personality: the universal and the culturally specificheine/docs/annual review.pdfself-enhancement,...

ANRV364-PS60-14 ARI 27 October 2008 16:18

Personality: The Universaland the Culturally SpecificSteven J. Heine and Emma E. BuchtelDepartment of Psychology, University of British Columbia, Vancouver, British Columbia,V6T 1Z4 Canada; email: [email protected], [email protected]

Annu. Rev. Psychol. 2009. 60:36994

The Annual Review of Psychology is online atpsych.annualreviews.org

This articles doi:10.1146/annurev.psych.60.110707.163655

Copyright c 2009 by Annual Reviews.All rights reserved

0066-4308/09/0110-0369$20.00

Key Wordsself-enhancement, cultural equilibria, Big 5, personality utility

AbstractThere appears to be a universal desire to understand individual differ-ences. This common desire exhibits both universal and culturally spe-cific features. Motivations to view oneself positively differ substantiallyacross cultural contexts, as do a number of other variables that covarywith this motivation (i.e., approach-avoidance motivations, internal-external frames of reference, independent-interdependent views of self,incremental-entity theories of abilities, dialectical self-views, and re-lational mobility). The structure of personality traits, particularly thefive-factor model of personality, emerges quite consistently across cul-tures, with some key variations noted when the structure is drawn fromindigenous traits in other languages. The extent to which each of theBig 5 traits is endorsed in each culture varies considerably, althoughwe note some methodological challenges with comparing personalitytraits across cultures. Finally, although people everywhere can con-ceive of each other in terms of personality traits, people in collectivisticcultures appear to rely on traits to a lesser degree when understand-ing themselves and others, compared with those from individualisticcultures.

369

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


ContentsINTRODUCTION . . . . . . . . . . . . . . . . . . 370THE EVALUATION OF THE SELF

ACROSS CULTURES . . . . . . . . . . . . 371What Processes Are Implicated

in Cultural Variation inSelf-EnhancementMotivations?. . . . . . . . . . . . . . . . . . . . 372

STRUCTURE AND CONTENTOF PERSONALITY ACROSSCULTURES . . . . . . . . . . . . . . . . . . . . . . 377Apparent Near Universality

of Personality Structure . . . . . . . . . 377Cross-Cultural Variability in Levels

of Personality Traits . . . . . . . . . . . . . 379THE UTILITY OF PERSONALITY

ACROSS CULTURES . . . . . . . . . . . . 382Content of the Self-Concept . . . . . . . 383Incremental Versus Entity

Theories of Self . . . . . . . . . . . . . . . . . 383Perceived Consistency of Traits . . . . . 383Attributions for Behavior . . . . . . . . . . . 384Spontaneous Trait Inferences . . . . . . . 385Personality Traits and Behavior . . . . . 386

CONCLUSION . . . . . . . . . . . . . . . . . . . . . 387

INTRODUCTIONIn previous decades, the study of culture waslargely limited to the work of anthropologists,who mainly sought evidence for culture in peo-ples social environments. More recently, thestudy of culture has also been taken up by psy-chologists, who primarily look for evidence ofculture in the person. These two complemen-tary efforts to understand the nature of culturalbeings have been fused in the field of culturalpsychology, which hinges on the assumptionthat personality and culture are mutually con-stituted (see Heine 2008, Shweder 1990). Thatis, one cannot fully understand the nature ofpeople without considering the cultural con-text within which they exist; nor can one fullyunderstand a cultural context without consid-ering the values and beliefs of the people who

inhabit it. Cultural psychologists seek to under-stand people as they are embedded within theircultures.

Over the past two decades, much cul-tural psychological research has revealed pro-nounced cultural variation in many psycholog-ical processes that were hitherto assumed tobe universal, such as the fundamental attribu-tion error (Choi et al. 1999) and preferencesfor choice (Iyengar & Lepper 1999). This cul-tural variation has important implications forstudying psychology across cultures. The studyof psychology in general and of personality inparticular has largely been guided by Westernresearch. For example, 92% of publications inthe Journal of Personality and Social Psychologyare from authors at North American institu-tions, and 99% are from authors at Westernschools (Quinones-Vidal et al. 2004). The nar-rowness of the sample upon which most per-sonality research has been conducted raises im-portant questions about the generalizability ofthis research (see Arnett 2008, Henrich et al.2008). Much cross-cultural personality researchhas been conducted to address these questions(for recent reviews, see Benet-Martinez 2007,Diener et al. 2003, Triandis & Suh 2002).

Personality psychology has been conceptu-alized by some as the study of human nature(e.g., Buss 1984). In this respect there is no bet-ter topic in psychology in which to investigatethe role of culture, as the nature of humansis very much that of a cultural species (Heine& Norenzayan 2006, Tomasello 1999). A keyquestion to consider is how cultural learningcomes to shape the ways that people under-stand themselves and others. In this article, weexplore the relation between culture and per-sonality by reviewing cross-cultural research in(a) how people evaluate themselves, (b) thestructure and content of personality across cul-tures, and (c) the utility of personality in-formation across cultures. There appears tobe a universal desire to understand individualdifferencesthat is, personality (Funder 2007).But culture has a large role to play in how weuse and understand information about individ-ual differences. In this review, we pay particular

370 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


attention to evidence suggesting universality orcultural variability of these different aspects ofpersonality and to describing how culture influ-ences individual differences.

THE EVALUATION OF THE SELFACROSS CULTURESThe mutual constitution of person and culturebecomes especially evident in the explorationof how people evaluate themselves across cul-tures, such as by considering trait-level self-esteem. That people are motivated to viewthemselves positively is one of the most deeplyheld assumptions about the self (Maslow 1943,Tesser 1988). However, much research revealsstrong variation in the strength of this moti-vation across cultures. For example, studies ofMexican-Americans (Tropp & Wright 2003),Native Americans (Fryberg & Markus 2003),and Bangladeshis (Schmitt & Allik 2005) revealsignificantly less positive self-views than thosefound in studies conducted with Westerners.

In particular, cross-cultural research findsthat East Asians evince far less motivation forself-enhancement than do Westerners. In a re-cently published meta-analysis, across 91 cross-cultural comparisons using 30 different meth-ods, the Western samples self-enhanced morethan the East Asian samples by an average effectsize of d = 0.84 (Heine & Hamamura 2007).Analyses within cultures of self-enhancementbiases (another indicator of motivation for self-esteem) also reveal striking differences. AmongWestern samples, the average effect size of self-enhancing biases was d = 0.87, a strong ef-fect that was evident in all 14 of the methodsthat were used; in contrast, for East Asians,the average effect was d = 0.01 (Heine &Hamamura 2007). Moreover, the methods thatdid yield a positive self-enhancing effect forEast Asians (i.e., those where people comparethemselves to the average other) appear tohave been largely driven by a methodologi-cal artifact: the everyone is better than theirgroups average effect (Klar & Giladi 1997; seeHamamura et al. 2007). Cultural differencesin self-enhancement between East Asians and

Self-enhancementbiases: the tendencyto evaluate the selfmore positively thanothers

Westerners are thus large and consistentlyfound across diverse methods.

These cultural differences in motivationsfor self-enhancement are not easily accountedfor by alternative explanations such as (a) EastAsians being motivated to esteem their groupsrather than their individual selves (much re-search finds that Westerners also evaluate theirgroups more positively than do East Asians;Crocker et al. 1994, Heine & Lehman 1997,Snibbe et al. 2003); (b) East Asians enhancethemselves in domains that are of most impor-tance to them [the most extensive meta-analysison this topic finds no correlation between self-enhancement and importance for East Asians,r = 0.01, in contrast to a positive correlationfor Westerners, r = 0.18 (Heine et al. 2007a),but see discussion regarding whether studiesshould be excluded from this meta-analysis(Heine et al. 2007b; Sedikides et al. 2007a,b)];and (c) East Asians are presenting themselvesself-critically, but are privately evaluating them-selves in a self-enhancing manner [the culturaldifferences are similarly pronounced with stud-ies using hidden behavioral measures (Heineet al. 2000, 2001), although the cultural dif-ferences are largely absent for measures of im-plicit self-esteem (Kitayama & Uchida 2003,Kobayashi & Greenwald 2003)]. These find-ings have led some to conclude that motiva-tions for high self-esteem are far weaker, if notlargely absent, among East Asians than amongWesterners (e.g., Heine et al. 1999).

