Culture, Interpersonal Perceptions, and Happiness in Social Interactions

http://psp.sagepub.com

Personality and Social Psychology Bulletin

DOI: 10.1177/0146167207311198 2008; 34; 307 Pers Soc Psychol Bull

Shigehiro Oishi, Minkyung Koo and Sharon Akimoto Culture, Interpersonal Perceptions, and Happiness in Social Interactions

http://psp.sagepub.com/cgi/content/abstract/34/3/307 The online version of this article can be found at:

Published by:

http://www.sagepublications.com

On behalf of:

Society for Personality and Social Psychology, Inc.

can be found at:Personality and Social Psychology Bulletin Additional services and information for

http://psp.sagepub.com/cgi/alerts Email Alerts:

http://psp.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://psp.sagepub.com/cgi/content/refs/34/3/307SAGE Journals Online and HighWire Press platforms):

(this article cites 34 articles hosted on the Citations

© 2008 Society for Personality and Social Psychology, Inc.. All rights reserved. Not for commercial use or unauthorized distribution. at UNIV OF VIRGINIA on July 29, 2008 http://psp.sagepub.comDownloaded from

http://www.spsp.org/

http://psp.sagepub.com/cgi/alerts

http://psp.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://psp.sagepub.com/cgi/content/refs/34/3/307


307

Culture, Interpersonal Perceptions,and Happiness in Social Interactions

Shigehiro OishiMinkyung KooUniversity of Virginia

Sharon AkimotoCarleton College

AFFECT IN INTERPERSONAL CONTEXTS

Several theorists have proposed that affective out-comes of social interactions are predicted by the degreeto which an individual feels understood by the interac-tion partner (e.g., Reis, Clark, & Holmes, 2004). Thisfeeling, when it emerges as a result of the initial interac-tion, is in turn likely to shape the quality of future inter-actions. The importance of feeling understood is alsowidely recognized in the literature on psychotherapy, asa client who feels misunderstood at the first session isunlikely to continue the client–therapist relationship(Elliott & James, 1989). This suggests that the subjec-tive feeling of being understood is an important factorin predicting affective outcomes of social interactionsand in the formation of close relationships.

What causes the subjective feeling of being under-stood? In dating and marital relationship contexts,Murray, Holmes, Bellavia, Griffin, and Dolderman(2002) found that perceived similarity in personalityand values was associated with a greater degree of felt

Authors’ Note: We gratefully acknowledge the support of theNational Institute of Mental Health Research Grant R01-MH066857and the Positive Psychology Young Scholar Grant. We also thank DonChoi, Lihn Tran, Jenny Reinke, Gary Sherman, and MargaritaKrochik for their help with the data collection. Correspondenceregarding this article should be addressed to Shigehiro Oishi atDepartment of Psychology, University of Virginia, P.O. Box 400400,Charlottesville, VA 22904-4400; e-mail: [email protected].

PSPB, Vol. 34 No. 3, March 2008 307-320DOI: 10.1177/0146167207311198© 2008 by the Society for Personality and Social Psychology, Inc.

The authors examined cultural differences in interper-sonal processes associated with happiness felt in socialinteractions. In a false feedback experiment (Study 1a),they found that European Americans felt happier whentheir interaction partner perceived their personal selfaccurately, whereas Asian Americans felt happier whentheir interaction partner perceived their collective selfaccurately. In Study 1b, the authors further demon-strated that the results from Study 1a were not becauseof cultural differences in desirability of the traits used inStudy 1a. In Studies 2 and 3, they used a 2-week eventsampling method and replicated Study 1. Unlike AsianAmericans, African Americans were not significantlydifferent from European Americans in the predictors ofhappiness in social interactions. Together, this researchshows that interpersonal affirmation of importantaspects of the self leads to happiness and that culturaldifferences are likely to emerge from the emphasisplaced on different aspects of the self.

Keywords: culture; happiness; positive affect; self

Some social encounters leave one feeling pleasant,whereas others leave one feeling unpleasant. What

predicts the affective outcome of social interactions?We report three studies that investigate this questionin the context of culture and show that the predictorsof affective outcome of social interactions vary sys-tematically across cultures, depending on whichaspects of the self are accurately perceived by an inter-action partner.



understanding by the romantic partner, which in turnpredicted relationship satisfaction. Aside from per-ceived similarity in personality and values, Reis andPatrick (1996) posited that the partner’s accurate per-ception of the target’s central self-concepts has to becommunicated to the target for the target to feel under-stood. Reis et al. (2004) recently extended the earliermodels by arguing that responsiveness to the centralaspects of the self is critical to affective outcomes ofsocial interactions. It is not surprising, then, that peoplelike their interaction partner when the partner accu-rately perceives important aspects of their self-concepts(e.g., Chen, Chen, & Shaw, 2004; Swann, 1990).

CULTURE AND SELF

Cross-cultural research found that central aspects of self-concept vary across cultures (e.g., Kanagawa, Cross, &Markus, 2001; Markus & Kitayama, 1991; Triandis,1989). For example, Cousins (1989) found that 58% ofAmericans’ top five sentence completions in the 20-state-ments test (“I am . . .”) were composed of personalitytraits such as honest, whereas only 19% of Japanese’s topfive responses included personality traits. In addition, amongAmericans, only 9% of the top five responses referencedsocial roles such as college student, whereas among Japanese27% of the top five responses mentioned social roles.

The literature on culture and the self suggests that thesources of interpersonal understanding are different forindividuals with different cultural backgrounds.Specifically, because personal aspects of the self (e.g., per-sonality traits and abilities) are known to be chronicallysalient and important to European Americans, we predictthat accurate recognition of the personal self by an inter-action partner should be the key to happiness in socialinteractions among European Americans. In contrast,because the collective self (e.g., group membership, role)is known to be chronically accessible to Asian Americans,we expect that accurate recognition of the collective selfby an interaction partner should be particularly impor-tant to happiness in social interactions among this group.Analogous to these cultural predictions, Oishi, Lun, andSherman (2007) recently found that individuals whomoved a lot while growing up (movers) deemed the per-sonal self more central than did those who did not move(nonmovers), whereas nonmovers deemed the collectiveself more central than did frequent movers. Furthermore,frequent movers felt happy when their interaction partnerperceived their personal self accurately, whereas non-movers felt happy when their interaction partner per-ceived their collective self accurately. We conducted alaboratory experiment and two event sampling studiesthat tested our cultural difference predictions in the rela-tionship between the type of self accurately perceived by

an interaction partner and the level of happiness experi-enced in that interaction.

THIS RESEARCH

In Study 1a, we experimentally manipulated the part-ner’s accurate perception of participants’ personal andcollective self and then measured how happy participantsfelt about the interaction. This manipulation allowed usto infer a causal direction from the accurate perception ofthe personal versus collective self to happiness. In Study1b, we tested cross-cultural equivalence of the personal-ity traits used in the manipulation of Study 1a to ascer-tain that the findings from Study 1a were not because ofcultural differences in the desirability of the personalitytraits used in the first study. Because laboratory experi-ments typically boast high internal validity but lack eco-logical validity, we also conducted event sampling studies(Studies 2 and 3) in which participants reported socialinteractions and their affective reactions to those interac-tions in their natural, daily contexts. Thus, Studies 2 and3 provide valuable insight into everyday interpersonalprocesses and affective experiences while minimizing ret-rospective bias and memory bias, which are major con-cerns in well-being and relationship research in general(Reis & Gable, 2000).

