japaness and american parents

来源：华佗小知识

Child Development, November/December 2002, Volume 73, Number 6, Pages 1788–1802

Japanese and American Children’s Evaluations of Peer Exclusion,

Tolerance of Differences, and Prescriptions for Conformity

Melanie Killen, David S. Crystal, and Hirozumi Watanabe

Children and adolescents (N ϭ 1,057), divided by gender, at fourth, seventh, and tenth grades, from two mid-sized cities in the United States and in Japan, were surveyed regarding their evaluations of peer group exclu-sion of atypical peers. Six reasons for atypicality were being aggressive, having an unconventional appearance,acting like a clown, demonstrating cross-gender behavior, being a slow runner, and having a sad personality.Analyses revealed signiﬁcant effects for age, gender, country membership, and the context of exclusion. Withage, children demonstrated context sensitivity, and believed that the excluded child should not change him- orherself to be accepted by the group. Across contexts, girls were less willing to exclude than were boys, andwere more tolerant of differences. The context of exclusion had an effect on all forms of judgments about exclu-sion, and there were very few overall effects for culture. Most children disagreed with the decision to exclude,believed that they were different from the atypical child, and believed that the excluded child shouldchange him- or herself to be accepted by the group. The results support a theory of developmental socialcognition in which multiple sources of inﬂuence have a signiﬁcant effect on social decision making involvingthe exclusion of others.

INTRODUCTION

Understanding how children and adolescents reasonabout exclusion reveals the ways in which youth con-ceptualize the relationship between the individualand the group (see Killen, McGlothlin, & Lee-Kim,2002). The legitimacy or wrongfulness of exclusiondepends on the nature of this conceptualization. Insome situations, for example, exclusion may be viewedas legitimate in order to preserve social group func-tioning (such as excluding a poor athlete from asports team); in other situations, exclusion may beviewed as wrong when individual rights are violated(such as excluding a club member because of ethnicbackground). Although social psychologists have longviewed exclusion solely in moral terms (see Opotow,1990), our research has indicated that exclusion fromgroups is a multifaceted phenomenon (Killen, Pisa-cane, Lee-Kim, & Ardila-Rey, 2001; Killen & Stangor,2001; Theimer, Killen, & Stangor, 2001). Different formsof reasoning are brought to bear on decisions aboutexclusion. In a similar manner, Wainryb and colleagues(Wainryb, Shaw, Laupa, & Smith, 2001; Wainryb,Shaw, & Maianu, 1998), also using a social–cognitiveframework, have found that tolerance is a multifac-eted construct; tolerance depends on the content andcontext of the message or value under consideration.This view is also distinct from prior research thatidentiﬁed tolerance as an absolute moral value. Fur-ther, our view stands in contrast to theories thatconceive of exclusionary attitudes as a personalitytrait, one that transcends context (Adorno, Frenkel-

Brunswick, Levinson, & Sanford, 1950; Pratto, Sida-nius, Stallworth, & Malle, 1994).

In the present study, the approach to the issue of ex-clusion was based, in large part, on social–cognitivedomain theory (Helwig, 1995; Killen, 1991; Smetana,1995; Tisak, 1995; Turiel, 1998), which suggests thatsocial reasoning about complex moral and socialproblems requires detailed analyses of context, andthat different forms of reasoning are brought to bearon such issues. The current investigation was de-signed to expand the domain-speciﬁc understandingof children’s and adolescents’ attitudes toward exclu-sion in two ways: ﬁrst, by adding a cross-culturalcomponent; and second, by using individual-level at-tributes of behavior or personality as criteria for ex-clusion rather than race or gender, which we haveused in prior studies (see Killen, McGlothlin, & Lee-Kim, 2002). If, as we have stated, decisions about ex-clusion necessarily involve judgments about the rela-tionship between the individual and the group, it isreasonable to expect that such decisions would varywidely by culture. Cultures, such as the United Statesand Japan, for example, are considered to have dis-parate orientations toward the individual–group rela-tionship. Theorists such as Markus and Kitayama(1991) and Triandis (19, 1995) have characterizedthe United States as an individualistic or independentculture, whose members are seen as giving priority to

the individual over the group. In contrast, Japan is de-scribed as a collectivistic or interdependent culture,whose members are thought to give priority to thegroup over the individual. Other theorists and re-searchers, however, have moved away from such adichotomous view, and have come to acknowledgethat collectivism and individualism coexist withincultures (Greenﬁeld & Cocking, 1994; Harkness, Super,& van Tijen, 2000; Killen & Wainryb, 2000; Turiel,1998; Turiel & Wainryb, 1994; Wainryb & Turiel, 1994).This latter approach has investigated the contexts inwhich members of diverse cultures give priority to in-dividualistic or collectivistic values (Wainryb & Tu-riel, 1994). Such an attempt necessarily involves ananalysis of the multiple contexts within any givenculture in which an individual makes a judgmentabout the relationship between the individual and thegroup, as, for instance, in contexts regarding exclusion.Much of the previous research on prejudice and ex-clusion has focused on social categories such as gen-der (e.g., boys excluding a girl from playing withtrucks; see Killen et al., 2001), race (e.g., a Black bas-ketball team excludes a White child from joining; seeKillen & Stangor, 2001), or social reference group (e.g.,“jocks” exclude a “nerd” from the team; see Horn, inpress). There are, however, criteria other than socialcategories on which decisions of exclusion may bebased—for example, certain types of personally un-pleasant behavior or various forms of unconventionalappearance. These individual-level attributes differfrom gender and race in that a person can, to varyingdegrees, potentially change him- or herself with re-gard to these characteristics to be less atypical andmore like other members of the group.

To examine the role of cultural membership on judg-ments about exclusion in nonsocial category contexts,we chose six reasons for exclusion, which have beenshown to be among the most common criteria for peergroup exclusion among youth in both the United Statesand Japan (see Crystal, Watanabe, & Chen, 2000;Rubin, Bukowski, & Parker, 1998; Shimizu, 1990). Thesesix categories—aggressiveness, unconventional appear-ance, poor athleticism, cross-gender identity, sad per-sonality, and social disruptiveness—reﬂected generaldomains used by Harter (1998) and colleagues for as-sessing self-development. Our goal was to ﬁnd six ex-amples of atypical peer behavior that ranged from rel-atively serious transgressions (e.g., aggressiveness ata shopping mall) to fairly minor annoyances (e.g., act-ing like a clown in a movie theater). We theorized thatpotentially unacceptable behavior, such as aggres-siveness, would be more salient to children and ado-lescents than would cultural ideology about individ-ual/group relationships. For example, we predictedKillen, Crystal, and Watanabe17

that children would judge it as legitimate for a group toexclude a child who was aggressive—and that thiswould be the case even for students from a culture suchas Japan, in which preserving the group is highly val-ued. Thus, we expected children’s evaluations of ex-clusion to vary depending on the reason for exclu-sion, and cultural ideologies to be a contributing—butnot a determining—factor in how children evaluatedthese reasons.

As noted above, decisions about exclusion aremultifaceted and based on numerous considerations.One possible consideration is the perception of simi-larity or difference between the excluder and the ex-cluded. Social comparison and social developmentalresearchers have indicated that individuals tend to befavorably biased toward others whom they perceiveto be similar to themselves (Boivin, Dodge, & Coie,1995; Osbeck, Moghaddam, & Perreault, 1997; Poulin,Cillessen, Hubbard, & Coie, 1997; Rubin, Lynch, Cop-lan, Rose-Krasnor, & Booth, 1994). Conversely, it fol-lows that when deciding whether to exclude a personfrom a group, the more an individual perceives thatperson to be different from him- or herself, the morelikely the individual will be willing to exclude thatperson. Assessing children’s perceptions of how sim-ilar or different they are with regard to the excludedchild provides a measure of what has been referred toas a tolerance of differences in others (see Crystal etal., 2000). Children who say they are different fromthe excluded child yet want to include that child intheir group are demonstrating an ability to be tolerantof others who are different from the self. Childrenwho say they are similar to the excluded child andwant to include him or her are acknowledging theirsimilarity with the excluded child. Children whoagree with a decision to exclude and judge them-selves to be different from the excluded child may beseen as expressing an intolerance of differences. Thus,in the current study, in addition to questions about ex-clusion, we also asked children how similar or differ-ent they were to the excluded child.