In support of this conclusion, some researchfinds that positive assessments of ones self ap-pear to be of less utility for East Asians than forWesterners. A number of studies find that pos-itive self-views are less correlated with subjec-tive well-being (Diener & Diener 1995, Kwanet al. 1997), self-concept clarity (Campbellet al. 1996), and depression (Heine & Lehman1999) in East Asia than they are in NorthAmerica. Moreover, whereas experimentallymanipulated positive self-views lead to en-hanced persistence among North Americans,such manipulations lead to less persistenceamong East Asians (Heine et al. 2001). In sum-mary, positive self-views appear to be associated

www.annualreviews.org Personality 371

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


with fewer positive consequences among EastAsians than among Westerners.

What Processes Are Implicatedin Cultural Variation inSelf-Enhancement Motivations?Why are self-enhancement motivations sucha salient and important feature of Westernpersonalities, but not of East Asians? Thatpronounced cultural differences between West-erners and East Asians in self-enhancing mo-tivations emerge so consistently across diversemethods raises the question of why these cul-tural differences exist. One way to assess thiskind of question is to consider the psychologicalprocesses that relate to the cultural differences.Thus far, in an effort to make sense of the ob-served cultural variation in self-enhancing mo-tivations, several different processes have beenexplored and assessed.

Approach-avoidance motivation. One rel-evant process contributing to the culturaldifferences in self-enhancing motivations isapproach-avoidance motivation. Approach mo-tivation focuses on advancement, accomplish-ments, and aspirations; it involves a concernwith the presence or absence of positive out-comes. In contrast, avoidance motivation fo-cuses on safety, responsibilities, and obligations;it is concerned with the presence or absence ofnegative outcomes (Higgins 1996).

There is much evidence that East Asians dif-fer from Westerners in the extent to which theyshow approach and avoidance motivations. Ingeneral, various studies find that in comparisonwith Westerners, East Asians show relativelymore evidence for avoidance motivation andrelatively less evidence for approach motiva-tion. For example, in comparison with NorthAmericans, East Asians embrace more personalavoidance goals (Elliot et al. 2001), rate oppor-tunities to lose as more important than oppor-tunities to win (Lee et al. 2000), persist more ona task after failure and less after success (Heineet al. 2001), and are motivated more by negativerole modelssomeone that people want to

ensure they do not become like (Lockwoodet al. 2005). Furthermore, this cultural differ-ence is evident in the ways that people processinformation: East Asians have been shownto have better memory for details regardingopportunities for losses than for opportunitiesfor gains (Aaker & Lee 2001), they recall eventsbetter if they contain prevention information,and they view book reviews to be more helpful ifthose reviews contain prevention information(Hamamura et al. 2008b). These reliablyobserved cultural differences in approach-avoidance motivation have been proposed tobe the result of the different kinds of positiveself-views (i.e., self-esteem and face) that areprioritized by Westerners and East Asians,respectively (see Hamamura & Heine 2008,Heine 2005).

Internal versus external frame of reference.Another mechanism that is implicated in cul-tural variation in self-enhancing motivationsis the perspective of the evaluator. In evaluat-ing themselves, people can attend to whetherthey are meeting their own internal standardsof competence (i.e., I think Im doing well), orthey can attend to whether they are meetingother peoples standards of competence (i.e.,others think Im doing well). Although thesetwo orientations are not independent, as peo-ples evaluations of themselves are influencedby their assessments of how they are meetingothers standards (Leary & Baumeister 2000),people can vary in the extent to which theymore closely attend to their own or to othersstandards. Elsewhere, we propose that a con-cern with maintaining face leads East Asiansto attend more to the standards of others whenevaluating themselves, whereas a concern withenhancing self-esteem leads Westerners to at-tend more to their own internal standards (seeHeine 2005, Heine et al. 2008b). This reason-ing suggests that East Asians should pay closerattention to the perspective of others than doWesterners.

There is much recent evidence for this cul-tural difference in perspective taking (for a re-view, see Cohen et al. 2007). For example,

372 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


Cohen & Gunz (2002) demonstrated that incomparison with Westerners, East Asians aremore likely to recall memories of themselveswhen they were at the center of attention froma third-person perspective. Apparently, EastAsians attention to an audience leaked into anddistorted their memories of themselves. Simi-larly, East Asians outperformed Westerners ona visual perspective-taking task, making fewervisual fixations on objects that were not visibleto a person who was giving instructions to them(Wu & Keysar 2007).

Cross-cultural research on self-awarenessalso identifies cultural divergences in framesof reference. When individuals are aware ofhow they appear to others, they are said to bein the state of objective self-awareness (Duval& Wicklund 1972), and this leads to a num-ber of predictable responses (e.g., people be-come more self-critical and are less likely to en-gage in counter-normative behaviors; Diener &Wallbom 1976, Fejfar & Hoyle 2000). In a stateof objective self-awareness, people are aware ofhow they appear as an object (a me) in con-trast to the experience of being a subject (anI). To the extent that East Asians are aware ofan audience and adjust their behaviors to thataudience, they would more likely be in a habit-ual state of objective self-awareness than wouldNorth Americans. If this is the case, then stim-uli that enhance objective self-awareness (forexample, seeing oneself in front of a mirror)should have little effect on East Asians. Evenwithout a mirror present, East Asians should beconsidering themselves in terms of how theyappear to others. Some recent cross-culturalresearch corroborates this hypothesis: whereasNorth Americans were more self-critical andwere less likely to cheat on a test when a mir-ror was present compared to when it was ab-sent, the presence of a mirror had no effect onJapanese for either dependent variable (Heineet al. 2008b). Moreover, although North Amer-ican self-evaluations were much more positivethan Japanese when the mirror was not present,they were at relatively similar levels to Japanesewhen they were in front of the mirror. One rea-son that self-evaluations tend to be so much

more positive for North Americans than forJapanese may be that North Americans are lesslikely to consider how they appear to others.Objectivity constrains the ability to maintain apositive self-view.

Independent versus interdependent viewsof self. Cultural variation in self-enhancementcan also be better understood when consideringthe kinds of self-concepts that are most com-mon in various cultures. One way of consider-ing the self is to see it as a relatively autonomous,self-sustaining collection of attributes that islargely independent from others. This inde-pendent view of self is more common inWestern cultures and has been the workingmodel for many of the theories of self that havebeen developed by a Western-dominated socialpsychology. In contrast, a second way of con-struing selves is to see them as being fundamen-tally interconnected, situationally variable, andgrounded in roles and relationships with signifi-cant ingroup others. This interdependent viewof self is more common in non-Western cul-tures and has been linked to a wide array of dis-tinct phenomena (for reviews, see Heine 2001,Markus & Kitayama 1991, Triandis 1989).

Measures of self-esteem and self-enhancingbiases tend to be positively associated with inde-pendence and negatively associated with inter-dependence (although these latter correlationstend to be weaker), regardless of the culture thathas been investigated (Heine et al. 1999, Heine& Renshaw 2002, Oyserman et al. 2002). Oneway to account for these correlations is to con-sider the consequences of elaborating a posi-tive self-view. Self-enhancement is associatedwith both costs and benefits to the individual.Paulhus (1998) makes the case that these ben-efits and costs are realized in two different do-mains. First, benefits of self-enhancement tendto be intrapsychic in nature. That is, focus-ing on what is good about the self tends tobe associated with subjective well-being andself-efficacy and is negatively associated withdysphoria and depression (Taylor & Armor1996, Taylor & Brown 1988). One clear ben-efit of self-enhancing, then, is that it feels good.


Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


However, the intrapsychic benefits that derivefrom self-enhancement come at the expense ofones relationships. A number of researchershave highlighted how self-enhancers risk at-tracting the scorn of those around them (Colvinet al. 1995, Paulhus 1998, Vohs & Heatherton2001; for a contrary view, see Taylor et al. 2003).To put it simply, most people do not partic-ularly like self-enhancers. These interpersonalcosts are especially evident in long-term rela-tionships (Robins & Beer 2001), the kinds ofrelationships that are particularly implicated ininterdependent selves (Adams 2005).

The costs and benefits of self-enhancementin these two domains suggest that to the extentan individuals culture prioritizes intrapsychicover interpersonal concerns, self-enhancementwould be a beneficial strategy. The positive feel-ings that arise from self-enhancement will beseen as worth the price of the alienation ofthose around one. In contrast, to the extentthat an individuals culture emphasizes interper-sonal relationships over intrapsychic rewards,self-improvement and face maintenance shouldbe a more beneficial strategy. The benefits ofdeepening relations with others outweigh thecosts of the negative feelings associated withself-improvement. There is much evidence thatpeople in Western cultures are more concernedwith positive feelings than are people in EastAsian cultures (Diener et al. 1995, Kitayamaet al. 2000, Mesquita & Karasawa 2002), andthat people in East Asian cultures are moreconcerned with maintaining interpersonal har-mony than are people in Western cultures(Morling et al. 2002, Suh et al. 1998). This evi-dence suggests that the cost-benefit ratio of self-enhancing is not as favorable for East Asians asit is for North Americans.

