. After providing their descriptions, participants listened RESULTS & DISCUSSION to and rated a series of four 40 second sound clips that served as a The evaluations were combined into two scales, evaluation buffer between the description and the evaluations of the experience. of the aspect (α = .97) and of the flight overall (α = .92). The two Finally, participants asked to evaluate the experience they had scales were submitted to a 2 (aspect valence) X 2 (recall aspect/ described earlier for its pleasantness, the extent to which they describe product) Multivariate Analysis of Variance (MANOVA). enjoyed it, and their satisfaction (1 not at all – 7 very). Participants Not surprisingly, when participants were asked to list a pleasant were instructed that their evaluations were for control purposes only aspect they evaluated that aspect more positively than when they and that they would not be viewed by the judges. 390 / Motivated Recall and the “Rosy View” in Retrospective Evaluations RESULTS & DISCUSSION takes place with the goal of appearing fun and interesting as a friend Participants’ ratings of the described experience were combined - a goal that is doubtless very common in everyday recall. into a single scale (α = .941). This scale was submitted to a 4 (detailed recall: none, friend, teacher, witness) X 3 (Experience valence: GENERAL DISCUSSION positive, negative, mixed) between subjects ANOVA. As expected, I have presented evidence that detailed recall – particularly there was a main effect of valence, with participants evaluating a social detailed recall – has the power to improve retrospective negative experience (M = 2.4) less positively than a mixed (M = evaluations of experiences, adding a rosy tint to experiences 4.8) or positive (M = 5.8) experience (F(2,256) = 89.7, p < .0005). that initially seemed negative. Experiment 1 provided an initial Neither the interaction nor the main effect of detailed recall was demonstration of the effect in the context of evaluations of a past significant, ps > .1. We also executed three planned contrasts. For the trip by airplane. Experiment 2 clarified that individual components negative experience, participants who recalled believing they would of an experience are unlikely to be reconstrued with detailed recall, be evaluated as a friend evaluated the experience more positively (M but that an overall evaluation is influenced by the components that = 2.9) than participants in the control condition (M = 2.0) (F(1, 256 are recalled, positive or negative. Next, Experiment 3 demonstrated = 5.82, p < .05). The planned contrast for the AREA model failed that detailed recall with the motive of seeming “fun and interesting” to reach significance however, with participants who believed they improves evaluations for initially negative experiences, mimicking would be judged on the quality of their explanation not seeing an the results from the Fading Affect Bias. increase in evaluation in the negative condition and a decrease in the This paper presents empirical evidence for the influence of evaluation in the positive condition over the control participants, p > detailed recall on retrospective evaluations, drawing its predictions .1. Finally, when all detailed recall conditions were pooled, they did from the intersection of past work on the Fading Affect Bias, not produce a significant improvement in evaluations over the control perspective taking, and motivated retrieval. These three lines condition, p > .1. The results of Experiment 3 show a remarkable of research complement each other to produce predictions for correspondence to the Fading Affect Bias, but for retrospective the plasticity of with broad theoretical and managerial evaluations rather than evoked affect. To test for selective retrieval of relevance. For example, in order to encourage pleasant memories positive aspects of the experience, participants written descriptions of the experiences they provide, firms might combine satisfaction were coded by two raters blind to the experiment’s hypotheses. surveys with social media. Consumers who believe that their Participants with a social motive during detailed recall listed more satisfaction report will be seen by their peers are likely to list more positive aspects of the experience. The positive aspects listed will positive aspects compared to negative aspects (Mpositive – negative = 0.76) have a double benefit to the firm: (1) improving retrospective than participants with accuracy (Mpositive – negative = -0.73) or explanation evaluations of the consumer and (2) increasing the positivity of word (Mpositive – negative = -0.61) motives, all ps < .05. Negative experiences are improved by detailed recall, but only when that detailed recall of mouth disseminated by the consumer.