Cultural theorists have characterized Japan as acollectivistic culture whose members value confor-mity, in contrast to the United States, which is an in-dividualistic culture whose members value auton-omy (Triandis, 19, 1995). Based on this dichotomy,it might be expected that Japanese children and ado-lescents would be more likely than their Americancounterparts to fall into the “intolerant” group (thosewho say they are different from the excluded childand want to exclude him or her), whereas Americanyouth would be more likely than their Japanese peersto be categorized as “tolerant” (those who say theyare different from the excluded child but who want to

1790Child Development

include him or her in the group). Previous studies,however, that have speciﬁcally investigated Ameri-can and Japanese children’s response to a variety ofatypical peers, have found that intolerance is highlydependent on the context (Crystal, Watanabe, & Chin,1997). These ﬁndings suggest that reactions to devi-ant or atypical peers cannot be predicted on the basisof group membership in so-called “individualistic” or“collectivistic” cultures, as has often been proposedin the cross-cultural literature (see Markus & Ki-tayama, 1991; Triandis, 19). In light of previousdomain-speciﬁc ﬁndings and the basic tenets ofsocial–cognitive domain theory (see Turiel, 1983, 1998),we expected that context rather than culture wouldbe most inﬂuential in determining exclusion judg-ments and the various response styles to tolerance ofdifferences among Japanese and American students.As noted previously, peer group exclusion basedon individual-level attributes differs from that basedon race or gender because of the potential alterabilityof the characteristics that make the target personatypical or unappealing. The alterability of the cri-teria for exclusion raises the question of conformity,and the extent to which the target person might be ex-pected to change his or her characteristics to gain peeracceptance. For this reason, we also asked childrenhow much the excluded child should change him- orherself to be accepted by the group.

Previous cross-cultural research on conformity hasproduced mixed results. For example, despite stereo-typic descriptions of members of collectivistic cul-tures as being strongly conformist, and members ofindividualistic cultures as being strongly indepen-dent in their actions and opinions (e.g., Triandis,19), Spiro (1993) found that members of a tradition-ally “collectivistic” society (e.g., Japanese villagers andfarmers) had strong individualistic goals and personalpriorities. In fact, Spiro argued that it is necessary todistinguish cultural ideologies from psychologicalrealities, and he questioned whether the constructsof individualism and collectivism actually reﬂectthe psychological values held by members of diversecultures. Thus, because of the contradictions in the lit-erature, we made no speciﬁc cross-cultural hypothe-ses regarding expectations that the excluded childchange him- or herself for group acceptance.

Given the substantial changes in the importanceand awareness of the peer group that occurs betweenchildhood and adolescence (Gavin & Furman, 19;O’Brien & Bierman, 1988), it is reasonable to expectthat judgments about peer group exclusion and pre-scriptions for conformity would be signiﬁcantly re-lated to the age of the respondent. To determinewhether such judgments change with age in similarways among students from diverse cultures, we sur-veyed children in fourth, seventh, and tenth grades inboth the United States and Japan. Because previousstudies have shown that older children are more con-cerned about group functioning than are youngerchildren (Killen & Stangor, 2001; Turiel, 1983), it washypothesized that younger children would be lesslikely than would older children to view individual-level differences as jeopardizing group harmony, andwould thus be less willing to exclude those differentfrom themselves in contexts in which group function-ing was at stake. Thus, we predicted that tolerance ofdifferences for both Japanese and American childrenwould be greater in fourth-grade than in tenth-gradejudgments. Additionally, in view of the importance ofphysical appearance to adolescent self-esteem in bothcultures (Harter, 1998; Mukai, 1996), and the depres-sive moods and symptoms to which both Japanese andAmerican adolescents are often prone (Takeuchi, Rob-erts, & Suzuki, 1995), we expected tolerance of differ-ences toward the peer who was unconventional in ap-pearance or personality to decrease with age, overall.With regard to prescriptions for conformity, de-velopmental theory suggests that the need for self-differentiation and autonomy increases throughoutthe adolescent period, as the person focuses on identityformation and individuation from parents and peers(Erikson, 1968; Nucci, 1996; Smetana, 1988). Such aprocess would seem to imply a greater recognition inolder than in younger children of the importance ofresisting peer group pressure and “being yourself.”Therefore, at the same time that adolescents are be-coming increasingly concerned with group function-ing and conformity, they are also concerned with au-tonomy and resisting conforming to the group. Thesetwo goals appear to be contradictory, especially if thecontext of the interaction is not taken into account.We proposed, however, that analyses of the contextof judgments about exclusion would reveal that thesetwo seemingly conﬂicting values—group functioningand autonomy—often coexist in adolescence. For ex-ample, in one context, adolescents may exclude some-one, and judge that that person needs to change theirbehavior to ﬁt the group (endorsing conformity). Inanother context, however, adolescents may deem itwrong to exclude someone, and may believe that theperson does not need to change their behavior to ﬁtthe group (endorsing autonomy). Thus, although ad-olescents may view exclusion as legitimate in somecontexts due to their increasing concerns with confor-mity and group functioning, they will also view theprescription that atypical peers change themselves toﬁt the group as an infringement on that peer’s auton-omous decision making. Only when an analysis takes

into account context will it effectively demonstratethe coexistence of these orientations. Further, an in-teresting question is whether this coexistence extendsto adolescents in non-Western cultures.

Studies conducted with Western participants haveshown that conformity peaks in ninth grade and ado-lescents then become increasingly dissatisﬁed withthe conformity demands of peer groups, and, in fact,support less group conformity (Berndt, 1992; Brown,Eicher, & Petrie, 1986; Gavin & Furman, 19). Al-though we could ﬁnd no developmental studies ofconformity in Japan, recent research on East Asian ad-olescents has shown a strong sense of individuality inthese samples (Helwig, Arnold, Tan, & Boyd, 2001;Yau & Smetana, 1993, 1996). In light of these theoreti-cal and empirical considerations, we predicted that,with age, both American and Japanese adolescentswould resist pressures to conform to the group insome contexts, and not in others.

We also examined gender differences becauseprior research has shown that females are often moresensitive to issues of exclusion than are males (Killen& Stangor, 2001; Theimer et al., 2001). Some theorieshave speculated that females’ more frequent experi-ence with exclusion—for example, in the realm ofsports—contributes to their greater sensitivity to it(Killen et al., 2002). Thus, in terms of tolerance of dif-ferences, we predicted that in both cultures, femaleswould be more likely than would males to evaluateexclusion using a tolerance of differences perspective.In view of females’ greater experience with and sensi-tivity to exclusion, we also expected that females, rel-ative to males, would be less likely to prescribe con-formity for the atypical child. However, we alsoexpected that gender would interact with culture. Giventhe stronger view of women’s rights in the UnitedStates than in Japan (Lebra, 1993), it was expected thatAmerican girls would express a stronger view of thewrongfulness of exclusion than would Japanese girls.How this would affect comparisons between Ameri-can and Japanese boys was not known.