Incremental versus entity theories of abil-ities. The value of self-enhancement also de-pends on the lay theories that people hold aboutthe nature of abilities. One way to conceive ofabilities is to view them as arising from a set ofrelatively fixed and innate attributes. This kindof entity theory (Dweck & Leggett 1988) ofabilities reflects beliefs in an underlying essence

that is tied to abilities. Within such a worldview,an individuals successes and failures directly re-flect upon his or her perceived capabilities andself-worth. To the extent that abilities are per-ceived to be largely immutable and reflectingessential aspects of the individual, having a pos-itive assessment of ones abilities would be ac-companied by subjective well-being and wouldprovide the individual with the requisite confi-dence to perform at his or her best on a task.Viewing ones abilities negatively, on the otherhand, would seem to be tied closely to depres-sion and would decrease any motivation to im-prove. There would be little reason to try harderif ones failures were perceived to be immutable(Dweck 1999).

A second way of conceiving of abilities is toview them as being malleable and ultimately im-provable. This kind of incremental theory ofabilities reflects a belief in the key role of ef-fort in abilities. Within this worldview, ratherthan successes and failures being diagnostic ofones capabilities and self-worth, they are in-stead perceived as revealing the extent of onesefforts. Doing poorly on a task does not indi-cate that one is lacking the potential, but ratherthat one needs to direct additional effort to im-provement. This suggests that those with incre-mental views of abilities should not find failuresas painful, or successes as pleasant, as those withentity theories, and hence performance on tasksshould be less tied to their self-esteem.

Cultural differences in entity and incremen-tal theories of abilities parallel those of self-enhancement motivations. For example, a num-ber of studies have identified greater tendenciesfor East Asians compared with North Amer-icans to attribute school achievement to ef-fort and not to abilities (e.g., Holloway 1988,Stevenson & Stigler 1992; but see mixed evi-dence on cultural comparisons of Likert scalemeasures of malleability, e.g., Heine et al.2001, Hong et al. 1999, Norenzayan et al.2002). Likewise, experimental manipulationsof incremental theories of abilities corroboratethe cultural differences. Japanese come to re-spond to failure in a way similar to Americanswhen entity-theories are primed, whereas

374 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


Americans come to respond to failure in theways Japanese do when incremental theories areprimed (Heine et al. 2001). It appears that an-other reason cultures differ in the positivity oftheir self-views is the cultural variation in laytheories of abilities.

Dialectical reasoning about the self. Cul-tural variation in self-enhancing motivationscan be understood in yet another way: EastAsian and Western cultures differ in their toler-ance for contradiction (Peng & Nisbett 1999).That is, whereas Westerners typically respondto contradictory statements by trying to dis-miss or transcend the contradiction, East Asiansare more content to accept the contradictionsas they are. The tendency to perceive andtolerate psychological contradiction has beentermed nave dialecticism (Peng & Nisbett1999). This cultural difference in attitudes to-ward contradiction is not limited to how peopleperceive contradictory logical arguments aboutthe world; the difference also generalizes tohow people view themselves. When describingthemselves, East Asians maintain more contra-dictory self-views than do Westerners. For ex-ample, compared with Westerners, East Asiansare more likely to endorse opposing statementsabout their personalities (e.g., they accept state-ments regarding being both introverted andextraverted; Choi & Choi 2002, Hamamuraet al. 2008a), they acknowledge experiencingpositive and negative affective states more si-multaneously (Bagozzi et al. 1999), they viewthemselves as acting less consistently across dif-ferent situations (Kanagawa et al. 2001, Suh2002), they have more contradictory knowledgeabout themselves that is simultaneously accessi-ble (Spencer-Rodgers et al. 2008), and they aremore likely to endorse both positive and neg-ative statements about their own self-esteem(Hamamura et al. 2008a, Spencer-Rodgerset al. 2004).

One reason, then, why East Asians mightshow self-views that are less self-positive thanthose of Westerners is that they hold dialecti-cal views of themselves (e.g., I am a good per-son, but I am also a bad person). A dialectical

view of the self would lead to moderately pos-itive views of the self rather than overwhelm-ingly positive self-views, which is precisely theway that East Asian self-enhancement scoresdiffer from those of North Americans (Heineet al. 1999). Importantly, Spencer-Rodgerset al. (2004) find that peoples scores on ameasure of dialecticism mediate the differ-ences in self-esteem between East Asians andNorth Americans. Cultural differences in self-enhancement thus also stem from cultural dif-ferences in attitudes toward self-consistency.

Relational mobility. Another more recenteffort to understand the mechanisms underly-ing cultural variation in self-enhancing motiva-tions comes from the study of relational mo-bility (see Oishi et al. 2007, Yuki et al. 2007b;cf., Adams 2005). Relational mobility refers tothe perceived amount of opportunity that anindividual has for forming new relationships.In many individualistic contexts, such as thoseof American undergraduates, for example, in-dividuals live in a high-relational-mobility con-text, in which they are frequently meeting newpeople and have the potential to forge new re-lationships on a day-to-day basis. In contrast, inmany collectivistic contexts, for example, muchof Japanese society, there is little relational mo-bility in that people tend to belong to nonover-lapping groups (such as a school club or an of-fice) where there is little movement betweensocial groups and the membership is largelystable (also see Adams 2005 for similar argu-ments in West African communities). Becauseself-esteem is influenced by the degree to whichone feels socially accepted (Leary & Baumeister2000), it has been proposed that people willrely on their self-esteem to predict when theywill be accepted by others (Sato et al. 2007). Incontexts where people have many opportuni-ties for forming new relationships, then, havinghigh self-esteem will serve to aid them in func-tioning well. Indeed, the perceived availabilityof opportunities for forming new relationshipshas been shown to significantly mediate East-West cultural differences in both self-esteem(Sato et al. 2007) and in the relation between


Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


self-esteem and well-being (Yuki et al. 2007a).Relational mobility thus is another com-pelling candidate for a mechanism that canexplain cultural variation in self-enhancementmotivations.

Summary of mechanisms related to self-enhancement. The above review reveals sixdifferent mechanisms that underlie the ob-served cultural difference in self-enhancement,and it is possible that additional mechanismswill prove to be relevant in the future. Thishardly provides a parsimonious account for cul-tural variation in positive self-viewsthe ten-dency for North Americans to self-enhancemore than East Asians thus appears to beoverdetermined. Why might there be so manydifferent mechanisms related to this culturaldifference?

We suggest that the similar pattern acrosscultures for each of the six phenomena reviewedabove indicates that it is not productive to thinkof these as independent mechanisms underlyingself-enhancement. Rather, we propose that wecan understand the cultural variation in each ofthese phenomena as indicating a stable equi-librium point in a dynamical system (Cohen2001, Kitayama 2002). That is, the elementsof a culture are not independent from eachother. One feature of a culture (such as hav-ing a norm where extended families live in thesame household) will influence another feature(such as the likelihood that other family mem-bers get involved in decisions regarding whoone will marry; Lee & Stone 1980). This in-terdependence among different features of cul-tures reduces the variability of possible culturalarrangements. Each aspect of a culture is influ-enced by, and in turn influences, other aspectsof the culture. This interdependence results ina relatively small number of stable equilibriawithin a system. If an individual deviates froman equilibrium point, the interrelations amongthe various parts of the system will constrainher options, and she will likely gravitate backtoward the cultural norm (Boyd et al. 1997).

In present East Asian cultural contexts, a dy-namical system exists such that people tend to

view themselves as interdependent with signif-icant others, have few opportunities to forgenew relationships, tolerate contradictions, havemore incremental theories of abilities, are espe-cially attentive to others perspectives, are vigi-lant of potential losses, and exhibit self-criticalmotivations. Each of these psychological vari-ables is sustained by the other variables, andthey represent a fairly stable system. It is un-likely, say, that just one of these variables couldbe changed without influencing the other vari-ables as well. The mutual interdependence ofthese variables suggests that there are few op-portunities for much change in any single vari-able because the presence of the other vari-ables would act to constrain and stabilize thesystem (Boyd et al. 1997). We submit that cul-tural change in these variables is only likely tooccur when the pressures for change are greatenough that the system reaches a tipping pointand then gravitates toward a new equilibrium(Cohen 2001). For example, another equilib-rium point is found in present North Americancontexts, where people tend to view themselvesas independent from others, have many oppor-tunities to develop new relationships, eschewcontradictions, have entity theories of abilities,primarily consider their own perspective, areattentive to opportunities for gain, and evinceself-enhancing motivations. The dynamic sys-tems of the cultures of East Asia and NorthAmerica are not best described as different fromeach other on a single variable, such as theirself-construals, but rather they represent dif-ferent systems that gravitate toward divergentequilibria. Cultural change in these cultures islikely to be noticed across the entire systemwhen a tipping point is reached, rather than be-ing restricted to any transformation of a singlevariable. This systems view of culture calls intoquestion the value of efforts to identify medi-ational variables that are theorized to under-lie cultural differences (Heine & Norenzayan2006).