Table 1 . Summary of Studies Study 1 Study 2 Study3 Participants write about one aspect Participants write about their most of their most recent flight on an Participants write about an recent flight on an airplane and then Description airplane and then rate the retrospective autobiographical event and rate its rate the retrospective pleasantness pleasantness of both the aspect and the retrospective pleasantness. of the flight. flight overall. Type of description (none / social Time writing (detailed description/ Time Writing motive / accuracy motive / explanation IV no description) X Valence of aspect (2 min/15 min) motive) X Valence of event (pleasant / described (pleasant/unpleasant) unpleasant / mixed) Pleasantness & Satisfaction for Pleasantness, Satisfaction & Enjoyment Pleasantness, Satisfaction & Enjoyment Primary DV whole flight, best moment, and for whole flight and the aspect for the experience. worst moment. described.

Describing unpleasant experiences with a social motive improved overall Describing negative aspects in detail retrospective evaluations of the produced poorer overall evaluations Evaluations of the whole flight and unpleasant experiences. Other types and rehearsing positive aspects in detail Result worst moment were higher when of detailed recall were ineffective in produced better overall evaluations. participants wrote for 15 min. changing overall evaluations. Detailed descriptions did not influence Participants with a social motive listed evaluations of the aspects themselves. more positive than negative aspects when describing their experience. Individual aspects of an experience do Social detailed recall encourages recall Detailed recall improves not change in valence with detailed of the positive aspects of unpleasant Implication retrospective evaluations by making recall. Overall evaluations change to fit experiences, resulting in improved the worst moments better. the aspects recalled. retrospective evaluations. Advances in Consumer Research (Volume 41) / 391 REFERENCES Pennebaker, J. W. (1997). Writing about emotional experiences as a Barber, S. J., Rajaram, S., & Marsh, E. J. (2008). Fact learning: therapeutic process. Psychological Science, 8(3), 162-166. how information accuracy, delay, and repeated testing change Pezdek, K., Whetstone, T., Reynolds, K., Askari, N., & Dougherty, retention and retrieval experience. Memory, 16(8), 934-946. T. (1989). Memory for real world scenes: The role of Bernsten, D. (1996). Involuntary autobiographical memories. consistency with schema expectation. Journal of Experimental Applied Cognitive , 10, 435-454. Psychology: Learning, Memory, and Cognition, 15(4), 587- Bernsten, D. (1998). Voluntary and involuntary access to 595. . Memory, 6(2), 113-141. Rook, D. W. (1985). The ritual dimension of consumer behavior. Brewer, W. F. & Treyens, J. C. (1981). Role of schemata in memory Journal of Consumer Research, 12, 251-264. for places. Cognitive Psychology, 13, 207-230. Schmidt, S. R. (2002). Outstanding memories: the positive and Bryant, F. B., & Brockway, J. H. (1997). Hindsight bias in reaction negative effects of nudes on memory. Journal of Experimental to the verdict in the O. J. Simpson criminal trial. Basic and Psychology: Learning, Memory, and Cognition, 28, 353-361. Applied , 19, 225–241. Schulz, L. S. (1971). Effects of high-priority events on recall and Dudokovic, N. M., Marsh, E. J., & Tversky, B. (2004). Telling a recognition of other events. Journal of Verbal Learning and story or telling it straight: the effects of entertaining versus Verbal Behavior, 10, 322-330. accurate retellings on memory. Applied Cognitive Psychology, Schutzwol, A. (1998). Surprise and schema strength. Journal of 18, 125-143. Experimental Psychology: Learning, Memory, and Cognition, Dutta, S., Kanungo, R. N., & Freibergs, V. (1972). Retention of 24(5), 1182-1199. affective material: Effects of intensity of affect on retrieval. Skowronski, J. J., Gibbons, J. A., Vogl, R. J., & Walker, W. R. Journal of Personality and Social Psychology, 23, 64-80. (2004). The effects of