In summary, our predictions in this study were asfollows: With regard to judgments about the wrong-fulness of exclusion, we expected such judgments tobe more a function of context than of cultural mem-bership; that, with age, children would be more likelyto view exclusion of an individual who disrupts thegroup functioning as legitimate; and that girls wouldbe more sensitive to exclusion than would boys. Withregard to tolerance of differences, we again expectedthat context would be more inﬂuential than wouldculture in determining whether children were toler-ant or willing to maintain the status quo; that fourth-grade students would be more likely than would

Killen, Crystal, and Watanabe1791

tenth-grade students to demonstrate a tolerant atti-tude toward differences, overall, and, in particular,toward the unconventional peer with regard to dressand personality; and that females would be morelikely than would males to hold a tolerance of differ-ences perspective on exclusion. Finally, with regard toprescriptions for conformity, we predicted that expec-tations for behavioral change would, with age, varyby context for both the United States and Japanesestudents, and that females would be less likely thanwould males to expect behavioral change.METHOD

Participants.A total of 1,057 students participatedin this study: 513 Japanese student participants (150fourth graders, 91 males and 59 females; 126 seventhgraders, 65 males and 61 females; and 237 tenth graders,110 male and 127 females) from the suburbs of Mat-suyama, Japan (a middle-sized metropolitan area), and544 U.S. students (119 fourth graders, 53 males and 66females; 230 seventh graders, 102 males and 126 fe-males; and 195 tenth graders, 103 males and 92 females)from the suburbs of Washington, DC (a middle-sizedmetropolitan area). There were no signiﬁcant differ-ences in the mean ages of the children in the UnitedStates and Japan (fourth grade, M ϭ 9,6; seventhgrade, M ϭ 12,6; tenth grade, M ϭ 15,3). These loca-tions were chosen because they reﬂected two diversecultures—East Asian and North American—andwere comparable in size and diversity. The ethnicbreakdown for the U.S. schools that participated wasas follows: 61.3% European American, 18.2% AfricanAmerican, 12.4% Hispanic, and 8% Asian American.The participants in the Japanese schools were 100%Japanese. All students from both cultures were fromworking-class to middle-class backgrounds.

Procedure.A professor and several graduate stu-dent assistants explained the purpose of the surveyand distributed it to all children during special class-time sessions devoted to the survey administration.Children who did not want to participate or who didnot have parental consent were given an alternativetask by the classroom teacher. Two of the coauthors,ﬂuent in both Japanese and English, supervised thetranslations of the surveys and all materials. Childrenwere allowed to raise their hands and ask questions atany time during the session. All names were speciﬁcto the culture and matched the gender of the partici-pant. Children were told that their answers would beconﬁdential and anonymous and that their participa-tion was strictly voluntary.

Survey items.Students were asked to evaluate sixtypes of reasons given by children for excluding a

1792Child Development

peer from a group activity: (1) aggressive behavior(Fight; excluding someone who ﬁghts with otherswhen going shopping at the mall), (2) unconvention-ality in dress (Hair; excluding someone who dyes hisor her hair green and wears unusual clothes when go-ing to a fancy restaurant), (3) unconventionality inpublic behavior (Clown; excluding someone who actslike a clown when going to the movies), (4) cross-genderbehavior (Cross-Gender; excluding a boy who actslike a girl or a girl who acts like a boy when forminga social club), (5) slowness in sports (Sports; exclud-ing someone who is a slow runner when entering atown race), and (6) personality (Sad; excluding some-one who acts sad and lonely when going on a picnic);see the Appendix for a complete description of eachscenario. Six random orders of presentation of theitems were generated and preliminary analyses indi-cated that there were no signiﬁcant order effects.Using 5-point Likert scales, students were asked threequestions: (1) “How much do you agree or disagreewith the decision not to ask X?” (1 ϭ “disagree com-pletely” to 5 ϭ “agree completely”); (2) “How muchshould X change him/herself to be accepted by thegroup?” (1 ϭ “not at all” to 5 ϭ “very much”); and (3)“How similar to or different from X are you?” (1 ϭ“very similar” to 5 ϭ “very different”). (X was thename of the excluded child; Japanese names wereused in Japan, and American names were used in theUnited States)

Tolerance of differences.To understand the ratingsof self-perceived difference in relation to the exclu-sion judgments, we created new variables represent-ing children’s tolerance of differences (see Crystal etal., 2000). Three categories were created for each sce-nario: (1) Tolerance of differences (students who rated1, 2, or 3—“disagreed with exclusion”—on the exclu-sion scale, and 4 or 5—“I’m different”—on the per-ceived difference scale), (2) Similarity with the ex-cluded child (students who rated 1, 2, or 3—“disagreedwith exclusion”—on the exclusion scale, and 1 or 2—“I’m similar”—on the perceived difference scale), and(3) Intolerance of differences (students who rated 4or 5—“agreed with exclusion”—on the exclusionscale, and 4 or 5—“I’m different”—on the perceiveddifference scale). A fourth category that included par-ticipants who rated 4 or 5 on the exclusion scale, and1, 2, or 3 on the similarity scale was also created. Wedid not include this category in the analyses, how-ever, because the frequencies were very low (for Tol-erance of differences, M ϭ .08, range ϭ .04–.15) andnot easily interpretable from our conceptual frame-work. For each context, variables were coded as “1”for children who fell into each category, and “0” for allother children in the sample who were not in the

category. These codes were mutually exclusive; inany given context, children could fall into one andonly one category.RESULTS

Repeated-measures ANOVAs were conducted on therating scores for all children. The between-subjectsfactors were country (Japan and United States), gender,and grade (fourth, seventh, and tenth); the within-subjects factor was context of exclusion (Fight, Hair,Clown, Cross-Gender, Sports, and Sad). Follow-upunivariate ANOVAs were conducted as well as Bon-ferroni and Tukey tests to test for within-subjects andbetween-subjects differences. Due to the abundanceof ﬁndings, summaries will refer primarily to resultsof hypothesis testing.

Ratings of peer exclusion.The ﬁrst set of analysestested the general hypothesis that context, rather thanculture, was a signiﬁcant factor contributing to chil-dren’s decision to exclude an atypical peer. A 2 (coun-try) ϫ 2 (gender) ϫ 3 (grade) ϫ 6 (contexts of exclu-sion) ANOVA with repeated measures on the lastfactor for the 1 (disagree) to 5 (agree) exclusion rat-ings indicated signiﬁcant between-subjects effects forgender, F(1, 4970) ϭ 50.852, p Ͻ .001, and a Country ϫGender interaction, F(1, 4970) ϭ 18.8, p Ͻ .019. Nooverall country effect emerged. In accord with predic-tions, within-subjects effects were found for the con-text of exclusion, F(5, 4970) ϭ 29.69, p Ͻ .001; Contextof Exclusion ϫ Country, F(5, 4970) ϭ 6.405, p Ͻ .001;Context of Exclusion ϫ Grade, F(10, 4970) ϭ 4.279,p Ͻ .001; Context of Exclusion ϫ Gender, F(5, 4970) ϭ16.785, p Ͻ .001; and Context of Exclusion ϫ Country ϫGender, F(5, 4970) ϭ 5.8, p Ͻ .001.