In Study 3 we also explored the interpersonalprocesses associated with happiness in social interac-tions among African Americans as well as amongEuropean and Asian Americans. Previous researchshowed that African Americans are more similar toEuropean Americans than to Asian Americans in theircultural orientation toward the self. For instance,African Americans are known to endorse more individ-ualistic values than do European Americans (d = 0.31)and Asian Americans (d = 0.55) and endorse collec-tivistic values as much as European Americans (d =0.04), and less than Asian Americans (d = –0.35;Oyserman, Coon, & Kemmelmeier, 2002). AfricanAmericans also report having higher levels of globalself-esteem than do European Americans (d = 0.19)and Asian Americans (d = 0.49; Twenge & Crocker,2002). In addition, African Americans and EuropeanAmericans often show self-protective attribution pat-terns for negative experiences (Crocker & Major, 1989;Zuckerman, 1979), whereas Asian Americans do not(Oishi, Wyer, & Colcombe, 2000). If AfricanAmericans in Study 3 showed patterns of happiness insocial interactions that were more similar to EuropeanAmericans than to Asian Americans, this would providefurther support for the cultural explanation of the dif-ference between European Americans and AsianAmericans in Studies 1 and 2. If, on the other hand,African Americans showed different patterns of happiness

308 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN



than did European Americans, then the differencebetween Asian and European Americans found inStudies 1 and 2 cannot be attributed to cultural factors.In short, the three-group comparison in Study 3 pro-vides a more stringent test of the cultural account ofhappiness felt in social interactions than a typical two-group comparison.

Our research is not the first attempt to investigatecultural differences in interpersonal processes. Aglimpse at the existing literature reveals work on every-thing from cultural similarities and differences in themeaning of friendship (e.g., Adams & Plaut, 2003); theaccuracy of interpersonal perception and liking (e.g.,Heine & Renshaw, 2002); social support seeking (e.g.,Taylor et al., 2004); and the frequency, duration, andintimacy of social interactions (e.g., Wheeler, Reis, &Bond, 1989) to love and marriage (e.g., Levine, Sato,Hashimoto, & Verma, 1995), interpersonal contextsand emotion (e.g., Mesquita & Karasawa, 2002; Oishi,Diener, Scollon, & Biswas-Diener, 2004), and relation-ship harmony and life satisfaction (Kwan, Bond, &Singelis, 1997). However, our research is unique in sev-eral respects. First, building on the close relationship lit-erature (e.g., Reis et al., 2004), we examined the linkbetween accurate perception of central aspects of theself and affect, a critical process in the formation andmaintenance of close relationships, for the first time ina cross-cultural context. Second, this is the first researchon cultural and interpersonal processes (to our knowl-edge) that employs both experimental and event sam-pling methods. Finally, our three-group comparison inStudy 3 provides a more rigorous test than a typicaltwo-group comparison does for the cultural account ofthe differences between European Americans’ and AsianAmericans’ felt happiness in social interactions.

STUDY 1A: LABORATORY EXPERIMENT

Method

Participants

Participants were 71 students who identified them-selves as European American (29 male, 41 female, 1 didnot specify sex) and 47 students who identified them-selves as Asian or Asian American (23 male, 22 female,2 did not specify sex). Participants were enrolled in anintroductory psychology course at the University ofIllinois at Urbana-Champaign.

Materials and Procedure

Participants completed a three-phase experiment inpairs. During the first phase, participants were seated inindividual cubicles where they completed a short survey

that included a list of 10 traits. The ten traits listed werehardworking, intelligent, fun-loving, friendly, stubborn,cooperative, relaxed, leader, emotional, and rational.We chose 8 traits (the 10 aforementioned traits exclud-ing leader and stubborn) from Suh’s (1999, Pilot Study2) earlier work on spontaneous personality descrip-tions. These are the traits most often spontaneously gen-erated by American and Korean college students. Weadded leader and stubborn to the list to include theindependent or assertive dimension that appears impor-tant for many American college students (Cantor,1994). Participants were first asked to choose two traitson the list that described them most accurately. Next,participants listed their group affiliations that wereimportant to them at the time (e.g., Psi Chi, sorority) orthat were important to them in high school (e.g., studentcouncil, baseball). They then listed the city or town wherethey were born and the cities or towns in which they grewup. They were asked to indicate whether the city or townin which they grew up was a large city (e.g., Chicago), thesuburb of a large city (e.g., Elmhurst), a medium-sizedcity (e.g., Peoria), the suburb of a medium-sized city, or arural town. Finally, they were asked to indicate theirdeclared or intended academic major.

During the second phase, participants were broughttogether in a larger room and instructed to engage in adiscussion task. Three discussion topics (illegal drug use,college dropouts, career choice) were given to them, andthey discussed each topic for about 3 to 5 minutes. Thistask was intended to allow participants to express theirintelligence, rationality, emotionality, assertiveness,stubbornness, and so forth. Next, the participants com-pleted a basketball task, in which they took 20 freethrows as a team. They were told that the team thatmade the most baskets would win two movie tickets.This task was devised to allow participants to expresstheir fun-loving side, cooperativeness, and other suchaspects of their personality. The structure of these inter-action tasks encouraged participants to express andcommunicate who they are to the extent that this is pos-sible in the restricted context of a laboratory interaction.

After these two interaction tasks, participants wereonce again separated to individual cubicles and beganthe third phase of the experiment. They were asked touse personality traits to describe their interaction part-ner as well as to guess what groups the interaction part-ner belonged to (e.g., debate team), her or his academicmajor (e.g., biology), and where the partner grew up(e.g., in a large city). This impression sheet was thenturned in to the experimenter.

Once both participants had completed the impressionsheet, each received another impression sheet that hadpresumably been completed by the partner. In reality, afalse impression sheet was switched with the actualimpression sheet before it was given to the participants.

Oishi et al. / CULTURE AND INTERPERSONAL PERCEPTIONS 309



Based on the assigned experimental condition and theself-descriptions provided by the participants during thefirst phase of the experiment, the experimenter createdthese false impression sheets while the participants wereinteracting with one another. In the accurate personalself condition, synonyms replaced the personality traitsactually chosen by the participant on the partner’s fakeimpression sheet (e.g., laid-back replaced relaxed). In theinaccurate personal self condition, traits that were qual-itatively different from participants’ self-descriptionswere used on the partner’s fake impression sheet (e.g.,passionate replaced rational; see Table 1 for the list of 16traits used in feedback). In the accurate collective selfcondition, the experimenter copied one group affiliationmentioned by the participant onto the fake impressionsheet as well as the correct academic major and the cor-rect size of the city or town where the participant grew up,whereas in the inaccurate collective self condition, theexperimenter mentioned a different group, major, and sizeof city or town. To make the collective self feedback asequivalent across cultures as possible, we did not use anyethnic groups (e.g., Chinese student association) in the falsefeedback. Participants were randomly assigned to one offour conditions: AA (both personal and collective selvesaccurate: 15 European Americans, 11 Asian Americans), IA(inaccurate personal, accurate collective selves: 18European Americans, 12 Asian Americans), AI (accuratepersonal, inaccurate collective selves: 20 EuropeanAmericans, 13 Asian Americans), or II (both personal andcollective selves inaccurate: 18 European Americans, 10Asian Americans).

After inspecting an impression sheet presumably com-pleted by the interaction partner, participants completedtwo manipulation check items: “How accurate was theinteraction partner’s impression of the personal aspectsof you (e.g., personality traits)?” and “How accuratewas the interaction partner’s impression of the collectiveaspects of you (e.g., group affiliations, academicmajor)?” They responded to these items using a 7-pointscale (1 = not at all accurate to 7 = absolutely accurate).Then participants rated how happy and good they feltabout the impression that the partner had formed ofthem using a 7-point scale (1 = not at all to 7 = verystrongly). The mean positive affect (happy and good)was 5.11 (SD = 1.21) and Cronbach’s alpha was .85.

Results and Discussion

Our manipulation was successful, as participants inthe accurate personal self conditions (AA, AI) indeedperceived that their interaction partner’s impression oftheir personal self was more accurate than did participantsin the inaccurate conditions (IA, II), Ms = 5.42 (SD =0.67) vs. 4.44 (SD = 1.11), t(117) = 5.78, p < .001,

d = 1.07 (t[69] = 5.74, p < .001, d = 1.38 forEuropean Americans and t[45] = 2.00, p = .05, d =0.60 for Asian Americans). Participants in the accuratecollective self conditions (AA, IA) indeed perceived thattheir interaction partner’s impression of their collectiveself was more accurate than did participants in the inac-curate conditions (AI, II), Ms = 5.36 (SD = 1.17) vs.2.33 (1.30), t(117) = 13.32, p < .001, d = 2.46 (t[69]= 9.34, p < .001, d = 2.25 for European Americans andt[45] = 9.44, p < .001, d = 2.81 for Asian Americans).