In summary, a cluster of interrelated vari-ables correlates with self-enhancement and dis-tinguishes East Asians from Westerners. Wesubmit that these variables mutually influence

376 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


each other and manifest in at least two differentstable equilibria within East Asian and Westerncultural contexts. It is possible that other cul-tural contexts possess different equilibria pointsamong these same variables.

STRUCTURE AND CONTENTOF PERSONALITY ACROSSCULTURESIn the above section, we outlined cultural dif-ferences in one aspect of individual difference,namely self-enhancement. The degree to whichself-enhancement is a salient and importanttrait depends on an intertwined set of culturalvariables. Self-enhancement, however, is onlyone type of individual difference. Are other in-dividual differences in fact equally important inall societies?

People tend to be curious and reflectiveabout the ways that individuals differ fromeach other. This curiosity may well be uni-versal across cultures, at least to a certain de-gree. Various different personality typologieshave been proposed over time and around theworld that serve to classify people into differ-ent types. For example, Hippocrates proposedthat there were four basic types of human tem-peraments, which depended upon the balanceof the four fluids, or humors, that were presentin the body: blood, yellow bile, phlegm, andblack bile. Ayurvedic medicine from India pro-poses that there are three metabolic body-types(vata, pita, and kapha), thus maintaining thatones metabolism rate provides the foundationof individual temperaments. Popular Japanesefolklore views the four blood types as underly-ing reliable differences in personality. In short,across cultures and history, people have comeup with a remarkably diverse array of ways forcarving up personalities.

Western psychologists have also made manytargeted research efforts toward developingpersonality typologies to classify the variety ofways to be a person. Several different schemeshave been proposed (e.g., Ashton et al. 2004,Cattell 1957, Eysenck 1975), each varying inthe number of core traits and the content of

those traits. However, the typology that is byfar the most widely accepted and researchedis the Five-Factor Model (McCrae & Costa1987; for criticisms of this model, see Block1995, McAdams 1992). According to this model[first derived by Fiske (1949)], there are fivecore personality traits: openness to experience,conscientiousness, extraversion, agreeableness,and neuroticism. The Big 5 are said to un-derlie the nearly 18,000 traits that exist in theEnglish language (Allport & Odbert 1936).Several hundreds of studies have explored thesetraits and their relation to other constructs.This research raises some interesting and im-portant questions regarding personality acrosscultures: Is the five-factor structure somethingbasic about human nature that we should findin the personalities of people in all cultures thatwe look? Or, alternatively, does the five-factormodel reflect ideas about personhood that arelimited to the West, where the vast majority ofthis research has been conducted?

Apparent Near Universalityof Personality StructureA number of evolutionary psychological per-spectives on personality maintain that the five-factor model reflects universal kinds of indi-vidual variation. Some have argued that theBig 5 are fundamental responses to core chal-lenges faced by humans (e.g., Ellis et al. 2002,Goldberg 1981). For example, it would be adap-tive for people to be able to identify who waslikely to rise in the social hierarchy (extraver-sion), who could be reliable and dependable(conscientiousness), who would have difficultycoping with adversity (neuroticism), who couldbe a good friend (agreeableness), and whomone could turn to for wise advice (openness;Buss 1991)that is, the accurate perception ofthe Big 5 in others could enhance ones fit-ness. However, arguments for why it is adap-tive for individuals themselves to vary in theBig 5 are currently incomplete, as the heri-tability of personality traits (typically around0.40; Plomin et al. 2001) makes it appearthat between-individual variability should be


Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


drastically reduced, as long as personality traitsuniformly afforded fitness across all environ-ments. Compelling evolutionary accounts forwhy individuals differ in the degree to whichthey possess adaptive personality traits may bedeveloped if we consider individuals responsesto their different environmental niches (Penkeet al. 2007). Regardless of the particular evolu-tionary theory that is applied, to the extent thatthe Big 5 evolved in response to core challengesfrom the ancestral environment, it follows thatthe model should be cross-culturally universalin its application.

Some evidence supporting the biologicaluniversality of the Five-Factor model can befound in comparative research, which has iden-tified markers of the Big 5 traits in a number ofanimal species (Gosling & John 1999). For ex-ample, behavioral patterns consistent with eachof the Big 5 traits have been identified in chim-panzees (King & Figueredo 1997), and sometraits, for example neuroticism, have been iden-tified in species as diverse as hyenas (Gosling1998), guppies (Budaev 1997), and octopuses(Mather & Anderson 1993). It is possible thatthe Big 5 (or at least some of the dimensions)represent fundamental responses to biologicalchallenges encountered by many, if not most,species. However, the vast majority of animalstudies have been conducted by Western re-searchers, and the similarity of the traits that areobserved between animals and humans mightbe due to people interpreting animal behav-ior through the lens of their most familiar waysof categorizing peoplean account that is ad-dressed to a degree by noting that evidence fortraits in animals is clearer for some traits andin some species than in others (Gosling 2001).Nonetheless, the best evidence for the uni-versality of a psychological construct requiresthe consideration of data from multiple cul-tures (Norenzayan & Heine 2005). The studyof the cross-cultural generalizability of the Big5 is one of the most ambitiously researched at-tempts to address the question of universalityfor any psychological phenomenon, and sev-eral large-scale multicultural studies have beenconducted.

Various measures of the Big 5 [e.g.,Neuroticism-Extroversion-Openness Person-ality Inventory-Revised (NEO-PI-R); Costa &McCrae 1992] have been translated into a num-ber of languages and have been distributedto thousands of people in dozens of culturesaround the world. Early cross-cultural com-parisons of the factor structure of the Big5 were promising: Four out of five factors(all except Openness) emerged in Hong Kong(Bond 1979), Japan (Bond et al. 1975), and thePhilippines (Guthrie & Bennett 1971), reveal-ing considerable similarity in the structure ofpersonality across these diverse cultures. Morerecent studies with some other cultures havefared even betterall five factors emerged incultures from countries as diverse as Israel(Montag & Levin 1994), Korea (Piedmont &Chae 1997), and Turkey (Somer & Goldberg1999). One large-scale study investigated peo-ple from 50 different cultures from all conti-nents except Antarctica and had participantsevaluate someone they knew well on trait ad-jectives that assessed the Big 5 (McCrae et al.2005). In most of the 50 cultures, the factorstructure of the Big 5 was replicated. In a num-ber of developing cultures (in countries includ-ing Botswana, Ethiopia, Lebanon, Malaysia,Puerto Rico, and Uganda), the factor structurewas not so evident. However, in these latter cul-tures, the quality of data was rather poor, whichsuggests that people may not have fully under-stood the questions or were unfamiliar with an-swering questions in that format (McCrae et al.2005). If unfamiliarity with Western measurescan account for the poor data fit found in somecultures, then there is good evidence that theBig 5 reflect the universal structure of person-ality (also see Allik & McCrae 2004, Yik et al.2002). Still, support for universality would bestronger if convergent evidence emerged fromstudies of developing and small-scale societies(cf., Henrich et al. 2005).

It is important to note that the measuresof the Big 5 (such as the NEO-PI-R) wereinitially developed through the exploration ofEnglish personality terms, and largely withAmericans. The challenge with factor analyses

378 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


is that they only speak to the structure thatemerges from the universe of items that wereconsidered. It is possible that a different set ofitems, particularly those that were more mean-ingful in other cultural contexts, might reveala different underlying personality structure. Animportant question to consider, then, is whetherthe Big 5 personality dimensions emerge re-gardless of what traits one considers, or whetherthey reflect the underlying structure of thekinds of personality traits that are discussed inEnglish.