As shown in Table 1, children and adolescentsevaluated the six contexts of exclusion quite differ-ently, and, in line with hypotheses, there were nooverall differences between the Japanese and Americanchildren’s evaluations. Follow-up one-way ANOVAsrevealed that the Scenario ϫ Country effect was a re-sult of higher ratings of exclusion (all right to ex-clude) among Japanese participants relative to Amer-ican participants for two of the six scenarios, Hair,F(1, 1207) ϭ 10.33, p Ͻ .001, and Sad, F(1, 1179) ϭ10.327, p Ͻ .001. Follow-up analyses for the Context ofExclusion ϫ Grade effect revealed grade differencesfor two of the six exclusion contexts, Fight, F(3, 1212) ϭ7.30, p Ͻ .001, and Sports, F(3, 1187) ϭ 10.02, p Ͻ .001.In both cases, consonant with hypotheses, tenth-grade participants judged it more legitimate to excludesomeone who was aggressive (fourth and tenth graders,M ϭ 2.90, SD ϭ 1.25 and M ϭ 3.30, SD ϭ 1.14, respec-tively) or who was a slow runner (fourth and tenth

Killen, Crystal, and Watanabe

Table 1Japanese and American Students’ Ratings of Agreement with Peer Exclusion

Scenario

Country by GenderJapanMaleFemaleTotalUnited StatesMale FemaleTotalTotal Male Female

Fight

Hair

Clown

Cross-Gender

Sports

Sad

1793

Total

3.15 (1.24)3.02 (10.6)3.09 (1.06)3.15 (1.22)3.21 (1.18)3.18 (1.20)3.14 (1.24)3.13 (1.13)

2.79 (1.26)2.73 (1.13)2.76 (1.20)2.93 (1.45)2.22 (1.21)2.56 (1.38)2.83 (1.36)2.47 (1.21)

2.83 (1.23)2.63 (1.03)2.73 (1.14)2.69 (1.34)2.58 (1.23)2.63 (1.28)2.77 (1.28)2.59 (1.13)

2.97 (1.25)2.28 (1.01)2. (1.19)3.31 (1.47)2.21 (1.27)2.73 (1.48)3.14 (1.37)2.25 (1.15)

2.79 (1.34)2.44 (1.20)2.62 (1.29)3.03 (1.50)2.50 (1.30)2.75 (1.42)2.91 (1.42)2.47 (1.26)

2.70 (1.25)2.69 (1.17)2.69 (1.21)2.66 (1.35)2.16 (1.21)2.40 (1.30)2.68 (1.30)2.42 (1.21)

2.76 (1.19)

2.70 (1.34)2.93 (1.32)2.55 (1.18)

Note:N ϭ 1,057. Responses to “How much do you agree or disagree with the decision not to ask X?” Rating scale: 1 ϭ “disagree com-pletely” to 5 ϭ “agree completely.” Values in parentheses are standard deviations.

graders, M ϭ 2.47, SD ϭ 1.35 and M ϭ 2.90, SD ϭ 1.35,respectively) than did fourth-grade participants.

As predicted, there was a signiﬁcant main effect forgender, with boys, overall, more willing to excludethan were girls (see Table 1). Follow-up ANOVAsshowed this pattern to occur, speciﬁcally, in four ofthe six contexts: Cross-Gender, F(1, 1026) ϭ 126.57,p Ͻ .001; Hair, F(1, 1038) ϭ 20.38, p Ͻ .001; Sad,F(1, 1011) ϭ 11.10, p Ͻ .001; and Sports, F(1, 1019) ϭ26.29, p Ͻ .001).

Overall, gender differences were greater in theUnited States than in Japan. Signiﬁcant Country ϫGender interactions, however, were found, in partic-ular, for three contexts: Cross-Gender, F(1, 1026) ϭ8.20, p Ͻ .01; Hair, F(1, 1038) ϭ 10.72, p Ͻ .001; andSad, F(1, 1011) ϭ 10.66, p Ͻ .001. Further analyses in-dicated that with regard to the cross-gendered peer,American boys were more willing to exclude thanwere their Japanese counterparts, F(1, 508) ϭ 8.09, p Ͻ.01, whereas no differences emerged between Ameri-can and Japanese girls. In contrast, in both the Hair,F(1, 524) ϭ 21.87, p Ͻ .001, and the Sad, F(1, 506) ϭ25.39, p Ͻ .001, scenarios, the only signiﬁcant cross-cultural differences were found among the girls, withJapanese participants more willing than Americanparticipants to exclude the atypical peers.

In summary, the prediction that context rather thanculture would be the most signiﬁcant and consistentfactor inﬂuencing exclusion judgments was borne outby the analyses of the ratings of exclusion. No overalldifferences between cultures, or between grade lev-els, emerged. Next to context differences, those relatedto gender proved to be most salient, with Americanboys most willing to exclude, followed by Japaneseboys, Japanese girls, and American girls. Unexpect-edly, gender differences were more pronounced in theUnited States than in the Japanese sample.

Ratings of self-perceived differences.Mean ratings ofstudents’ self-perceived differences from the atypicalpeers, broken down by culture and gender, are pre-sented in Table 2. A 2 (country) ϫ 2 (gender) ϫ 3(grade) ϫ 6 (context of exclusion) ANOVA with re-peated measures on the last factor (for the 1 ϭ “verysimilar” to 5 ϭ “very different” judgments) was con-ducted to analyze the self-perceived difference rat-ings. Rather than report detailed results of these anal-yses we will simply summarize the main ﬁndings,given that we did not formulate speciﬁc hypothesesregarding self-perceptions of difference.

In general, children in both countries differentiatedamong the reasons for exclusion when judging self-perceived differences as indicated by a within-subjectscontext effect, F(5, 4940) ϭ 91.33, p Ͻ .001. Childrenjudged that they were most like the child who actedlike a clown (M ϭ 3.42, SD ϭ 1.27) and least like thechild who dyed his or her hair green (M ϭ 4.33, SD ϭ1.07), p Ͻ .001. There were no between-subjects effects.Other within-subjects effects were found for Context ϫCountry, F(5, 4940) ϭ 8.93, p Ͻ .001; Context ϫ Grade,F(10, 4940) ϭ 9., p Ͻ .001, Context ϫ Gender,F(5, 4940) ϭ 46.03, p Ͻ .001; and Context ϫ Countryϫ Grade, F(10, 4940) ϭ 4.73, p Ͻ .001. Follow-up testsrevealed that in terms of culture, American studentsrated themselves as being more different from thecross-gendered peer than did Japanese students,who, in turn, rated themselves as being more differ-ent from the sad peer than did their American coun-terparts, ps Ͻ .001. In terms of development, with age,

1794Child Development

Table 2Japanese and American Students’ Ratings of Self-Perceived Differences

Scenario

Country by GenderJapanMaleFemaleTotalUnited StatesMaleFemaleTotal

Fight

Hair

Clown

Cross-Gender

Sports

Sad

3.58 (1.10)3.81 (1.11)3.69 (1.11)3.71 (1.31)4.16 (1.18)3.95 (1.26)

4.32 (1.03)4.43 (.91)4.37 (.98)4.23 (1.17)4.30 (1.13)4.27 (1.15)

3.49 (1.20)3.42 (1.16)3.46 (1.18)3.17 (1.34)3.55 (1.33)3.37 (1.36)

4.38 (.94)3.35 (1.19)3.88 (1.19)4.46 (1.09)3.77 (1.36)4.10 (1.29)

3.76 (1.23)3.33 (1.29)3.55 (1.28)3.81 (1.36)3.62 (1.34)3.71 (1.35)

4.10 (1.09)4.19 (1.00)4.14 (1.05)3.86 (1.26)4.07 (1.20)3.97 (1.24)

Note:N ϭ 1,057. Responses to “How similar to or different from X are you?” Rating scale: 1 ϭ “very similar” to 5 ϭ “very different.”Values in parentheses are standard deviations.

students increasingly saw themselves as more differ-ent from the aggressive child, p Ͻ .001, but less differentfrom the child who dyed his or her hair green, theslow runner, and the child with the sad personality,ps Ͻ .001. Finally, in terms of gender, girls reportedthat they were more different from the aggressivechild, but less different from the cross-gendered childand the slow runner than did the boys, ps Ͻ .001.Tolerance of differences.We predicted that contextwould be more inﬂuential than would culture in de-termining whether children were tolerant or intoler-ant of differences in others. We also hypothesized thatyounger students would be more likely than wouldolder students to respond in a tolerant manner, andlastly, that females would be more likely than wouldmales to be tolerant of differences.