We went on to conduct a 3-way analysis of variance(ANOVA) with cultural group (European vs. AsianAmerican), personal self (accurate vs. inaccurate), andcollective self (accurate vs. inaccurate) as between-subject variables. As expected, participants in the accu-rate condition felt more positive affect (PA) than didthose in the inaccurate condition both in terms of thepersonal self, F(1, 109) = 5.61, p < .01, d = 0.45, andthe collective self, F(1, 109) = 14.86, p < .01, d = 0.74.There were no 2-way or 3-way interactions, Fs(1, 109)< 1.39, ps > .24, ds < .23. Also, European Americanand Asian American participants did not differ onPA overall across all the conditions, F(1, 109) = 0.62,p = .43, d = 0.15.

Some readers might be surprised not to see an interac-tion in the above analysis. This is not surprising, however,considering that we expected no cultural differences in theAA and II conditions (i.e., expected no cultural differencesin two of the four conditions). As Rosenthal, Rosnow,and Rubin (2000) argued, the omnibus ANOVA oftenmasks interesting interactions, and the focused hypothesisshould be tested with the focused contrast. Our hypothe-sis was that (a) there would be no cultural differences inthe AA and II conditions, (b) European Americans wouldbe happier than Asian Americans in the AI condition, and(c) Asian Americans would be happier than EuropeanAmericans in the IA condition. Thus, we tested the signif-icance of the following two orthogonal contrasts simulta-neously, following the guidelines of Rosenthal et al. ForEuropean Americans, the contrast was 1, –1, 1, –1for the AA, IA, AI, and II conditions, respectively. ForAsian Americans, the contrast was 1, 1, –1, –1, for theAA, IA, AI, and II conditions, respectively. Our hypothe-sis regarding the pattern of cultural differences across thefour experimental conditions was clearly supported, asthere was a significant effect of this set of contrasts, F(1,115) = 17.53, p < .001, d = 0.78 (see Figure 1). We alsotested sex differences and sex-by-contrasts interactions.There were no sex differences in overall PA, F(1, 115) =0.04, ns, d = 0.02. There were also no interactionsbetween sex and this set of contrasts, F(1, 115) = 2.25,p = .14, d = 0.42.

In addition to the contrast analysis presented above,the II and AA conditions provided meaningful tests for




our hypothesis. Specifically, by comparing the IA and AIconditions to the II condition, we were able to assess thedegree to which accurate perception of each aspect of theself was beneficial. Consistent with our prediction, accu-rate perception of the personal self was beneficial forEuropean Americans, as European Americans in the AIcondition felt significantly happier than did those in theII condition, t(36) = 2.22, p < .05, d = 0.74. In con-trast, accurate perception of the collective self did notbenefit European Americans, as there was no differencebetween the IA condition and the II condition amongEuropean Americans, t(34) = 1.28, ns, d = 0.44.Contrary to European Americans, accurate perceptionof the collective self tended to be beneficial for Asianparticipants, as Asian participants in the IA conditionwere marginally happier than were those in the II condi-tion, t(20) = 1.95, p = .066, d = 0.87. Again contrast-ing with European Americans, accurate perception ofthe personal self did not make Asian Americans happier,as there was no difference between AI and II conditionsamong Asian Americans, t(21) = 0.15, ns, d = 0.07.

Finally, we compared the AI and IA conditions to theAA condition for each cultural group to examine thedegree of detrimental impact of each type of inaccurateperception on PA. As predicted, inaccurate perception ofthe personal self was detrimental to European Americans’happiness, as European Americans in the IA conditionfelt significantly less positive about the feedback than didthose in the AA condition, t(31) = 2.20, p < .05, d =0.79. In contrast, inaccurate perception of the collectiveself was not harmful for European Americans as there wasno difference between AA and AI conditions, t(33) =1.78, p = .09, d = 0.62. Differing from EuropeanAmericans, inaccurate perception of the collective selfwas detrimental to Asian Americans’ happiness, as AsianAmerican participants in the AI condition felt significantlyless PA than did those in the AA condition, t(22) = 2.36,p < .05, d = 1.01. In contrast, inaccurate perception ofthe personal self did not have a damaging effect on thePA of Asian Americans, as there was no differencebetween the AA and IA conditions among AsianAmericans, t(21) = 0.72, ns, d = 0.31.


TABLE 1: Desirability Ratings of 26 Personality Traits for Asian and European Americans in Study 1b

Asian Americans European Americans

Trait M SD M SD t d

Hardworking 5.94 1.19 6.19 1.03 –0.97 –0.22Intelligent 6.00 1.02 6.22 0.87 –0.99 –0.23Fun-loving 6.19 0.97 6.25 0.88 –0.27 –0.06Friendly 6.36 0.82 6.31 0.82 0.26 0.06Stubborn 3.28 1.72 3.00 1.05 0.89 0.20Cooperative 6.04 1.02 5.78 0.94 1.15 0.26Relaxed 5.77 1.13 5.61 1.17 0.58 0.13Leader 4.96 1.57 5.31 1.20 –1.08 –0.25Emotional 4.39 1.64 4.13 1.24 0.82 0.19Rational 5.02 1.30 5.59 1.13 –2.03* –0.46Laid-back 5.06 1.63 5.56 1.16 –1.49 –0.34Carefree 4.98 1.42 4.47 1.34 1.60 0.36Playful 6.00 1.06 5.59 1.07 1.67 0.38Smart 6.17 0.99 6.19 1.00 –0.08 –0.02Assertive 5.30 1.33 5.28 0.99 0.06 0.01Agreeable 5.45 1.10 5.06 1.16 1.49 0.34Serious 4.66 1.37 4.48 1.00 0.66 0.15Logical 5.40 1.12 5.41 0.91 –0.01 –0.00Cautious 5.38 1.24 4.63 1.01 2.87** 0.66Open 6.15 1.00 6.00 1.08 0.63 0.14Passionate 6.04 0.91 6.09 0.96 –0.24 –0.05Independent 5.85 1.22 5.63 1.07 0.85 0.19Flexible 5.72 1.14 5.53 1.16 0.73 0.17Tough 5.06 1.39 4.72 1.51 1.05 0.24Warm 5.91 1.06 5.53 1.16 1.52 0.35Calm 5.81 1.12 5.63 1.36 0.66 0.15

NOTE: The first 10 traits were on the original list (in the self-description phase of Study 1a). The next 16 traits were used in the feedbackin Study 1a.*p < .05. **p < .01.



In sum, accurate perception of the personal self was acritical predictor of happiness for European Americans,whereas accurate perception of the collective self was acritical predictor of happiness for Asian Americans. Ourfindings suggest, for example, that European Americanswho think of themselves as laid-back would not be happyif their interaction partners perceived them as serious,whereas Asian Americans who think of themselves aslaid-back would be relatively unaffected if their interac-tion partners perceived them as serious. This might bebecause Asian Americans can think of the situations inwhich they were serious and see why their interactionpartners thought of them as serious (see Choi & Choi,2002, for empirical demonstration). In contrast, AsianAmericans for whom being psychology majors is animportant aspect of their collective selves would not behappy if their interaction partners perceived them to beeconomics majors, whereas European American psychol-ogy majors would not be affected if their interaction part-ners perceived them to be economics majors. Theseexamples illustrate important cultural differences in inter-personal conditions that lead to a sense of felt under-standing and happiness in social interactions.