social disclosure on the affective Erasmus, A. C., Boshoff, E., & Rousseau, G. G. (2002). The intensity provoked by autobiographical memories. Self and potential of using script theory in consumer research. Journal Identity, 3, 285-309. of Family Ecology and Consumer Sciences, 30, 1-9. Smith, D. A. & Graesser, A. C. (1981). Memory for actions in Farrel, S., & Lewandowsky, S. (2003). Dissimilar items benefit scripted activities as a function of typicality, retention interval, from phonological similarity in serial recall. Journal of and retrieval task. Memory & Cognition, 9, 550-559. Experimental Psychology: Learning, Memory, & Cognition, Kemp, S., Burt, C. D. B., & Furneaux, L. (2008). A test of the peak- 29, 838-849. end rule with extended autobiographical events. Memory & Frederickson, B. L., & Kahneman, D. (1993). Duration neglect Cognition, 36(1), 138-138. in retrospective evaluations of affective episodes. Journal of Sutton, R.I. (1992). Feelings about a Disneyland visit: Photographs Personality and Social Psychology, 65, 45-55. and reconstruction of bygone emotions. Journal of Geraci, L., & Manzano, I. (2010). Distinctive items are salient Management Inquiry, 1, 278-287. during : Delayed judgements of learning predict Trafimow, D., & Wyer, R. S. (1993). Cognitive representation of the isolation effect. The Quarterly Journal of Experimental mundane social events. Journal of Personality and Social Psychology, 63(1), 50-64. Psychology, 64(3), 365-376. Hubbert, A. R., Sehorn, A. G., & Brown, S. W. (1995). Service Tversky, B., & Marsh, E. J. (2000). Biased retellings of events yield expectations: the consumer versus the provider. International biased memories. Cognitive Psychology, 40, 1-38. Journal of Service Industry Management, 6(1), 6-21. Walker, W. R., & Skowronski, J. J. (2009). The fading affect bias: Hunt, R. R. (2009). Does salience facilitate longer-term retention? But what the hell is it for? Applied Cognitive Psychology, 23, Memory, 17(1), 49-53. 1122-1136. Kelley, M. R., & Nairne, J. S. (2001). Von Restorff revisited: Walker, W. R., Skowronski, J. J., & Thompson, C. P. (2003). Life Isolation, generation, and memory for order. Journal of is pleasant – and memory helps to keep it that way! Review of Experimental Psychology: Learning, Memory, & Cognition, General Psychology, 7(2), 203-210. 27, 54-66. Walker, W. R., Vogl, R.J., & Thompson, C. P. (1997). Kemp, S., Burt, C. D. B., & Furneaux, L. (2008). A test of the peak- Autobiographical memory: Unpleasantness fades faster than end rule with extended autobiographical events. Memory & pleasantness over time. Applied Cognitive Psychology, 11, Cognition, 36(1), 138-138. 399-413. Kunda, Z. (1987). Motivated inference: Self-serving generation Wallendorf, M. & Arnould, E. J. (1991). “We Gather Together”: and evaluation of causal theories. Journal of Personality and Consumption rituals of Thanksgiving day. Journal of Social Psychology, 53, 636–647 Consumer Research, 18, 13-31. Marsh, J.E. (2007). Retelling is not the same as recalling. Current Wilson, T. D., & Gilbert, D. T. (2008). Explaining away: A model Directions in Psychological Science, 16(1), 16-20. of affective adaptation. Perspectives on Psychological Science, Mitchell, T.R., Thompson, L., Peterson, E., & Cronk, R. (1997). 3(5), 370-386. Temporal adjustments in the evaluation of events: The “rosy Wirtz, D., Kruger, J., Scollon, C. N., & Diener, E. (2003). What view.” Journal of Experimental Social Psycholoogy, 33, 421- to do on spring break? The role of predicted, on-line, and 448. remembered experience in future choice. Psychological Nelson, L. D., & Meyvis, T. (2008). Interrupted consumption: Science, 14(5), 520-524. Disrupting adaptation to hedonic experiences. Journal of White, J. D., & Carlston, D. E. (1983). Consequences of Schemata Marketing Research, 45, 654-664. for , impressions, and recall in complex social Nottenburg, G., & Shoben, E. J. (1980). Scripts as linear orders. interactions. Journal of Personality and Social Psychology, Journal of Experimental Social Psychology, 16, 329-347. 45(3), 538-549.