A number of investigations have exploredthis question. For example, Cheung et al. (1996)sought to identify what kinds of personality di-mensions would emerge if they factor-analyzedindigenous Chinese personality traits ratherthan relied on translations of English traits.The researchers first explored the kinds ofpersonality traits that were common in Chi-nese by examining Chinese novels, Chineseproverbs, peoples personality descriptions, andthe Chinese psychology literature. These ef-forts revealed 26 unique personality constructs(as well as another 12 clinical constructs).The constructs were then put into a personal-ity questionnaire (the Chinese Personality As-sessment Inventory), which was completed byChinese participants. The resultant factorstructure was not the same as the Big 5; rather,four factors emerged that were captured bythe following labels: dependability (reflectingresponsibility, optimism, and trustworthiness),interpersonal relatedness (reflecting harmony,thrift, relational orientation, and tradition), so-cial potency (reflecting leadership, adventur-ousness, and extraversion), and individualism(reflecting logical orientation, defensiveness,and self-orientation). Further analyses includedthe Chinese Personality Assessment Inventorytogether with a measure of the Big 5 (Cheunget al. 2003). That analysis revealed that therewas substantial overlap between three of thefactors; namely, neuroticism correlated withdependability, extraversion correlated with so-cial potency, and individualism correlated withagreeableness. Openness to experience did notcorrelate with any of the Chinese factors, and

interpersonal relatedness was not correlatedwith any of the Big 5 factors. Perhaps, then,interpersonal relatedness may be a sixth person-ality factor that is especially salient in Chineseculture. Whether interpersonal relatedness is areliable sixth factor in Western samples has yetto be demonstrated.

Similar approaches have been taken withother cultures. For example, Church et al.(1997; also see Church et al. 1998) developedan indigenous list of Filipino personality traitsand explored their underlying factors throughfactor analysis. This analysis revealed five traitsthat were highly similar to the Big 5; how-ever, they also revealed two additional factors:temperamentalness and a negative valence di-mension, which did not correlate strongly withany of the Big 5. Likewise, Benet-Martinez &Waller (1995, 1997) found that an investiga-tion of Spanish personality constructs revealedseven underlying personality factors, althoughthese did not map on so well to the Big 5. Simi-larly, Saucier et al. (2005) found that a six-factorsolution emerged from indigenous Greek termsand was somewhat at odds with the Big 5. Ingeneral, investigations with indigenous traitsreveal that although the Big 5 personality traitsappear to be cross-culturally robust, they maynot be an exhaustive list of the ways that per-sonality can emerge in other cultures. Somealternative dimensions have emerged from ex-plorations of personality structures using in-digenous personality terms, and future researchis necessary to determine the robustness anduniversality of these other factors.

Cross-Cultural Variability in Levelsof Personality TraitsGiven the evidence that the Five-Factor modelof personality appears to adequately capturethe structure of personality traits in many cul-tures, researchers have recently begun to com-pare mean levels of personality traits acrosslarge samples of cultures (e.g., McCrae 2002,McCrae et al. 2005, Schmitt et al. 2007). Thisburgeoning research program has resulted indebate about the meaning and validity of such


Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


cross-cultural comparisons. Below, we outlinesome of the findings of these large cross-cultural comparisons and discuss the debateabout their validity.

Some of the most thorough multinationalcomparisons that have been conducted in psy-chology have compared Big 5 traits across cul-tures. As of this writing, aggregate person-ality means from the NEO-PI-R (Costa &McCrae 1992) have been reported for self-ratings from 36 cultures (McCrae 2002) andfor peer-ratings from 51 cultures (McCrae et al.2005), and a modified Big 5 measure was usedto collect peoples perceptions of their compa-triots in 49 cultures (Terracciano et al. 2005).Another popular measure, the Big Five In-ventory (BFI; Benet-Martnez & John 1998),has been used to collect self-ratings in 56 na-tions (Schmitt et al. 2007). This hard-wonwealth of data has attracted much interest andsparked further research (e.g., McCrae & Allik2002). It has shown, for example, that accord-ing to the self-report means, the most neuroticpeople on the planet are Spaniards, the mostextraverted are Norwegians, the least consci-entious are Japanese, the most open to new ex-periences are Austrian, and the most agreeableare Malaysian (McCrae 2002).

Part of the promise of these kinds of multi-national comparisons of mean levels of person-ality traits is that they stand to map out the per-sonality profiles of cultures across the globe.The value of this research enterprise would beespecially noteworthy to the extent that it of-fered cultural profiles that were of greater valid-ity than those profiles formed on the basis of in-ferior or biased methods, such as those formedon the basis of peoples stereotypes. To demon-strate this point, Terracciano, McCrae, and col-leagues investigated how well peoples percep-tions of the national character of their countrycorrelated with the means from self-reports andpeer reports on the NEO-PI-R discussed above(McCrae & Terracciano 2006, Terracciano et al.2005). The results indicated that there were es-sentially no correlations between the nationalcharacter profileswhat people believe theiraverage compatriot is likeand the actual na-

tional average self-ratings or peer ratings on theNEO-PI-R. The investigators argued that thefindings provided strong evidence that com-mon perceptions of national character in facthave little to no connection with reality; peo-ples views of their compatriots do not appearto contain even a kernel of truth (McCrae &Terracciano 2006, p. 160).

The assertion that aggregate self-reports orpeer reports are appropriate validity criteria inthemselves, and that perceptions of nationalcharacter are therefore illusory, has been metwith some resistance (Ashton 2007, Heine et al.2008a, McGrath & Goldberg 2006, Perugini& Richetin 2007). Indeed, the literature oncross-cultural methodology raises a number ofcaveats that should make one cautious in draw-ing conclusions from direct comparisons ofmean levels of personality traits across cultures.For example, there are questions of whetheritems are interpreted in the same way by peo-ple from all cultures (e.g., Church & Katigbak2002, Grimm & Church 1999, Poortinga et al.2002), whether people respond to items in thesame way (Chen et al. 1995, Greenfield 1997,Hamamura et al. 2008a, Poortinga et al.2002), and whether individuals in different cul-tures compare themselves to different stan-dards when making ratings (e.g., Heine et al.2002, 2008a; Peng et al. 1997). Nevertheless,some personality researchers have optimisti-cally maintained that most of these potentialbiases can be controlled for (e.g., the acquies-cence bias; McCrae 2001, McCrae et al. 2005)or that these differences still yield largely inter-pretable results (McCrae et al. 2005, Schmittet al. 2007). The difficulties in comparing meanscores on subjective Likert scales across culturesmeans that researchers must seriously considerwhat kinds of data could actually validate suchcross-cultural comparisons.

What are the sources of evidence for andagainst the validity of such cross-national per-sonality comparisons? Though evidence fromdata clustering of national personality profilesand some correlations with other national-level variables have been put forth as validat-ing mean nation-level scores, other evidence

380 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


suggests that this kind of national profilingmay be inaccurate, such as the low reliabilitybetween different measures of the Big 5, dis-agreement with expert ratings, and bizarre cor-relations with behavioral measures. We discussthis evidence below.

Cluster analyses indicate some reasonablerelationships emerging from the cross-culturalcomparisons of the traits. For example, analy-ses of profile similarity reveal that cultures ofsimilar geographical or historical backgroundstend to cluster together (Allik & McCrae 2004,McCrae et al. 2005, Schmitt et al. 2007).Though suggestive of validity, we note thatcluster analyses are difficult to examine as valid-ity evidence. For example, Schmitt et al. (2007)find that although most of the closest pair-ings on BFI personality profiles are predictable(e.g., Botswana and South Africa, Cyprusand Greece), some others are not explain-able by geographic or historical similarity (e.g.,Estonia and Mexico, Israel and Finland). Moreproblematic, similar personality profiles couldreflect either actual personality similarities orsimply similar cultural standards for compar-ison and are therefore not necessarily goodevidence of validity (Heine et al. 2008a).

Researchers have also calculated correla-tions of mean trait levels with other country-level data to establish validity of the cross-cultural comparisons. For example, Schmittet al. (2007) found that extraversion correlatedwith liberal views toward sexuality both withinand between cultures. McCrae (2002) foundthat Hofstedes (2001) cultural dimensions cor-related with some of the Big 5 measures. Thisconvergence with other criteria would appear tobe a good demonstration of the validity of thecountry scores. However, we note a few pointsabout using other kinds of country scores tovalidate the personality data. First, it is crucialthat any validity criteria be theoretically rele-vant a priori. For example, noting that neuroti-cism and masculinity are correlated (McCrae2002) does not provide validity unless thereare clear theoretical reasons to anticipate suchcorrelations beforehand. Second, validity cor-relations should be reliable across different

Reference-groupeffects: implicitcomparison to theaverage, or ideal,amount of a constructwithin your groupwhen makingself-ratings

measurements of the Big 5. We note that nosignificant correlations exist between any of theBig 5 and Hofstedes (2001) five dimensions thatreplicate across three independent measures ofthe Big 5 (McCrae 2002, McCrae et al. 2005,Schmitt et al. 2007). Third, we emphasize thatthe strongest kind of criteria that one could seekto validate country mean scores would be thosethat utilized different methods. Finding signifi-cant correlations between two sets of self-reportmeasures could reflect the fact that both mea-sures are compromised by the same kinds ofculturally specific reference-group effects andresponse biases.