To test these hypotheses, separate 2 (country) ϫ 2(gender) ϫ 3 (grade) ϫ 6 (contexts of exclusion)ANOVAs with repeated measures on the last factorwere conducted for each new variable (Tolerance,Similarity, Intolerance; see Table 3 for the proportionsfor each variable for each context by culture). Becausethe actual proportions for the Similarity categorywere very low, the analyses conducted on this vari-able are not reported (see Table 3).

Analyses for the Tolerance of differences categoryrevealed signiﬁcant between-subjects effects for grade,F(2, 991) ϭ 7.583, p Ͻ .001; gender, F(1, 991) ϭ 10.520,p Ͻ .001; and Country ϫ Gender, F(1, 991) ϭ 9.508,p Ͻ .002. As hypothesized, with age, children wereless likely to be tolerant of differences, and girlswere more likely to be tolerant of differences thanwere boys. The gender gap was signiﬁcantly larger inthe United States than in Japan, especially for the so-cially disruptive issues, such as wearing strangeclothes and acting like a clown, and also for display-ing a sad personality.

Consistent with hypotheses, within-subjects effectswere found for the context of exclusion, F(5, 4955) ϭ36.401, p Ͻ .001; Context ϫ Grade, F(10, 4955) ϭ 3.582,p Ͻ .001; and Context ϫ Gender, F(5, 4955) ϭ 4.136,p Ͻ .001. There were no between-subjects effects forculture. The largest grade effects were found for theHair, F(2, 1001) ϭ 5.33, p Ͻ .005 (M ϭ .61, SD ϭ .48;

Table 3Proportions of Japanese and American Students Using Each Tolerance of Differences Category

Country by Tolerance of Differences Category

Japan

ScenarioFightHairClownGenderSportsSad

Tolerance.32 (.47).53 (.50).33 (.47).43 (.50).37 (.48).49 (.50)

Similarity.10 (.30).39 (.19).19 (.39).13 (.34).19 (.39).06 (.24)

Intolerance.22 (.41).23 (.42).14 (.35).18 (.39).16 (.37).21 (.41)

Tolerance.35 (.48).52 (.50).30 (.46).41 (.49).37 (.48).48 (.50)

United StatesSimilarity.14 (.34).09 (.29).29 (.45).15 (.35).16 (.36).13 (.33)

Intolerance.33 (.47).24 (.43).19 (.39).30 (.46).23 (.42).18 (.38)

Note:N ϭ 1,057. Proportions cannot exceed 1.00. Values in parentheses are standard deviations.

M ϭ .52, SD ϭ .50; and M ϭ .48, SD ϭ .50 for fourth,seventh, and tenth grades, respectively); Sports,F(2, 1001) ϭ 9.42, p Ͻ .001 (M ϭ .46, SD ϭ .49; M ϭ .39,SD ϭ .48; and M ϭ .30, SD ϭ .45, respectively); and Sad,F(2, 1001) ϭ 8.129, p Ͻ .001 (M ϭ .58, SD ϭ .49; M ϭ .49,SD ϭ .48; and M ϭ .42, SD ϭ .45, respectively) scenar-ios, in all of which, in accord with predictions, fourthgraders were more tolerant than were tenth graders.Analyses of the Intolerance of differences categoryrevealed between-subjects effects for country,F(1, 992) ϭ 18.31, p Ͻ .001, and gender, F(1, 992) ϭ29.17, p Ͻ .001. Contrary to hypotheses, Americanstudents (M ϭ .25, SD ϭ .24) were more likely to en-dorse the Intolerance of differences category thanwere Japanese students (M ϭ .19, SD ϭ .22). In linewith predictions, males (M ϭ .26, SD ϭ .24) weremore willing to exclude others who were differentfrom them than were females (M ϭ .18, SD ϭ .21).Within-subjects interaction effects, however, qualiﬁedthese ﬁndings. Consonant with expectations, within-subjects effects were found for context, F(5, 4830) ϭ9.22, p Ͻ .001; Context ϫ Country, F(5, 4830) ϭ 6.60, p Ͻ.001, Context ϫ Grade, F(10, 4830) ϭ 2.65, p Ͻ .05, Con-text ϫ Gender, F(5, 4830) ϭ 17.80, p Ͻ .001; and Con-text ϫ Country ϫ Gender F(5, 4830) ϭ 4.21, p Ͻ .001.The interaction effects revealed that U.S. childrenwere more intolerant of differences than were theirJapanese counterparts for four of the six scenarios:Clown, F(1, 1034) ϭ 4.15, p Ͻ .05; Fight, F(1, 1045) ϭ16.94, p Ͻ .001; Cross-Gender, F(1, 1025) ϭ 18.87, p Ͻ.001; and Sports, F(1, 1019) ϭ 7.43, p Ͻ .001. A gradeeffect was found only for the Fight scenario, F(2, 1045) ϭ11.30, p Ͻ .001. With age, Japanese and U.S. childrenwere more willing to be intolerant of someone who isaggressive (M ϭ .18, SD ϭ .38; M ϭ .27, SD ϭ .44; andM ϭ .34, SD ϭ .47, for fourth, seventh, and tenthgrades, respectively); follow-up t tests showed signif-icance between fourth and seventh grades, p Ͻ .03;between fourth and tenth grades, p Ͻ .001; and be-tween seventh and tenth grades, p Ͻ .05. Regardinggender differences, U.S. boys were more intolerant ofdifferences for the Cross-Gender (M ϭ .48, SD ϭ .50),Hair (M ϭ .32, SD ϭ .47), and Sad (M ϭ .22, SD ϭ .42)scenarios than were U.S. girls (M ϭ .13, SD ϭ .34; M ϭ.17, SD ϭ .37; and M ϭ .14, SD ϭ .34, respectively). Incontrast, Japanese boys were more intolerant of dif-ferences in the Clown (M ϭ .20, SD ϭ .40), Cross-Gender (M ϭ .28, SD ϭ .45), and Sports (M ϭ .20,SD ϭ .40) scenarios than were their female peers (M ϭ.09, SD ϭ .29; M ϭ .08, SD ϭ .27; and M ϭ .11, SD ϭ.32, respectively)

In summary, our hypotheses regarding culturewere partially supported. No between-subjects effectsof culture were found for the Tolerance of differencesKillen, Crystal, and Watanabe1795

category, but, overall, more American than Japanesestudents endorsed the Intolerance of differences re-sponse style (for four of the six scenarios). Within-subjects effects for context and its interactions withculture, grade, and gender were found for both re-sponse styles. Consistent with hypotheses, fourthgraders were more likely than were tenth graders tohold a tolerance of differences perspective for half ofthe scenarios, and speciﬁcally so, with regard to thechild who had green hair, the child who was a slowrunner, and the child who was sad. Lastly, as hypoth-esized, females were more likely than were males torespond in a tolerance of differences manner.