In Study 1a we established cultural differences in theeffect of accurate perception of the personal and collec-tive selves on happiness in social interactions. However,one weakness was that the interaction consisted ofspending 20 to 30 minutes with a stranger and thus wasquite artificial. This leaves open the possibility that ourfindings are simply a product of this type of laboratoryinteraction rather than indicative of some general pat-tern in social interactions. The second weakness of thisstudy was that PA was measured in reaction to specificfeedback. In their daily lives, people often learn how

others perceive them through a third party (e.g., yourfriend lets you know that someone thinks you are a seri-ous person), and they affectively react to this informa-tion. However, we rarely receive direct feedback fromthe interaction partner. It is therefore important to exam-ine whether the findings from Study 1a can be generalizedto PA felt in daily social interactions. Finally, there is anartifactual explanation of the findings with regard to thepersonal self. If the personality traits used in this study weremore desirable for European American participants thanfor Asian Americans, this explains why the accuracy ofthe feedback regarding the personal self led to greater hap-piness among European Americans than among AsianAmericans. In terms of the collective self, however, wefound no differences in the type of group affiliations eitherlisted by participants or received in the false feedback.1

STUDY 1B

We conducted Study 1b to test the artifactual alter-native explanation for Study 1a by examining cross-cultural comparability of the traits that we used in thefirst study. Participants in Study 1a chose 2 traits thatdescribed them best out of 10 possible traits. For themanipulation of the accuracy or inaccuracy of the per-sonal self, we used 16 personality traits that were verysimilar or dissimilar to the original 10 traits. In total, weused 26 traits in Study 1a (see Table 1 for the entire list).Participants in Study 1b evaluated each of these traits interms of desirability.

Method

Participants

Eighty-eight students (36 male, 52 female) at CaliforniaState University, East Bay, participated in this study. Forty-seven of them identified themselves as Asian Americans, 32as European Americans, and 9 as Others.

Procedure

Participants were provided with 26 personality traits andasked to indicate the desirability of each trait on a 7-pointscale (1 = not at all desirable; 7 = extremely desirable).


Only 1 trait of the 10 traits in the original list showedany cultural difference in desirability (see Table 1).European Americans viewed rational as more desirablethan did Asian Americans, t(77) = 2.03, p < .05, d =0.46. In the actual feedback that participants received inStudy 1a, the term logical was used for those who chose


4

4.5

5

5.5

6

6.5

AA IA AI IIFeedback Condition

PA

Asian Euro

Figure 1 How happy and good participants felt by experimentalconditions and cultural groups in Study 1a.

NOTE: PA = positive affect; AA = accurate personal and collectiveselves feedback condition; IA = inaccurate personal, accurate collectiveselves condition; AI = accurate personal, inaccurate collective selvescondition; II = inaccurate personal and collective selves condition.



rational as self-descriptive. It is interesting that the termlogical did not differ in desirability across the twogroups, t(77) = 0.01, ns, d = 0.00. The only other traitthat differed in desirability was the term cautious,which was used in the inaccurate feedback for thosewho had chosen fun-loving as self-descriptive, t(77) =2.87, p < .05, d = 0.66. On average, Asian Americansviewed cautious as more desirable than did EuropeanAmericans. Considering that 15 of the 16 traits used inthe feedback showed no cultural differences in desir-ability, it is unlikely that the findings from Study 1a arebecause of cross-cultural differences in desirability.

We conducted another analysis, however, to directlytest the cross-cultural equivalence in the desirability ofthe personal self-feedback in Study 1a. We computed themean desirability ratings of the personality feedback foreach participant by replacing each individual rating withthe mean desirability ratings for the group. Specifically,we took the average desirability rating of each culturalgroup for each trait obtained in Study 1b, assigned thesegroup scores to each participant’s actual trait feedback inStudy 1a, then computed the desirability ratings for eachparticipant in Study 1a. For example, an Asian Americanparticipant who received open and playful as feedback inStudy 1a was given a score of 6.075 (because amongAsian Americans in Study 1b, open had a mean desir-ability rating of 6.15 and playful had a mean desirabilityrating of 6.00, and the average of these two scores is6.075), whereas a European American participant whoreceived laid-back and agreeable was given a score of5.31 (because among European Americans in Study 1b,laidback had a mean desirability rating of 5.56 and agree-able had a mean desirability rating of 5.06, and the aver-age of these scores is 5.31). It is most important that wedid not find any difference between the two groups in themean desirability ratings of the actual personality traits thatparticipants received in their feedback in Study 1a, AsianAmericans M = 5.51, SD = 0.34; European AmericansM = 5.41, SD = 0.48); t(67) = 0.93, ns, d = 0.23.

In sum, Study 1b demonstrated that the cultural dif-ference we observed in Study 1a with regard to the per-sonal self-feedback cannot be due to cultural differencesin the desirability of the personality traits that we usedin Study 1a. Once we eliminated the artifactual alterna-tive explanation for Study 1a, we went on to address theremaining issues from Study 1a in the next two studies.

STUDY 2: EVENT SAMPLING STUDY

We conducted Study 2 to address two limitations ofStudy 1a: (a) a contrived social interaction in the labo-ratory and (b) affective outcome specific to the feed-back. This time we used an event sampling method to

examine the relationship between the accurate percep-tion of the personal and collective selves and happinessin the context of natural, daily social interactions. Overa 2-week period, participants completed mood ratingsafter each naturally occurring social interaction thatlasted more than 10 minutes, and they indicated thedegree to which they felt understood by the interactionpartner in terms of their personal and collective selves.

Method

Participants

Participants were 107 students at the University ofMinnesota who responded to an ad in the student news-paper. Of the 107 original participants, 7 (6.5%) com-pleted less than 10 valid reports for the 2-week periodand were excluded from our analyses. Three (2.8%)additional individuals’ data were lost because they forgotto recharge the personal digital assistant (PDA) duringthe 2-week period. Thus, the final sample consisted of 97(87% of the original) participants, 56 (20 men, 33 women,and 3 did not provide this information) of whom identifiedthemselves as European American and 41 (22 men, 18women, and 1 did not provide this information) of whomidentified themselves as Asian or Asian American. Participantsreceived $25 upon completion of the study.

Materials and Procedure

Participants came to a research laboratory and metindividually with an experimenter. The experimentergave each participant a PDA that was programmed witha short survey and told participants to complete thissurvey each time they engaged in a social interactionthat lasted more than 10 minutes. Participants were toldto document their social interactions in this way for thenext 2 weeks. The experimenter defined a social inter-action as one that involves a face-to-face conversation,a shared activity (e.g., playing cards together), a phoneconversation, or an instant message conversation.Participants were also informed that all of their entrieswould be automatically time stamped.

Event sampling items. Participants indicated howhappy and pleasant they felt during the interactionusing a 7-point scale ranging from not at all (1) toextremely (7). We computed a PA score by taking theaverage of happy and pleasant (α = .82). Participantsthen reported the extent to which their interaction partnerunderstood their personal self (i.e., their “personality”and their “abilities and skills”; α = .88) and the extentto which the interaction partner understood their col-lective self (i.e., their “social and cultural background”and their “social roles and situations”; α = .86). In




addition, participants reported who the interactionpartner was from the following list: friend, roommate,romantic partner, family member, stranger, coworker,or other. Among the valid reports, on average 45.7% ofthe interactions were with a friend, 10.4% were with afamily member, 9.6% were with a roommate, 9.3% werewith a romantic partner, 9.3% were with a coworker,9.1% were with an other, and 6.6% were with a stranger.

In 2 weeks, the mean number of reports that participantscompleted was 57.62 (SD = 36.63). There were no culturalgroup differences in the number of reports, t(95) = 1.26, ns,d = 0.26. Female participants completed more reports thandid male participants, t(91) = 2.82, p < .01, d = 0.59.There was no culture-by-sex interaction in the number ofreports completed, F(1, 89) = 1.09, ns, d = 0.22.