Many sources of evidence call into ques-tion the validity of these cross-cultural compar-isons. One first step to demonstrating validityis to establish the reliability of the findingsitis difficult to make the case that one rank or-dering of means is valid if it is not reproducedthrough other methods. However, the rank or-derings that have emerged from the above en-deavors to compare personality traits acrosscultures do not correlate particularly strongly.For example, correlations between the coun-try scores for the self-report measures of theBig 5 with the NEO-PI-R and the BFI rangedfrom 0.22 to 0.45 (Schmitt et al. 2007), whichare quite modest given that these are mea-sures of the same constructs. Perhaps more dis-turbing is that the correlations between thecountry scores from the BFI and NEO-PI-Rmeasures correlate more weakly for the cor-responding traits than they do for their non-corresponding traits in four of the Big 5 traits(e.g., the BFI measure of openness correlates0.73 with the NEO-PI-R measure of extraver-sion, but only 0.27 with the NEO-PI-R mea-sure of openness; Schmitt et al. 2007). This is indirect violation of the multitrait-multimethodmatrix approach to validating personality traits(Campbell & Fiske 1959). Furthermore, as de-scribed above, the country scores from the per-ceptions of national character showed no sig-nificant positive correlations with the countryscores from the NEO-PI-R for any of the Big5 traits (Terracciano et al. 2005). In sum, thereis little convergence among the country scores


Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


across different assessments of the same person-ality traits.

One method of validating conflicting cross-cultural data has been to utilize expert ratings(e.g., Heine et al. 2002, John & Robins 1994).The relative rankings of cultures on mean self-ratings of traits has been shown to disagreewith the judgments of cultural experts (Church& Katigbak 2002, McCrae 2001), whereas thenational character profiles are closer to ex-pert ratings (Terracciano et al. 2005, footnote#26). However, a weakness of this validationstrategy is that the expert ratings, such as theNational Character ratings, may draw on thesame invalid cultural stereotypes (McCrae &Terracciano 2006, Terracciano et al. 2005).

We submit that the strongest evidence forvalidity would come from actual observationsof personality-related behavior frequency in thedifferent cultures, but such data are difficult tofind or produce (e.g., Ashton 2007). A recentexample of using behavioral data to validate thecountry scores of conscientiousness (the traitwith the clearest behavioral markers; also seeRoberts et al. 2007), and the only one for whichwe could find cross-national data, found thatNational Character ratings correlated highlywith national rankings on conscientious-relatedbehaviors such as clock accuracy and efficiencyof postal clerks (average r = 0.61), whereas theNEO-PI-R and BFI self- and peer-report ag-gregate means correlated negatively or not atall with these behaviors (average rs ranged from0.43 to 0.06; Heine et al. 2008a). These find-ings indicate that the National Character rat-ings are more accurate than average self-reportsor peer reports at predicting the conscientiousbehaviors of average citizens. These findingsdovetail with other evidence that comparisonsof self-report measures across cultures sufferfrom some serious methodological confounds(Cohen 2007; Heine 2008; Heine et al. 2001,2002; Kitayama 2002). We suggest that in theabsence of convergent evidence from other de-signs, any cultural differences in means on sub-jective Likert scales should be taken with a grainof salt. At the least, future cross-cultural com-parisons of personality need to more seriously

consider validity criteria and the developmentof improved methods.

Although methodological artifacts such asthe reference-group effect make it problematicto compare means across cultures, it is impor-tant to underscore that those same problemsdo not typically emerge when using self-reportscales within cultures. Within a culture, peopletend to evaluate themselves in contrast to sim-ilar referents, a method that preserves the va-lidity of the rank order of individuals within aculture as well as with correlations both withinand between self-report scales. In fact, as dis-cussed above, cross-cultural studies of the struc-ture of personality have revealed much evidencefor universality. Arguably, it is part of the hu-man condition to perceive personality in termsof universal traits. However, another questionto consider is the extent to which people at-tend to and rely on personality information intheir efforts to understand themselves and oth-ers. Are personality traits of equal utility acrosscultures?

THE UTILITY OF PERSONALITYACROSS CULTURESMarkus & Kitayama (1991) played a key rolein relaunching the field of cultural psychologywhen they posited that the self-concept variedin significant ways across cultures. Although inthe West the self tends to be identified more asan independent entity, importantly grounded ininternal traits, the interdependent self-conceptthat is more common in the rest of the worldis largely based on its relationships and roleswith others. This difference in self-definitionacross cultures raises the possibility that in so-cieties more characterized by interdependentselves, personality traits might be of less util-ity for understanding oneself or in predictingthe behavior of others than are more relation-ally defined aspects of the selfsuch as socialroles (see Markus & Kitayama 1998). Belowwe consider evidence that speaks to the ques-tion of whether personality is of comparableutility between individualistic and collectivisticsocieties.

382 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


Content of the Self-ConceptOne source of information germane to thequestion of the utility of personality is the kindof information that people spontaneously con-sider when describing themselves. Open-endeddescriptions of the self-concept measured us-ing the Twenty Statements Test (Kuhn &McPartland 1954) have consistently revealedevidence for a weaker tendency to list pure psy-chological attributes (largely personality traits)among people from various collectivistic cul-tures (e.g., Native Americans, Cook Islanders,Masai, Samburu, Malaysians, and East Asians)than among those from individualistic cultures(e.g., Australians, Americans, Canadians, andSwedes) when describing themselves. Instead,people from various non-Western cultures aremore likely to describe themselves in termsof their social roles (Ip & Bond 1995, Ma &Schoeneman 1997) or specific descriptors thatare not abstract trait terms (Rhee et al. 1995).Evidence from these studies suggests that theself-concepts of people in collectivistic culturesmay not emphasize abstract personality traits inthe same way that self-concepts common in in-dividualistic societies do. Personality traits maynot be useful to the same degree everywhere fordescribing the self.

Incremental Versus EntityTheories of SelfAnother phenomenon related to the perceivedutility of personality trait knowledge is the laytheory that people tend to embrace regard-ing the nature of their selves. As discussedabove, people tend to view the self as be-ing either a rather stable and immutable en-tity or as more fluid and changing. Dweckand colleagues (Dweck & Leggett 1988, Honget al. 1999) have described these views as entityand incremental theories of self, respectively.Typical views of personality in individualisticcultures are grounded in the notion that per-sonality traits are inherited and somewhat sta-ble across the lifespanideas that are concep-tually consistent with an entity theory of self.

Twenty StatementsTest: a method ofmeasuring the contentof the self-concept byasking participants tocomplete twentyI am . . . statements

The notion of an ever-changing and incremen-tal theory of the self would seem to be at oddswith the notion of trait theories (Levy et al.1998, Molden & Dweck 2006). As describedabove, past cross-cultural research on theoriesof self finds that in comparison with Western-ers, East Asians are less likely to conceptualizetheir selves in entity terms (Heine et al. 2001,Norenzayan et al. 2002). The incremental na-ture of the self-views of East Asians is incon-sistent with Western views of stable and innatepersonality traits. It remains to be seen whetherpeople from collectivistic cultures outside ofEast Asia also demonstrate incremental viewsof themselves. In summary, lay theories of theself, at least in East Asia, are at odds with theprevailing view of personality as consisting ofstable traits, and such a view may be utilizedless in such cultures for the understanding ofself and others.

Perceived Consistency of TraitsAnother perspective on the utility of personalitytraits is the consistency that those traits man-ifest across situations. To the extent that peo-ples perceptions about their personality varyconsiderably across situations, this would ren-der personality traits to be less useful for un-derstanding the person (Mischel 1968; thoughsee Fleeson 2004 for new interpretations of theperson-situation debate). Indeed, the lay the-ories of personality and personhood in collec-tivistic contexts may in fact be closer to thatproposed by Mischel & Shoda (1995), in whicha persons traits shift across situations in anindividually characteristic pattern. The powerof the situation over behavior is acknowledgedmore in collectivistic cultures than it is withinindividualistic cultures. Much cross-cultural re-search has explored the extent to which person-ality is consistent across situations.