Prescriptions for conformity.To determine the effectsof culture, grade, gender, and context on prescrip-tions for conformity, a 2 (country) ϫ 2 (gender) ϫ 3(grade) ϫ 6 (type of exclusion) ANOVA with the re-peated measures on the last factor was conducted onthe conformity ratings, and revealed signiﬁcantbetween-subjects effects for culture, F(1, 986) ϭ7.947, p Ͻ .005; grade, F(2, 986) ϭ 23.00, p Ͻ .001; gender,F(1, 986) ϭ 91. 814, p Ͻ .001; and Culture ϫ Grade,F(2, 986) ϭ 3.70, p Ͻ .025. The within-subjects effectswere all signiﬁcant except for the four-way interac-tion and 1 three-way interaction. We report the three-way interactions, which qualiﬁed the two-way inter-actions: Type of Exclusion ϫ Country ϫ Grade,F(5, 4930) ϭ 5.596, p Ͻ .001, and Type of Exclusion ϫCountry ϫ Gender, F(5, 4930) ϭ 3.65, p Ͻ .003. Pro-portions for the Prescriptions for Conformity variableare presented in Table 4.

In general, the between-subjects ﬁndings indicatedthat fourth graders (M ϭ 2.78, SD ϭ 1.24) expectedmore behavioral conformity from the atypical childthan did older children (M ϭ 2.59, SD ϭ 1.22); and, aspredicted, males (M ϭ 3.0, SD ϭ 1.29) expected morebehavioral conformity than did females (M ϭ 2.51,SD ϭ 1.16). In addition, Japanese children (M ϭ 2.86,SD ϭ 1.17) thought that the child should change him-or herself for group acceptance more than did theAmerican children (M ϭ 2.68, SD ϭ 1.32). The inter-action effects qualiﬁed these general ﬁndings. Due tothe multitude of ﬁndings, we only report the ﬁndingsthat were exceptions to the general pattern of results.Thus, although conformity decreased with age over-all, Japanese students’ expectations of behavioral con-formity in the Fight scenario increased with age,F(2, 846) ϭ 5.24, p Ͻ .005. At the same time, Americanstudents’ expectations for conformity in the Fightscenario decreased with age, F(2, 805) ϭ 9.63, p Ͻ.002. American students expected more behavioralconformity than did Japanese students for the scenarioabout excluding the slow runner (Sports), F(1, 1003) ϭ5.036, p Ͻ .025. With regard to the cross-gendered

1796Child Development

Table 4Japanese and American Students’ Prescriptions for Conformity

Scenario

Fight

Japan Male

Grade 4Grade 7Grade 10TotalFemaleGrade 4Grade 7Grade 10TotalTotal United StatesMale

Grade 4Grade 7Grade 10TotalFemaleGrade 4Grade 7Grade 10TotalTotal

Hair

Clown

Cross-Gender

Sports

Sad

3.45 (1.13)3.63 (1.02)3.52 (1.09)3.52 (1.09)3.12 (1.07)3.28 (1.08)3.48 (1.00)3.34 (1.04)3.44 (1.07)

3.56 (1.07)3.29 (1.45)2.72 (1.36)3.15 (1.34)3.32 (1.15)3.31 (1.20)2.38 (1.05)2.84 (1.20)3.00 (1.28)

2. (1.00)2.66 (1.03)2. (1.27)2.83 (1.13)2.63 (.96)2.39 (.90)2.50 (.)2.50 (.91)2.67 (1.04)

3.47 (1.20)3.12 (1.39)3.16 (1.30)3.26 (1.29)2.63 (1.03)2.49 (1.19)2.10 (.88)2.32 (1.02)2.81 (1.26)

2.71 (1.36)2.68 (1.37)2.48 (1.44)2.61 (1.39)2.02 (.96)2.11 (1.17)1.70 (1.03)1.88 (1.06)2.26 (1.30)

3.25 (1.22)3.15 (1.34)3.12 (1.24)3.17 (1.25)2.78 (1.18)3.10 (1.03)2.83 (1.11)2.88 (1.11)3.03 (1.19)

3.77 (1.19)3.23 (1.22)3.31 (1.06)3.38 (1.16)3.71 (1.05)2.99 (1.31)2.98 (1.11)3.16 (1.22)3.26 (1.20)

2.90 (1.51)2.77 (1.43)2.76 (1.48)2.79 (1.46)2.71 (1.43)2.43 (1.36)1.72 (1.07)2.26 (1.35)2.51 (1.43)

3.25 (1.25)2.80 (1.25)2.35 (1.09)2.71 (1.23)3.14 (1.24)2.37 (1.18)1.98 (.96)2.42 (1.20)2.56 (1.22)

3.94 (1.25)3.53 (1.45)3.16 (1.51)3.46 (1.46)2.35 (1.22)2.26 (1.30)1.78 (1.11)2.12 (1.24)2.76 (1.51)

3.00 (1.45)2.81 (1.44)2.56 (1.40)2.75 (1.43)2.37 (1.32)2.20 (1.17)1. (1.15)2.13 (1.21)2.43 (1.35)

2.92 (1.38)2.86 (1.33)2.72 (1.13)2.81 (1.26)2.48 (1.04)2.57 (1.15)2.23 (1.21)2.43 (1.15)2.61 (1.22)

Note:N ϭ 1,057. Responses to “How much should X change him or herself to be accepted by the group?” Rating scale: 1 ϭ “not at all” to 5 ϭ“very much.” Values in parentheses are standard deviations.

peer, Japanese students did not expect more con-formity than did American students (no signiﬁcantdifferences).

In summary, analyses revealed that, as hypothe-sized, American fourth graders prescribed more con-formity than did their tenth-grade counterparts in allbut one of the contexts (see Table 4). Japanese stu-dents showed the same pattern, except in the Fightscenario, in which conformity decreased with age.Also, as hypothesized, boys were more likely thanwere girls to expect conformity from the atypicalchild in all six scenarios. Further, Japanese studentsexpected more behavioral change than did Americanstudents in the scenarios about the aggressive child,the green-haired child, and the sad child; whereasAmerican students expected more behavioral changethan did Japanese students in the scenario about theslow runner.DISCUSSION

This study examined Japanese and American chil-dren’s judgments regarding peer group exclusion,their tolerance for differences in others, and their pre-scriptions for conformity as manifested in expecta-tions of behavioral change in atypical peers. Becausethe analyses tested hypotheses related to culture, gen-der, and age, we will discuss each of these dimensionsseparately, beginning with culture.

In accord with the premises of social–cognitive do-main theory (Turiel, 1983, 1998), context, rather thanculture, was the most consistent predictor of chil-dren’s evaluations of exclusion, tolerance of differ-ences, and prescriptions for conformity. Children’sjudgments about exclusion, tolerance for differences,and prescriptions for conformity depended on the con-text of exclusion, and there were few overall culturaldifferences. Collectivism, characterized by groupism,conformity, and adherence to duty, has been attrib-uted to Japanese culture (see Gudykunst, Yoon, &Tsukasa, 1987; Matsumoto, 1990; Peak, 19; Trian-dis, 1995). Based on this view, one would expect thatJapanese children would respond in a uniform man-ner regarding the exclusion of a member from agroup, and that such a response would differ dramat-ically from American children’s evaluations. How-

ever, this was not the case in the ﬁndings of thepresent study. The majority of all children judgedthat it was wrong to exclude a child from a group,and there was also variability within both Japaneseand American cultures regarding when exclusionwas legitimate. In both countries, children gave prior-ity to group functioning in some contexts (e.g., mak-ing a sports team successful, not disrupting the atmo-sphere of a fancy restaurant by wearing strangeclothes), and gave priority to the individual in othercontexts (e.g., not excluding a child just because heor she acts like a clown, or just because he or she actslike someone of the opposite gender). On the whole,these ﬁndings support the coexistence view of cultureespoused by theorists who have moved beyond theuse of dichotomous templates, such as individual-ism and collectivism, to describe cultures (Greenﬁeld& Cocking, 1994; Harkness et al., 2000; Killen & Wain-ryb, 2000; Turiel, 1998; Turiel & Wainryb, 1994; Wainryb& Turiel, 1994).