We tested our hypotheses using hierarchical linearmodeling (HLM 5.04 program; Raudenbush, Bryk,Cheong, & Congdon, 2001) because our data consistedof two levels: within-person and between-person levels.The specific model that we tested was as follows:

Level 1: within-person

PA = β0 + β1 × Personal Self + β2 ×Collective Self + Error

Level 2: between-person

β0 = γ00 + γ01 × (Culture) + γ02 × (Sex) + u0β1 = γ10 + γ11 × (Culture) + γ12 × (Sex) + u1β2 = γ20 + γ21 × (Culture) + γ22 × (Sex) + u2,

where PA denotes positive affect, Personal Self denotesthe degree to which their personality and abilities andskills were understood by the interaction partner, andCollective Self denotes the degree to which their socialand cultural background and social roles and situationswere understood by the interaction partner. At Level 2,Culture was coded such that European American par-ticipants were 0 and Asian American participants were1, and Sex was coded such that male participants were0 and female participants were 1. Personal self and col-lective self scores were centered around each individ-ual’s mean. As can be seen in Table 2, there were no sexdifferences in the average PA, in the association betweenfelt understanding of the personal self and PA, or in theassociation between felt understanding of the collectiveself and PA, |t|s < 0.69, ps > .48, ds < 0.15.

Our hypothesis with regard to the personal self wassupported; European Americans’ PA was more stronglyassociated with the understanding of their personal self

than was Asians’, (γ11 = –.19, t = –3.51, p < .01, d =–0.73).2 As can be seen in Figure 2, a one-unit increase(i.e., 1 point increase in the 7-point scale) in the under-standing of the personal self was associated with a .45increase in PA among European American participants.In contrast, a one-unit increase in felt understanding ofthe personal self was associated with a .25 increase inPA among Asian American participants. Namely, whenEuropean American participants felt that their personal


TABLE 2: Cultural Differences in the Relations Between theAccurate Perception of Personal and Collective Selvesand Positive Affect in Study 2

Fixed Effect Coefficient (SE) t Ratio p Value

Intercept, β0

Intercept, γ00 5.13 (0.11) 44.71 .000Culture, γ01 –0.54 (0.13) –4.11 .000Sex, γ02 –0.07 (0.12) –0.58 .560

Personal self slope, β1

Intercept, γ10 0.45 (0.05) 9.63 .000Culture, γ11 –0.19 (0.05) –3.51 .001Sex, γ12 –0.03 (0.05) –0.69 .488

Collective self slope, β2

Intercept, γ20 –0.03 (0.05) –0.65 .514Culture, γ21 0.09 (0.05) 1.69 .091Sex, γ22 0.01 (0.05) 0.28 .779

NOTE: In Culture, European Americans were coded as 0, and AsianAmericans were coded as 1. Approximate df was 91 for this analysis.

Figure 2 Average within-person association between the under-standing of personal self and positive affect for EuropeanAmericans and Asian Americans in Study 2.

NOTE: –2, –1, +1, and +2 indicate the difference from participants’overall mean understanding of the personal self on a 7-point scale. PA= positive affect.



self was well understood by the interaction partner, theyalso felt substantially more PA than when they felt thattheir personal self was not well understood. AlthoughAsian Americans felt more PA when their personal selfwas well understood by the interaction partner thanwhen it was not well understood, the link between feltunderstanding of the personal self and PA was signifi-cantly weaker among them.

Our hypothesis concerning the collective self wasalso partially supported, as the understanding of thecollective self was positively associated with PA amongAsian Americans (β2 = .06), whereas it was not relatedto PA at all among European Americans (β2 = –.03;γ21 = 0.09, t = 1.69, p = .09, d = 0.35). As can be seenin Figure 3, the accurate understanding of the collectiveself was a marginally stronger predictor of happiness indaily social interactions among Asian Americans thanamong European Americans.

We next examined whether the patterns found abovewere generalizable across different types of interactions.Because specific types of interactions were limited innumber with the possible exception of interactions withfriends, it was not possible to repeat the above HLManalysis for each type of social interaction. Thus, wecreated two types of interactions out of the seven typesassessed: (a) close others (i.e., friend, romantic partner,family member, or roommate) and (b) distant others

(i.e., stranger, coworker, and other), and repeated theabove analysis. The HLM analysis on interactions onlywith close others again showed that EuropeanAmericans felt more PA than did Asian Americans whentheir personal self was accurately perceived by theirinteraction partner, γ11 = –.24 (SE = .06), t = –3.62,p < .01, d = 0.76, and Asian Americans felt more PAthan did European Americans when their collective selfwas accurately perceived by their interaction partner,γ21 = .15 (SE = .07), t = 2.24, p < .05, d = 0.47. TheHLM analysis on interactions only with distant othersalso showed that European Americans felt more PAthan did Asian Americans when their personal self wasaccurately perceived by their interaction partner, γ11 =–.28 (SE = .07), t = –3.75, p < .01, d = 0.79. AsianAmericans, however, did not feel more PA thanEuropean Americans did when their collective self wasaccurately perceived by a distant interaction partner,γ21 = .05 (SE = .08), t = 0.62, ns, d = 0.13.

By and large, Study 2 replicated the findings fromStudy 1a using a very different method. EuropeanAmerican participants felt happier and more pleasantthan Asian Americans did when their personal self wasaccurately perceived by their interaction partner. Incontrast, Asian Americans felt marginally happier thanEuropean Americans did when their collective self wasaccurately perceived by their interaction partner. Thefindings from Study 2 indicate that cultural differencesin the interpersonal conditions associated with happi-ness were not limited to a brief laboratory interactionwith a stranger as found in Study 1a or to reactions tothe specific impression feedback, but they were general-izable to the social interactions that college studentstypically encounter in their daily lives.

STUDY 3: A THREE-GROUP COMPARISON

Our main predictions were largely supported in botha laboratory and an event sampling study. In Studies 1and 2, however, we used a two-group comparison. Thisleaves alternative explanations for the main findings(Norenzayan & Heine, 2005). For instance, Asian andEuropean Americans are different not only in their cul-tural orientation toward the self (e.g., values and self-evaluation) but also in the minority-majority status. Assummarized in the introduction, African Americans’cultural orientation toward the self is on average moresimilar to European Americans’ than to AsianAmericans’ (e.g., Oyserman et al., 2002; Twenge &Crocker, 2002). Thus, if the cultural explanation isvalid, African Americans should show patterns of hap-piness similar to European Americans. In contrast, ifnoncultural factors, such as the minority status, underlie


Figure 3 Average within-person association between the under-standing of collective self and positive affect for EuropeanAmericans and Asian Americans in Study 2.

NOTE: –2, –1, +1, and +2 indicate the difference from participants’overall mean understanding of the collective self on a 7-point scale.PA = positive affect.



the findings from Studies 1 and 2, African Americansshould not show patterns of happiness similar toEuropean Americans. In sum, we conducted Study 3 to(a) further examine the replicability of the findings fromStudies 1 and 2 and (b) test the cultural account morerigorously using a three-group comparison method.

Method

Participants

Participants were 146 students at the University ofVirginia. Fifty-two of them (17 men, 35 women) self-identified as European Americans, 48 (10 men, 37women, one did not provide this information) self-iden-tified as African Americans, and 46 (12 men, 34women) self-identified as Asian Americans. Of the orig-inal 146 participants, 11 participants (7.5%) completedless than 10 valid reports and were excluded from ouranalyses. In addition, 10 participants’ data (6.8%) werelost because of participants’ errors (e.g., forgot torecharge PDA, broke or lost PDA) or the experimenter’serrors (e.g., overwrote the data). The final sampleincluded 124 participants: 41 European Americans (12men, 30 women), 41 African Americans (7 men, 34women), and 42 Asian Americans (9 men, 33 women).They were paid $25 for their participation.