For example, Kanagawa et al. (2001) exam-ined how much the testing situationfillingout a questionnaire in front of a profes-sor versus in front of one peer, a group ofpeers, or aloneinfluenced self-descriptions.They found that Japanese self-descriptions (on


Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


the Twenty Statements Test) varied signifi-cantly more depending on the testing situationthan did Americans self-descriptions. Like-wise, in an experience-sampling study, Oishiet al. (2004) asked participants in India, Japan,Korea, and the United States to record theirmood and who they were with (i.e., their sit-uation) at random moments during the day.Cultural differences emerged in the effect ofsituations on mood. For example, whereasJapanese participants felt much happier whenwith a romantic partner than otherwise, Amer-icans did not experience as much of a moodchange. Mood was more influenced by situa-tion in collectivistic cultures than in individ-ualistic cultures. Similarly, Suh (2002) askedKorean and American participants to reportwhat they believed their personality to be likewith five different people (e.g., parents, closefriend, or stranger) as well as in general. Theresults indicated that Korean participants re-ported much less consistency among these sixratings than did American participants; more-over, consistent selves were more strongly cor-related with positive outcomes for Americansthan they were for Koreans. Relatedly, a num-ber of studies have found that East Asians tol-erate more contradiction in their thoughts ofself, including variation across contexts, thando Westerners (Choi & Choi 2002; Hamamuraet al. 2008a; Spencer-Rodgers et al. 2004, 2008).

These studies suggest that the East Asian selfis not as consistent across situations in com-parison with the Western self. This raises thequestion of how the East Asian self might main-tain enough coherence to be even considered aself. One possibility is that despite being unsta-ble across situations, one might display a stablepersonality within situations across time. Onesglobal traits might not be a good way to de-fine ones self, but ones traits within a certainsocial rolearound a certain relationshipmight be. To investigate this question,English & Chen (2007) asked Asian Ameri-can and Euro-American participants to ratetheir personality traits within certain relation-ship contexts. As found by Suh (2002), the cor-relation of traits between relationship contexts

was smaller for Asian Americans than it was forEuro-Americans. Importantly, however, AsianAmericans showed as much consistency withinthat relationship situation over time as Euro-Americans did. In other words, their self-ratingsof traits within a certain relationship contextwere quite stable over time; that is, who one iswith ones mother does not change, even if thisis quite different from who one is with onesroommate. Likewise, in other research, whenEast Asians were asked if they had a true self,they considered a context-sensitive self. In con-trast, Westerners responded to this question byconsidering their feelings of self that were in-variant across situations (Kashima et al. 2004;also see Tafarodi et al. 2004). This researchhighlights how the self-concept in East Asiancontexts appears to be grounded in ones rolesand relationships rather than something thatprimarily derives from component traits. Fu-ture research is necessary to see whether theWestern and East Asian patterns generalize toother cultural contexts.

Attributions for BehaviorThe above review has considered how peoplein collectivistic cultures appear to rely on per-sonality traits less than do those from individ-ualistic cultures in understanding themselves.Other research indicates that people from col-lectivistic cultures might rely on personalitytraits less than Westerners do for understand-ing others as well. This research on how peopleexplain the behavior of others reveals anotherway that utility of personality varies across cul-tures. A number of classic studies have foundthat when asked to explain the behavior of oth-ers, people tend to largely attend to the per-sons disposition as a means for explaining thebehavior, even when there are compelling sit-uational constraints available ( Jones & Harris1967, Ross et al. 1977). This tendency to ig-nore situational information in favor of per-sonality information when explaining the be-haviors of others is so commonly observed thatit has been termed the fundamental attribu-tion error. However, as with so many other

384 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


psychological phenomena, this original re-search had been conducted almost exclusivelywith Western participants. Observations with anumber of collectivistic cultures have painteda different picture regarding peoples preferredways of making sense of the behavior of oth-ers. Geertz (1975) described how Balinese donot tend to conceive of peoples behavior interms of underlying dispositions, but instead seeit as emerging out of the roles that they have.Shweder & Bourne (1982) found that Indianstended to eschew trait descriptions of othersbehaviors but rather would explain their be-haviors in descriptive terms. Building upon thisidea, Miller (1984) found that Indians showedevidence for a reverse fundamental attributionerror in that Indian adults tended to favor situ-ational information over personality accounts.More recently, several studies conducted withEast Asians and Americans reveal that whereasAmericans attend to dispositions first, regard-less of how compelling the situational infor-mation may be (Gilbert & Malone 1995), EastAsians are more likely than are Americans toinfer that behaviors are strongly controlled bythe situation (Norenzayan et al. 2002) and aremore likely to attend to situational informa-tion (Miyamoto & Kitayama 2002, Morris &Peng 1994, Van Boven et al. 1999), particu-larly when that information is especially salient(Choi & Nisbett 1998). They may even auto-matically consider the situational informationprior to the personality information (Knowleset al. 2001; but for contrary findings, seeLieberman et al. 2005). Furthermore, in an in-vestigation of peoples lay beliefs about per-sonality across eight cultures, Church et al.(2006) found that people from individualisticcultural backgrounds (i.e., American and Euro-Australian) strongly endorsed implicit-traitbeliefs, such as the notions that traits remainstable over time and predict behavior over manysituations. In contrast, they found that thosefrom collectivistic cultural backgrounds (i.e.,Asian Australian, Chinese Malaysian, Filipino,Japanese, Mexican, and Malay) more stronglyendorsed contextual beliefs about personality,such as ideas that traits do not fully describe a

person as well as roles or duties and that trait-related behavior will change from situation tosituation. In summary, people in collectivisticcultures appear to be less likely than are peoplefrom individualistic cultures to utilize person-ality information in explaining the behavior ofothers.

Spontaneous Trait InferencesDo these cultural differences also exhibit them-selves at an automatic, cognitive level? Much re-search has revealed that people spontaneouslyencode observed behaviors in terms of under-lying traits: For example, learning of one per-son giving money to another person in needmay be encoded as generous (Uleman 1987).However, until recently, the majority of this re-search had been conducted in Western cultures,thus failing to shed light on the question of theuniversality of this tendency. More recent cross-cultural studies suggest that such spontaneoustrait inferences might not be so common else-where. For example, Maass et al. (2006) foundthat whereas Italians inferred traits from be-haviors and viewed trait adjectives to be pre-dictive of future behaviors, Japanese did thissignificantly less so. Rather, Japanese tended torely more on behavior-descriptive verbs in theirperson descriptions and memories of targetevents. Likewise, Zarate et al. (2001) found thatLatinos showed evidence of fewer spontaneoustrait inferences compared with North Ameri-cans. Similarly, tendencies to make spontaneoustrait inferences have been shown to correlatewith trait measures of independence (Duff &Newman 1997), which are more common inindividualistic cultures. In a study that exploredhow well people encoded trait versus role in-formation about themselves, Wagar & Cohen(2003) utilized the self-reference effect, inwhich words encoded in relation to elaboratedareas of self-concept are remembered better, todetermine whether social or personality traitswere more cognitively elaborated areas of theself-concept. This study revealed that AsianCanadians, compared to Euro-Canadians, re-membered social-role words better than they


Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


remembered trait words when they were en-coded in relation to the self, a finding that sug-gests that the social role aspect of identity wasmore cognitively elaborated than were person-ality traits. These studies converge to suggestthat people from collectivistic cultures are lesslikely to spontaneously encode trait informa-tion either about others or about themselves.

Personality Traits and BehaviorThe above review is consistent with our thesisthat personality, defined as situation-consistenttraits, is of less importance in collectivistic cul-tures than it is in individualistic ones. Ulti-mately, however, the most compelling kind ofevidence in support of this claim would be evi-dence that personality traits are less predictiveof behavior in collectivistic cultures. Among themany forces that prompt and guide behavior,such as norms, role obligations, peer pressure,and situational influences, we should expect thatpersonality traits play a less central role amongcollectivists than they do among individualists.Is there any evidence for a greater decouplingof personality and behavior among people fromcollectivistic cultures?

This question is challenging to address be-cause evidence for the relationship between be-haviors and personality is relatively rare evenin Western contexts owing to the practical dif-ficulties of assessing behaviors. Some evidencefor the predictive validity of personality traitsin behaviors among Westerners includes be-havioral residue, such as how one decoratesones dorm room (Gosling et al. 2002), life out-comes, such as health and occupational success(Roberts et al. 2007), and discreetly observedbehaviors, such as whether one cheats on anexam (e.g., Nathanson et al. 2006). Thus far,however, such direct behavioral evidence hasbeen limited to studies with Westerners.