Coexistence of cultural orientations was even moreevident in the analyses on children’s prescriptions forconformity. A widely held cultural stereotype foundin the Western psychological literature as well as thepopular media is that members of collectivistic cul-tures, particularly the Japanese, are highly conformistin their attitudes and behavior. However, the presentstudy’s data for ratings of conformity revealed age,gender, context, and culture differences; culture was afactor, but so were all of the other participant vari-ables. Moreover, no overall cross-cultural disparitiesin conformity ratings were found in the Clown or theCross-Gender contexts, and, in the Sports scenario,American students prescribed more conformity for theatypical child than did Japanese students. The patternof within-cultural heterogeneity and cross-culturalsimilarity in expectations of behavioral change is con-sistent with other research that presents a complexand nuanced view of Japanese and American socie-ties (e.g., Crystal, Watanabe, Weinfurt, & Wu, 1998;Hamilton, Blumenfeld, Akoh, & Miura, 1991; Lewis,1995; Rosenberg, 1992; Spiro, 1993).

Next to context, gender was the strongest predictorof judgments about exclusion, tolerance of differ-ences, and prescriptions for conformity. This studyextends previous work that has shown that femalesare more sensitive to issues of exclusion than aremales (Killen et al., 2001; Killen & Stangor, 2001;Stretch, 2000; Theimer et al., 2001) by ﬁnding thesame gender pattern in a non-U.S. sample. One sur-prising ﬁnding was that the gender gap for judg-ments about exclusion was larger in the United Statesthan in Japan. As an example, American boys had thehighest exclusion ratings (willingness to exclude)Killen, Crystal, and Watanabe1797

with regard to the child who dyed his or her hairgreen and American girls had the lowest ratings; Jap-anese boys and girls fell in between. This examplealso illustrates the interaction between gender andculture, because the Japanese participants, on the whole,were more willing to exclude a child for dying his orher hair green than were the Americans; it was onlybecause American girls were so unwilling to excludefor this reason—in contrast to the greater willingnessof American boys—that the overall American scorewas lower than that of the Japanese.

The differences between American and Japanesegirls may be due to the different climates regardingwomen’s rights in Japan and the United States (Lebra,1993). Japan is a more traditional culture in terms ofgender roles, and women’s equality is only currentlygaining momentum. As an example, gender role ste-reotypes are quite prevalent in Japanese children’stelevision programming, and this has only recentlybeen viewed as an issue to be discussed at the na-tional level (Rolandelli, 1991). This could account forwhy American girls were more sensitive to issues ofexclusion than were Japanese girls, but it does not ex-plain why American boys were the least sensitive tothese issues. The item on which American boys weremost likely to condone exclusion was that of the childwho acts like a girl (Cross-Gender context). Researchon gender-segregated play (Ruble & Martin, 1998)has shown that there is a very strong taboo amongboys in the United States with regard to cross-genderbehavior, and this may lead American boys to viewsuch behavior as a legitimate basis for exclusion.However, American boys were also more willing thanwere the other groups to exclude someone for being aslow runner. Perhaps this latter ﬁnding reﬂects thesocietal pressure that American boys experience inthe area of sports, as evidenced by media reports onthe increase in parental pressure for boys to excel insports as early as 6 and 7 years of age. The pressure toexcel may have its price in the manifestation of a lackof an inclusive orientation by boys in peer groupsports contexts. More research is needed to under-stand these gender ﬁndings.

We also found that boys expected more confor-mity from the atypical peers than did girls. These re-sults accord with previous investigations that reportedlower levels of conformity and greater intrinsic moti-vation among girls than boys (Boehnke, Silbereisen,Eisenberg, & Reykowski, 19; Werrbach, Grotevant,& Cooper, 1990). Our prediction regarding genderdifferences in conformity ratings was based on the as-sumption that girls would themselves have experi-enced more exclusion than would have boys andwould, therefore, identify with the excluded children.

1798Child Development

We did not, however, ﬁnd that girls rated themselvesas more similar to the atypical peer than did boys(with the exception of two categories—Sports andCross-Gender behavior). We did ﬁnd, however, thatboys were more likely to be intolerant than were girls.Expecting conformity and being intolerant of dif-ferences are highly related constructs. Thus, furtherresearch is needed to provide a comprehensive expla-nation for the gender differences that were found.Contrary to predictions, there were no overall agedifferences for exclusion judgments. Age-related ﬁnd-ings on ratings of exclusion emerged for only twocontexts pertaining to group functioning. With age,children were more willing to exclude an aggressivechild and a slow runner. The Sports ﬁnding is consis-tent with prior research, which has shown age-relatedresults for the willingness to exclude an unqualiﬁedplayer from a team sport (e.g., basketball; see Killen &Stangor, 2001). The ﬁnding for the Aggressive contextsupports prior research showing that, with age, ag-gressive children are increasingly rejected by otherchildren (Rubin et al., 1998).

Developmental hypotheses also received partialsupport in the analyses on Tolerance of differences.As predicted, with age, students were less likely toendorse the tolerant response style overall, and, spe-ciﬁcally, in the Hair, Sad, and Sports scenarios. Stu-dents who gave responses that were coded in theTolerant category were those who disagreed with ex-clusion, yet perceived themselves to be different fromthe atypical child. Our hypotheses regarding this cat-egory derived from the notion that, with age, stu-dents would become more sensitive to the impor-tance of group functioning (Killen & Stangor, 2001),and would thus consider it as increasingly legitimateto exclude someone who might disturb such func-tioning, in this case, someone who was different.However, as discussed above, exclusion ratings, ingeneral, did not change with age. It follows, therefore,that the developmental decrease in endorsement ofthe tolerance of differences perspective was due notto students’ greater willingness to exclude, but ratherto an increased identiﬁcation with the atypical peers.It was the older students’ perception of greater simi-larity to the excluded child, particularly in those areasof self that become especially important during theadolescent period—appearance, emotional stability,and athletic competence (Harter, 1990)—that pre-vented them from being categorized as “tolerant ofdifferences.” The greater degree of perceived similar-ity to the atypical peer (for three of the six contexts)also explains why we found no corresponding devel-opmental increase in endorsement of the intolerantperspective. This is because perceived difference is acriterion for an intolerant viewpoint. The only gradeeffect for the Intolerance category was for the aggres-sive child. With age, children were more intolerant ofsomeone who is aggressive.

Developmental theory suggests that the need forautonomy and self-individuation increases duringthe period of identity formation in early and middleadolescence (Nucci, 1996; Smetana, 1988). In supportof this theory, we found a decrease with age in chil-dren’s expectations that the atypical peer alter his orher behavior for group acceptance, among bothAmerican and Japanese students. In ﬁve of the six sce-narios, older Japanese students expected less confor-mity than did younger Japanese students. The onlyexception was found with regard to the aggressivechild, in which there were no signiﬁcant age-relatedchanges in responses by Japanese students. The Fightscenario described someone who engaged in disrup-tive behavior that might disturb group functioning,and also might bring both the individual and thegroup into conﬂict with adult authorities (e.g., the se-curity police at the mall). In light of long-standingcultural sanctions in Japan against direct interper-sonal confrontation (Peak, 19), it is reasonable thatpressures to avoid such confrontation would not di-minish, but rather increase with age. Japanese culturehas been described as emphasizing the importance ofhierarchy and respect for authority that are tied toseniority and age (Roland, 1988; Tinsley, 1998). Thenegative consequences for violating these values mayincrease as children get older. Nevertheless, fromthe perspective of developmental and cross-culturaltheory, it is interesting that the developmental pro-cesses governing such group dynamics appear to begenerally similar in both societies, despite the greatervalue placed on behavioral conformity by Japanese,relative to American, students. These results, in con-junction with the ﬁndings on exclusion and the sa-lience of group functioning for adolescents in somecontexts, support our contention that concerns withindividual autonomy and group functioning coexistin adolescent decision making in American and Japa-nese cultures.