Materials and procedure were exactly the same as inStudy 2. Cronbach’s alphas was .86 for PA, .85 for thepersonal self, and .90 for the collective self. On average,participants completed 36.74 reports (SD = 21.86).There were no differences in the number of reportscompleted among European, African, and AsianAmericans, F(2, 118) = 1.84, p = .16, d = 0.25. As inStudy 2, women completed more reports than did men,F(1, 118) = 4.85, p < .05, d = 0.41. There was nogroup-by-sex interaction in the number of reports com-pleted, F(2, 118) = .63, ns, d = 0.14. Among the validreports, on average, 48.9% of the interactions reportedwere with a friend, 14.5% were with a roommate,12.2% were with a family member, 12% were with aromantic partner, 4.5% were with a stranger, 4.4%were with an other, and 2.4% were with a coworker.


We tested our hypotheses again using HLM (5.04program). The specific model that we tested was as follows:

Level 1: within-person

PA = β0 + β1 × Personal Self + β2 × CollectiveSelf + Error

Level 2: between-person

β0 = γ00 + γ01 × (Sex) + γ02 × (Code 1)+ γ03 × (Code 2) + u0

β1 = γ10 + γ11 × (Sex) + γ12 × (Code 1)+ γ13 × (Code 2) + u1

β2 = γ20 + γ21 × (Sex) + γ22 × (Code 1)+ γ23 × (Code 2) + u2.

The Level 1 model was exactly the same as in Study2. At Level 2, the three groups were coded using dummycoding in which European American was the referencegroup. In Code 1, European and Asian Americans werecoded as 0 and African Americans were coded as 1. InCode 2, European and African Americans were codedas 0 and Asian Americans were coded as 1. Thus, Code1 indicates the difference between European Americansand African Americans, and Code 2 indicates thedifference between European Americans and AsianAmericans. In Sex, male participants were coded as 0and female participants were coded as 1. Personal selfand collective self scores were centered around eachindividual’s mean.

European Americans Versus Asian Americans

Replicating Study 2, the HLM analysis again showedthat European Americans’ PA was more strongly associ-ated with the understanding of the personal self thanwas Asians’, γ13 = –.14, t = –2.07, p < .05, d = 0.38(see Table 3). As can be seen in Figure 4, accurate per-ception of the personal self by an interaction partner wasmore strongly associated with European Americans’ PAthan with Asian Americans’ PA. Also replicating Study2, the understanding of the collective self was positivelyassociated with PA among Asians (β2 = .20), whereas itwas not related to PA among European Americans (β2 =.06; γ23 = .14, t = 1.92, p = .055, d = 0.36; see Figure5). As in Studies 1a and 2, therefore, accurate percep-tion of the collective self was marginally more associ-ated with Asian Americans’ PA than with EuropeanAmericans’ PA.

European Americans Versus African Americans

Whereas we found consistent differences betweenAsian Americans and European Americans, we did notfind a significant difference between African Americans(β1 = .45) and European Americans (β1 = .56) in thestrength of association between the understanding of




the personal self and PA, γ12 = –.11, t = –1.49, p =0.136, d = 0.28.3 Likewise, there were no differencesbetween African Americans (β2 = .11) and EuropeanAmericans (β2 = .06) in the association between theunderstanding of the collective self and PA, γ22 = .05,t = 0.69, p = .489, d = 0.13.

As in Study 2, we also examined whether the patternsof results we obtained above were specific to interac-tions with close others (i.e., friend, romantic partner,roommate, family member) by repeating the aboveHLM analysis. Consistent with Study 2, the under-standing of the personal self was again more stronglyassociated with PA among European Americans thanamong Asian Americans when we only examined thesocial interactions with close others, γ13 = –.15, t =–2.10, p < .05, d = 0.39. Furthermore, the understand-ing of the collective self was more strongly associatedwith PA among Asian Americans than among EuropeanAmericans when we analyzed only the social interac-tions with close others, γ23 = .15, t = 2.05, p < .05,d = 0.38. Consistent with the above analyses, therewere no differences between African Americans andEuropean Americans either in the strength of associa-tion between the understanding of the personal self andPA, γ12 = –.04, t = –0.53, p = .60, d = –0.09, or in theassociation between the understanding of the collectiveself and PA, γ22 = –.01, t = –0.12, p = .90, d = –0.02.We were unable to conduct the above HLM analysis


TABLE 3: Cultural Differences in the Relations Between theAccurate Perception of Personal and Collective Selvesand Positive Affect in Study 3

Fixed Effect Coefficient (SE) t Ratio p Value

Intercept, β0

Intercept, γ00 4.86 (0.14) 35.82 .000Sex, γ01 0.20 (0.14) 1.45 .146Code 1, γ02 –0.19 (0.14) –1.33 .184Code 2, γ03 –0.22 (0.14) –1.59 .111

Personal self slope, β1

Intercept, γ10 0.56 (0.08) 7.69 .000Sex, γ11 –0.11 (0.07) –1.50 .135Code 1, γ12 –0.11 (0.07) –1.49 .136Code 2, γ13 –0.14 (0.07) –2.07 .038

Collective self slope, β2

Intercept, γ20 0.06 (0.08) 0.77 .440Sex, γ21 –0.09 (0.07) –1.19 .235Code 1, γ22 0.05 (0.07) 0.69 .489Code 2, γ23 0.14 (0.07) 1.92 .055

NOTE: Sex = 0 for male, 1 for female participants. In dummy code1, European Americans and Asian Americans were coded as 0 andAfrican Americans were coded as 1. In dummy code 2, European andAfrican Americans were coded as 0 and Asian Americans were codedas 1. Thus, code 1 represents the comparison between EuropeanAmericans and African Americans, whereas code 2 represents thecomparison between European Americans and Asian Americans.Approximate df for this analysis was 117.

4

4.5

5

5.5

6

6.5

−2 −1 avg 1 2

Understanding of Personal Self

PA

EuroAm AfricanAm AsianAm

Figure 4 Average within-person association between the understandingof personal self and positive affect for European Americans,African Americans, and Asian Americans in Study 3.

NOTE: –2, –1, +1, and +2 indicate the difference from participants’overall mean understanding of the personal self on a 7-point scale.PA = positive affect.

Figure 5 Average within-person association between the understandingof collective self and positive affect for European Americans,African Americans, and Asian Americans in Study 3.

NOTE: –2, –1, +1, and +2 indicate the difference from participants’overall mean understanding of the collective self on a 7-point scale.PA = positive affect.



with only the interactions with distant others becausethere were only 505 valid event reports for 100 partici-pants in total (or 5 reports per person, which is not suf-ficient to obtain the Level 1 coefficients). This would beequivalent to running a multiple regression with 5 cases.

In sum, Study 3 replicated the main findings fromStudies 1a and 2, again using an event samplingmethod. European American participants felt happierthan Asian Americans did when their personal self wasaccurately perceived by their interaction partner. Incontrast, Asian Americans tended to feel happier thanEuropean Americans did when their collective self wasaccurately perceived by their interaction partner.Furthermore, the patterns of happiness felt in socialinteractions among African Americans were not differ-ent from those among European Americans.

GENERAL DISCUSSION

In three studies, we examined cultural differences ininterpersonal perceptions that are linked to happiness insocial interactions. Both in a laboratory study (Study 1a)and in event sampling studies (Studies 2 and 3), we foundthat European Americans felt happier than Asians didwhen their interaction partner perceived their personal selfaccurately, whereas Asian Americans felt happier thanEuropean Americans did when their interaction partnerperceived their collective self accurately. Furthermore, inStudy 3 we found no significant differences betweenAfrican Americans and European Americans in the linkbetween the understanding of specific aspects of the selfand happiness felt in social interactions.

These findings have several important implicationsfor research on well-being, social relationships, the self,and culture. First, although previous research hasshown a strong association between relationship qualityand well-being, most of the previous research of thistype has been correlational (see Myers, 1999, for areview). Thus, it was possible that satisfied people sim-ply viewed their relationships in a more positive light,and the factors that might lead to satisfaction in actualsocial interactions as well as the potential individual orcross-cultural differences in these factors remainedunclear. Study 1a demonstrated that the accurate per-ception of important aspects of the self has a causalimpact on happiness felt in social interactions.Moreover, it showed that the type of accurate percep-tion of the self that leads to happiness varies systemati-cally between European and Asian Americans.