Some indirect evidence speaks to the ques-tion of the predictive validity of personalitytraits across cultures. For example, considerone cross-cultural difference that was reviewedabove: When explaining other peoples behav-iors, those from collectivistic cultures rely on

personality information less than do those fromindividualistic cultures, and they are less likelyto communicate that information (e.g., EastAsian newspapers tend not to report on in-formation about peoples personalities as muchas do Western newspapers; Morris & Peng1994). Although it is possible that people arewrong in their theories about what are theactual causes for others behaviors, it is in-formative that in comparison with individual-ists, collectivists believe that personality is aless compelling explanation for peoples behav-ior; collectivists may indeed perceive a weakercorrelation between observed personalities andbehaviors.

A second indirect source of evidencecomes from studies that compare peoples self-reported personality with peer ratings. Peerratings are often used as an index for be-haviors because peers are in the position toform personality assessments on the basis ofobserved behaviors (e.g., Gosling et al. 1998,John & Robins 1994). For example, a studyby Suh (2002) compared self-reported person-ality traits with ratings made by friends andparents of Korean and American participants.Results indicated that the self-peer correlationswere lower among Koreans than they wereamong Americans. Moreover, correlations be-tween parent ratings and friend ratings werealso lower among Koreans than among Amer-icans. These findings are consistent with thenotion that personality traits, as perceived bythe self, are less predictive of behavior (as wit-nessed by the observers) among East Asiansand Americans. East Asians act more differentlyacross contexts than do Americans, suggestingthat contextual factors are guiding their behav-ior relatively more so than are traits. Before wecan draw any firm conclusions on any culturaldifferences in the predictive validity of traits,it will be necessary to utilize more direct mea-sures of behavior and to consider a wider ar-ray of cultural samples. Indeed, thus far almostall of the literature relevant to the question ofthe cross-cultural utility of personality has fo-cused only on North American and East Asiansamples.

386 Heine Buchtel

Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


CONCLUSIONPersonality research has been greatly informedby investigations outside of Western culture,and such data provide new perspectives to ad-dress important questions. Personality researchhas taken an important step in advancing thefield from what was largely the study of Amer-ican undergraduates to the study of humannature. We applaud this move, and we urgethe field to consider a much broader spectrumof samples, including those from other socialclasses and other age groups, and to target non-literate subsistence populations as well. Suchkinds of investigations have the potential toidentify what appear to be human universals(e.g., the structure of personality) and what isculturally variable (e.g., the positivity of evalua-tions of personality, the distribution of person-ality traits, and the utility of personality).

An understanding of what is universal andwhat is variable about human personality is not

some tangential question, but rather stands toilluminate fundamental concerns of the field(for more discussion, see Norenzayan & Heine2005). Evidence for universality is particularlyinformative for guiding evolutionary theoriesregarding the adaptiveness of certain facetsof personality, whereas evidence for variabil-ity provides important information regardingboundary conditions, mechanisms, and the roleof contextual variables in influencing aspectsof personality. The fact that so little work onpersonality has been conducted outside ofWestern samples (Quinones-Vidal et al. 2004)or has employed methods other than self-report[more than 95% of papers in the Journal ofPersonality rely on self-report methods (Kagan2007), which are particularly problematic forcross-cultural comparisons; Heine et al. 2002,2008a; Peng et al. 1997)] means that there arestill vast lacunae in our understanding of humanpersonality.

DISCLOSURE STATEMENTThe authors are not aware of any biases that might be perceived as affecting the objectivity of thisreview.

LITERATURE CITED

Aaker JL, Lee AY. 2001. I seek pleasures and We avoid pains: the role of self-regulatory goals in informationprocessing and persuasion. J. Consum. Res. 28:3349

Adams G. 2005. The cultural grounding of personal relationship: enemyship in West African worlds. J. Personal.Soc. Psychol. 88:94868

Allik J, McCrae RR. 2004. Toward a geography of personality traits: patterns of profiles across 36 cultures.J. Cross-Cult. Psychol. 35:1328

Allport GW, Odbert HS. 1936. Trait-names: a psycho-lexical study. Psychol. Monogr. 47:171220Arnett JJ. 2008. The neglected 95%: why American psychology needs to become less American. Work. pap.,

Clark Univ., Worcester, MAAshton MC. 2007. Self-reports and stereotypes: a comment on McCrae et al. Eur. J. Personal. 21:98386Ashton MC, Lee K, Perugini M, Szarota P, de Vries RE, et al. 2004. A six-factor structure of personality-

descriptive adjectives: solutions from psycholexical studies in seven languages. J. Personal. Soc. Psychol.86:35666

Bagozzi R, Wong N, Yi Y. 1999. The role of culture and gender in the relationship between positive andnegative affect. Cogn. Emot. 13:64172

Benet-Martnez V. 2007. Cross-cultural personality research: conceptual and methodological issues. In Hand-book of Research Methods in Personality Psychology, ed. RW Robins, CR Fraley, RF Krueger, pp. 17089.New York: Guilford

Benet-Martnez V, John OP. 1998. Los cinco grandes across cultures and ethnic groups: multitrait multi-method analysis of the Big Five in Spanish and English. J. Personal. Soc. Psychol. 75:72950


Ann

u. R

ev. P

sych

ol. 2

009.

60:3

69-3

94. D

ownl

oade

d fro

m a

rjour

nals.

annu

alre

view

s.org

by U

nive

rsity

of B

ritish

Col

umbi

a Li

brar

y on

12/

16/0

8. F

or p

erso

nal u

se o

nly.


Benet-Martinez V, Waller NG. 1995. The Big Seven factor model of personality description: evidence for itscross-cultural generality in a Spanish sample. J. Personal. Soc. Psychol. 69:70118

Benet-Martinez V, Waller NG. 1997. Further evidence for the cross-cultural generality of the Big Seven factormodel: indigenous and imported Spanish personality constructs. J. Personal. 65:56798

Block J. 1995. A contrarian view of the five-factor approach to personality description. Psychol. Bull. 117:187215

Bond MH. 1979. Dimensions of personality used in perceiving peers: cross-cultural comparisons of HongKong, Japanese, American, and Filipino university students. Int. J. Psychol. 14:4756

Bond MH, Nakazato H, Shiraishi D. 1975. Universality and distinctiveness in dimensions of Japanese personperception. J. Cross-Cult. Psychol. 6:34657

Boyd R, Borgerhoff-Mulder M, Durham WH, Richerson PJ. 1997. Are cultural phylogenies possible? InHuman by Nature: Between Biology and the Social Sciences, ed. P Weingart, pp. 35586. Mahwah, NJ:Erlbaum

Budaev SV. 1997. Personality in the guppy (Poecilia reticulata): a correlational study of exploratory behaviorand social tendency. J. Comp. Psychol. 111:399411

Buss DM. 1984. Evolutionary biology and personality psychology: toward a conception of human nature andindividual differences. Am. Psychol. 39:113547

Buss DM. 1991. Evolutionary personality psychology. Annu. Rev. Psychol. 42:45991Campbell DT, Fiske DW. 1959. Convergent and discrimination validation by the multi-trait multimethod

matrix. Psychol. Bull. 56:81105Campbell JD, Trapnell PD, Heine SJ, Katz IM, Lavallee LF, Lehman DR. 1996. Self-concept clarity:

measurement, personality correlates, and cultural boundaries. J. Personal. Soc. Psychol. 70:14156Cattell RB. 1957. Personality and Motivation Structure and Measurement. New York: World Book Co.Chen C, Lee S-Y, Stevenson HW. 1995. Response style and cross-cultural comparisons of rating scales among

East Asian and North American students. Psychol. Sci. 6:17075Cheung FM, Cheung SF, Leung K, Ward C, Leong F. 2003. The English version of the Chinese Personality

Assessment Inventory. J. Cross-Cult. Psychol. 34:43352Cheung FM, Leung K, Fan RM, Song W, Zhang J-X, Zhang J-P. 1996. Development of the Chinese

Personality Assessment Inventory. J. Cross-Cult. Psychol. 27:18199Choi I, Choi Y. 2002. Culture and self-concept flexibility. Personal. Soc. Psychol. Bull. 28:150817Choi I, Nisbett RE. 1998. Situational salience and cultural differences in the correspondence bias and in the

actor-observer bias. Personal. Soc. Psychol. Bull. 24:94960Choi I, Nisbett RE, Norenzayan A.

personality: the universal and the culturally specificheine/docs/annual review.pdfself-enhancement,...

Documents