Second, although many factors associated with cul-tural differences in well-being have been brought tolight in previous research (see Diener, Oishi, & Lucas,2003, for review), the vast majority of this research has

employed global self-reports of well-being. Previousresearch in this area has rarely examined how peopleactually felt and behaved in their daily lives or in natu-rally occurring social interactions (see Kitayama,Mesquita, & Karasawa, 2006; Mesquita & Karasawa,2002; Oishi et al., 2004, for exceptions). Thus, theprocesses underlying well-being and interpersonalunderstanding remained ambiguous until now. Thisresearch demonstrates that felt understanding of eitherthe personal or the collective self can be critical to theexperience of PA in social interactions, thereby delin-eating the specific interpersonal conditions that produceinterpersonal happiness. Our findings provide supportfor self-verification theory (Swann, 1990) and intimacymodel (Reis et al., 2004) in that responsiveness to impor-tant aspects of the self is of great import in interpersonalunderstanding and happiness felt in social interactions.We further extended these theories by demonstrating thatbecause important aspects of the self differ systematicallyacross cultures (Markus & Kitayama, 1991; Triandis,1989) the aspects of the self that need to be understoodby an interaction partner differ systematically acrosscultures as well.

Third, our main findings were replicated when weanalyzed only interactions with close others both in termsof the personal and the collective selves in Studies 2 and3. Considering that the interactions in Study 1a were withstrangers, it appears that the main interpersonalprocesses that we proposed apply to interactions withstrangers as well as to interactions with close others.However, one exception existed in that Asian Americanparticipants in Study 2 did not feel particularly happywhen distant others (i.e., strangers, coworkers, and oth-ers) accurately understood their collective selves. BecauseAsian Americans tend to make a sharper distinctionbetween close and distant others than do EuropeanAmericans (Oishi et al., 2004; Suh, 2002), the predictorsof happiness in social interactions might vary more radi-cally among Asian Americans than European Americans,depending on the type of interaction partner. This possi-bility needs to be clarified in the future.

Finally, our three-culture comparison strategy(Norenzayan & Heine, 2005) used in Study 3 clarifiedthat our main findings are not simply because of theminority status of Asian Americans in the United States.If the minority status per se was a driving force for thedifference between Asian Americans’ affective patternsand European Americans’, then African Americansshould have shown patterns similar to those of AsianAmericans. Although African Americans and AsianAmericans showed similar affective patterns in terms ofthe personal self (see Figure 4), they were quite differentin terms of the collective self (see Figure 5). In addition,research with Japanese in Japan (i.e., the majority group)




has shown a convergent pattern of results with our find-ings with Asian Americans (e.g., Hoshino-Browne et al.,2005, Study 2; Kitayama et al., 2006). For instance,Kitayama et al. (2006) showed that Japanese in Japanfelt happy when they also felt socially engaging emotions,which are presumably felt when an interaction partneraffirms the experiencer’s interdependence of the self.

The difference between Asian Americans and AfricanAmericans in the understanding of the collective selfcould be best attributed to sociohistorical factors. Onaverage, African Americans’ cultural background is his-torically devalued to a greater degree than is AsianAmericans’ cultural background in the United States.(e.g., Crocker & Major, 1989; Ogbu, 1978; Takaki,1989). These historical differences might be a reason whythe public and private aspects of collective self-esteem werepositively correlated among Asian Americans, whereasthey were not among African Americans (Crocker,Luhtanen, Blaine, & Broadnax, 1994). In a related vein,the stigmatized collective identity might also be responsi-ble for the stereotype threat effect found among AfricanAmericans (e.g., Steele & Aronson, 1995). Our findings,then, suggest that the understanding of the collective selfleads to happiness in social interactions when the collec-tive self is chronically accessible, personally important,and positively valued in society.

Before we reach our conclusions, the limitations ofour research should be addressed. First, all the traitsused in Study 1a, and most self-defining traits andgroup affiliations, are typically perceived to be positiveby participants. Thus, it is possible that participants inour studies felt that their interaction partner thoughthighly of them when they felt that their interaction part-ner accurately perceived their personally importantaspects of the self. In a sense, the effect we obtained ofaccurate understanding of important aspects of the selfon PA might be mediated by perceived positive regard.This needs to be clarified in future research. Second, ourfindings regarding the collective self were marginal inStudies 2 and 3. This might be because of imprecision inassessment of the collective self. It is important to har-ness the assessment of the collective self in the future.Third, we grouped together all self-identified AsianAmericans. As we develop this program of research inthe future, it will be important to determine whetherthese findings can be replicated using specific ethnicgroups (e.g., Chinese, Vietnamese), different genera-tions of Asian Americans, and various other types ofcultural groups (e.g., those who share a religious,sexual, or political orientation). Finally, we focused ona narrow quality of PA in our investigation. It is impor-tant to examine various qualities of PA (e.g., excitementvs. calmness; Oishi, Schimmack, & Colcombe, 2003;Tsai, Knutson, & Fung, 2006) in the future.

CONCLUSION

Despite some limitations, this study reveals consistentcultural differences between European Americans andAsian Americans in the interpersonal processes associ-ated with happiness. In conclusion, this research demon-strates that accurate perception of the personal self is keyto happiness in social interactions among EuropeanAmericans, whereas accurate perception of the collectiveself is critical to happiness in social interactions amongAsians. It is well known that social relationships areimportant building blocks of a general sense of well-being(Ryan & Deci, 2000; Ryff & Singer, 1998). Consideringthat affective outcomes of social interactions are associ-ated with the continuation and maintenance of relation-ships (Elliott & James, 1989), these outcomes are likelyto be associated with a general sense of well-being aswell. Our findings, then, indicate that the affective basisof relationship development and maintenance differsacross cultures and that the interpersonal conditions thatare conducive to a general sense of well-being might bedifferent across cultures. In the future, it will be importantto extend this research to various relationship contexts,including husband-wife, client-therapist, student-teacher,and employee-supervisor. These types of investigationswill deepen people’s understanding of the link betweeninterpersonal processes and happiness and will help themdevelop ways to improve many interactions and relation-ships in their daily lives.

NOTES

1. We classified group affiliations into student organization, ath-letic, Greek, arts, religious, volunteer, or ethnic group, and wecounted the number of each type of group. There were no differencesin any type of group affiliation either listed by participants themselves(|t|s < 1.40, ps > .17) or received in feedback (|t|s < 1.76, ps > .08).

2. To our knowledge, a consensus has not been reached regardingthe appropriate effect size for multilevel analysis (Roberts & Monaco,2006). To provide some idea regarding the magnitude of the effectsfrom the hierarchical linear modeling analyses reported in Studies 2and 3, however, we converted t-values to rs, using formula 2.5 inRosenthal, Rosnow, and Rubin (2000) and then converted rs to dsusing formula 2.14. Because hierarchical linear modeling analysis pro-vides only approximate dfs, we used the approximate df in place of dfin the formula 2.5. This might result in some bias in the estimation.

3. The degree of association between the understanding of the per-sonal self and positive affect was very similar between AfricanAmericans and Asian Americans, γ12 = .03, t = 0.48, p = .63, d =0.10. The degree of association between the understanding of the col-lective self and positive affect was also not significantly differentbetween these two groups, γ22 = –.08, t = –1.11, p = .27, d = 0.25.

REFERENCES

Adams, G., & Plaut, V. C. (2003). The cultural grounding of personalrelationship: Friendship in North American and West Africanworlds. Personal Relationships, 10, 333-347.




Cantor, N. (1994). Life task problem solving: Situational affordancesand personal needs. Personality and Social Psychology Bulletin,20, 235-243.

Chen, S., Chen, K. Y., & Shaw, L. (2004). Self-verification motives atthe collective level of self-definition. Journal of Personality andSocial Psychology, 86, 77-94.

Choi, I., & Choi, Y. (2002). Culture and self-concept flexibility.Personality and Social Psychology Bulletin, 28, 1508-1517.

Cousins, S. D. (1989). Culture and self-perception in Japan and theUnited States. Journal of Personality and Social Psychology, 56,124-131.

Crocker, J., Luhtanen, R., Blaine, B., & Broadnax, S. (1994).Collective self-esteem and psychological well-being among White,Black, and Asian college students. Personality and SocialPsychology Bulletin, 20, 503-513.

Crocker, J., & Major, B. (1989). Social stigma and self-esteem: Theself-protective properties of stigma. Psychological Review, 96,608-630.

Diener, E., Oishi, S., & Lucas, R. E. (2003). Personality, culture, andsubjective well-being: Emotional and cognitive evaluations of life.Annual Review of Psychology, 54, 403-425.

Elliott, R., & James, E. (1989). Varieties of client experience in psy-chotherapy: An analysis of the literature. Clinical PsychologyReview, 9, 443-467.

Heine, S. J., & Renshaw, K. (2002). Interjudge agreement, self-enhancement, and liking: Cross-cultural divergences. Personalityand Social Psychology Bulletin, 28, 578-587.

Hoshino-Browne, E., Zanna, A. S., Spencer, S. J., Zanna, M. P.,Kitayama, S., & Lackenbauer, S. (2005). On the cultural guises ofcognitive dissonance: The case of Easterners and Westerners.Journal of Personality and Social Psychology, 89, 294-310.

Kanagawa, C., Cross, S. E., & Markus, H. R. (2001). “Who am I?”The cultural psychology of the conceptual self. Personality andSocial Psychology Bulletin, 27, 90-103.

Kitayama, S., Mesquita, B., & Karasawa, M. (2006). Cultural affor-dances and emotional experience: Socially engaging and disengag-ing emotions in Japan and the United States. Journal ofPersonality and Social Psychology, 91, 890-903.

Kwan, V. S. Y., Bond, M. H., & Singelis, T. M. (1997). Panculturalexplanations for life satisfaction: Adding relationship harmony toself-esteem. Journal of Personality and Social Psychology, 73,1038-1051.

Levine, R., Sato, S., Hashimoto, T., & Verma, J. (1995). Love andmarriage in eleven cultures. Journal of Cross-Cultural Psychology,26, 554-571.

Markus, H. R., & Kitayama, S. (1991). Culture and the self:Implications for cognition, emotion, and motivation.Psychological Review, 98, 224-253.

Mesquita, B., & Karasawa, M. (2002). Different emotional lives.Cognition and Emotion, 16, 127-141.

Murray, S. L., Holmes, J. G., Bellavia, G., Griffin, D. W., &Dolderman, D. (2002). Kindred spirits? The benefits of egocen-trism in close relationships. Journal of Personality and SocialPsychology, 82, 563-581.

Myers, D. G. (1999). Close relationships and quality of life. InD. Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: Thefoundations of hedonic psychology (pp. 374-391). New York:Russell Sage.

Norenzayan, A., & Heine, S. J. (2005). Psychological universals:What are they and how can we know? Psychological Bulletin,131, 763-784.

Ogbu, J. U. (1978). Minority education and caste: The Americansystem in cross-cultural perspective. New York: Academic Press.

Oishi, S., Diener, E., Scollon, C. N., & Biswas-Diener, R. (2004).Cross-situational consistency of affective experiences across cul-tures. Journal of Personality and Social Psychology, 86, 460-472.

Oishi, S., Lun, J., & Sherman, G. D. (2007). Residential mobility, self-concept, and positive affect in social interactions. Journal ofPersonality and Social Psychology, 93, 131-141.

Oishi, S., Schimmack, U., & Colcombe, S. (2003). The contextual andsystematic nature of life satisfaction judgments. Journal ofExperimental Social Psychology, 39, 232-247.

Oishi, S., Wyer, R. S., Jr., & Colcombe, S. (2000). Cultural variationin the use of current life satisfaction to predict the future. Journalof Personality and Social Psychology, 78, 434-445.

Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinkingindividualism and collectivism: Evaluation of theoretical assump-tions and meta-analyses. Psychological Bulletin, 128, 3-72.

Raudenbush, S., Bryk, A., Cheong, Y.-F., & Congdon, R. (2001).HLM5: Hierarchical linear and nonlinear modeling. Lincolnwood,IL: SSI Scientific Software International.

Reis, H. T., Clark, M. S., & Holmes, J. G. (2004). Perceived partnerresponsiveness as an organizing construct in the study of intimacyand closeness. In D. J. Mashek, & A. P. Aron (Eds.), Handbook ofcloseness and intimacy (pp. 201-225). Mahwah, NJ: LawrenceErlbaum.

Reis, H. T., & Gable, S. L. (2000). Event-sampling and other methodsfor studying everyday experience. In H. T. Reis & C. M. Judd (Eds.),Handbook of research methods in social and personality psychology(pp. 190-222). Cambridge, UK: Cambridge University Press.

Reis, H. T., & Patrick, B. C. (1996). Attachment and intimacy:Component processes. In E. T. Higgins & A. Kruglanski (Eds.),Social psychology: Handbook of basic principles (pp. 523-563).New York: Guilford.

Roberts, J. K., & Monaco, J. P. (2006, April). Effect size measures forthe two-level linear multilevel model. Paper presented at the annualmeeting of the American Educational Research Association, SanFrancisco.

Rosenthal, R., Rosnow, R. L., & Rubin, D. B. (2000). Contrasts andeffect sizes in behavioral research: A correlational approach. NewYork: Cambridge University Press.

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facil-itation of intrinsic motivation, social development, and well-being.American Psychologist, 55, 68-78.

Ryff, C. D., & Singer, B. (1998). The contours of positive humanhealth. Psychological Inquiry, 9, 1-28.

Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intel-lectual test performance of African Americans. Journal ofPersonality and Social Psychology 69, 797-811.

Suh, E. M. (1999). Culture, identity consistency, and subjective well-being. Unpublished dissertation, University of Illinois, Urbana-Champaign.

Suh, E. M. (2002). Culture, identity consistency, and subjective well-being. Journal of Personality and Social Psychology, 83, 1378-1391.

Swann, W. B., Jr. (1990). To be adored or to be known: The interplayof self-enhancement and self-verification. In R. M. Sorrentino &E. T. Higgins (Eds.), Handbook of motivation and cognition:Foundations of social behavior (Vol. 2., pp. 408-448). New York:Guilford.

Takaki, R. (1989). Strangers from a different shore: A history ofAsian Americans. New York: Penguin.

Taylor, S. E., Sherman, D. K., Kim, H. S., Jarcho, J., Takagi, K., &Dunagan, M. S. (2004). Culture and social support: Who seeksit and why? Journal of Personality and Social Psychology, 87,354-362.

Triandis, H. C. (1989). The self and social behavior in differing cul-tural contexts. Psychological Review, 96, 506-520.

Tsai, J. L., Knutson, B. K., & Fung, H. H. (2006). Cultural variationin affect valuation. Journal of Personality and Social Psychology,90, 288-307.

Twenge, J., & Crocker, J. (2002). Race, ethnicity, and self-esteem:Meta-analyses comparing Whites, Blacks, Hispanics, Asians, andNative Americans, including a commentary on Gray-Little andHafdahl (2000). Psychological Bulletin, 128, 371-408.

Wheeler, L., Reis, H. T., & Bond, M. H. (1989). Collectivism-indi-vidualism in everyday social life: The middle kingdom and themelting pot. Journal of Personality and Social Psychology, 57,79-96.

Zuckerman, M. (1979). Attribution of success and failure revisited:Or the motivational bias is alive and well in attribution theory.Journal of Personality, 47, 245-287.

Received August 24, 2006Revision accepted August 31, 2007




Culture, Interpersonal Perceptions, and Happiness in Social Interactions

Documents