The tolerance of differences measure provided in-formation about how willing children were to includesomeone who was different from themselves. Socialcomparison theory predicts that, under certain cir-cumstances, individuals prefer to compare them-selves to others who are similar to themselves (Ruble& Frey, 1991). Most of the research on social compari-son has been conducted with Western samples (Butler& Ruzany, 1993, and Crystal et al., 2000 are excep-tions), and has focused on the academic domain. Thepresent study examined the extent to which children

were willing to exclude from a peer group others whowere different from the self. We found that childrenwere more tolerant of differences than they were in-tolerant (on the whole) for all contexts except that ofthe aggressive child. Thus, we found that childrenwere not willing to exclude someone from a group,even when they perceived that child to be very differ-ent from themselves. These results provide a view ofthe possibilities of positive peer interactions (seeRubin et al., 1998), and, in addition, reveal that toler-ance of differences depends on the issue being dis-cussed, as shown by Wainryb et al. (1998). This is be-cause age-related increases were shown for only half,not all, of the six contexts.

It is important to acknowledge the limitations ofthe present research methodology. No single measureis without limitations and in this case we used thesurvey method, which did not allow for in-depthprobing of participants’ reasoning for their evalua-tions. A follow-up study could be conducted to obtainmore extensive data regarding how children and ad-olescents reason about exclusion in different contexts.The use of the survey method, however, enabled us tocollect a fairly large sample of students (513 in Japan,544 in the United States) with which we were able totest within-culture differences that included age andgender as relevant variables in addition to examiningbetween-culture differences. Thus the survey methodenabled us to test our hypotheses about culture de-spite the drawbacks of not being able to conduct indi-vidual in-depth interviews with each student.

In future research, interviewing participants wouldenable probing of children’s justiﬁcations for their de-cision making about exclusion, as mentioned above,which would provide more information about theirinterpretations of the varied contexts of exclusion.Further, observational data of actual instances of ex-clusionary behavior could also serve as an importantcomplement to the interview data. Although docu-menting behavioral correlates of exclusion is difﬁcult(exclusion becomes more implicit and psychologicalwith development, thus hindering attempts to wit-ness or videotape it), setting up exclusionary situa-tions in the laboratory could reveal the social interac-tional processes that occur when one individualexcludes another. This research could also be extendedto other contexts, such as the family and school. Thepresent study focused on exclusion from peer groupactivities. Examining judgments about exclusion in thehome (e.g., fathers excluding a daughter from an ac-tivity) and exclusion in the school setting (e.g.,teachers excluding particular children from specialopportunities) would enhance the knowledge abouthow children evaluate exclusion in multiple contexts.

Killen, Crystal, and Watanabe1799

ACKNOWLEDGMENTS

Part of this project was presented as a poster at the In-ternational Society for the Study of Behavioral Devel-opment, Beijing, China, 2000, and at the Annual Sym-posium of the Jean Piaget Society, Berkeley, CA, June2001. The authors thank the following students fortheir assistance with data collection in the UnitedStates: Stacey Horn, Jennie Lee-Kim, Heidi McGloth-lin, B. Sefton Price, and Micah Stretch; and with dataanalyses for the entire dataset: Stacey Horn, YoonjungPark, and Micah Stretch. They also thank ChristinaEdmonds for her help on the manuscript prepara-tion. This project was supported, in part, by a grantfrom the National Science Foundation (BCS9729739)awarded to the ﬁrst author. The authors extend theirappreciation to the Montgomery County PublicSchools in Maryland, U.S.A., for their willingness toallow this study to be conducted in their school sys-tem, as well as to the principals and teachers who par-ticipated. They are grateful to the parents who gavetheir consent, and to the students in the United Statesand in Japan for their willingness to participate in thisproject.

ADDRESSES AND AFFILIATIONS

Corresponding author: Melanie Killen, Departmentof Human Development, 3304 Benjamin Building,University of Maryland, College Park, MD 20742-1131; e-mail: mkillen@umd.edu. David S. Crystal isat Georgetown University, Washington, DC; andHirozumi Watanabe is at Ehime University, Mat-suyama, Ehime, Japan. APPENDIX

TYPES OF EXCLUSION SCENARIOS

(Note: Japanese and American names, matched for the gen-der of the participant, were used in each country for theexcluded child in the scenario.)

1.Fight: There is a group of ﬁve boys who are goodfriends. One of the boys, named “A,” sometimes gets intophysical ﬁghts with other boys. One day, the group decides togo shopping at the mall. All the other boys in the group de-cide not to ask “A” to come along because they are afraid hewill hurt someone.

2.Hair: There is another group of ﬁve boys who aregood friends. One of the boys, named “B,” wears unusualclothes and dyes his hair green. One day, the group decidesto go to a fancy restaurant. All the other boys in the groupdecide not to ask “B” to come along because of his unusualclothes and green hair.

3.Clown: One of the boys, named “C,” often gets intotrouble for acting like a clown in class. One day, the group

1800Child Development

decides to go to the movies. All the other boys in the groupdecide not to ask “C” to come along because he might startacting like a clown in the movie theater.

4.Gender: One of the boys, named “D,” sometimes talksand acts like a girl. One day, the group decides to join aboys’ club. All the other boys in the group decide not to ask“D” to join the club because he acts like a girl. (Reversed for“girl” scenario.)

5.Sports: One of the boys, named “E,” is not a goodathlete and is a slow runner. One day, the group decides tocompete in a town race. The winners of the race get freeathletic equipment as a prize. All the other boys in thegroup decide not to ask “E” to join the team because he isa slow runner.

6.Sad:One of the boys, named “F,” often looks sad andrarely laughs at people’s jokes. One day, the group decidesto go on a picnic with a group of girls. All the other boys inthe group decide not to ask “F” to come along because helooks sad and rarely laughs.

REFERENCES

Adorno, T., Frenkel-Brunswick, E., Levinson, D., & San-ford, N. (1950). The authoritarian personality. New York:Harper.

Berndt, T. (1992). Friendship and friends’ inﬂuence in ado-lescence. Current Directions in Psychological Science, 1,156–159.

Boehnke, K., Silbereisen, R. K., Eisenberg, N., & Reykowski,J. (19). Developmental patterns of prosocial motiva-tion: A cross-national study. Journal of Cross-Cultural Psy-chology, 20, 219–243.

Boivin, M., Dodge, K., & Coie, J. D. (1995). Individualand group behavioral similarity and peer status in ex-perimental playgroups of boys: The social misﬁt revis-ited. Journal of Personality and Social Psychology, 69,269–279.

Brown, B. B., Eicher, S. A., & Petrie, S. (1986). The impor-tance of peer group (“crowd”) afﬁliation in adolescence.Journal of Adolescence, 9, 73–96.

Butler, R., & Ruzany, N. (1993). Age and socialization effectson the development of social comparison motives andnormative ability assessment in kibbutz and urban chil-dren. Child Development, , 532–543.

Crystal, D. S., Watanabe, H., & Chen, R. (2000). Reactions tomorphological deviance: A comparison of Japanese andAmerican children and adolescents. Social Development,9, 40–61.

Crystal, D. S., Watanabe, H., & Chin, W. (1997). Intoleranceof human differences: A developmental and cross-cultural study of American, Japanese, and Chinese chil-dren. Journal of Applied Developmental Psychology, 18, 149–167.

Crystal, D. S., Watanabe, H., Weinfurt, K., & Wu, C. (1998).Concepts of human differences: A comparison of Amer-ican, Japanese, and Chinese children and adolescents.Developmental Psychology, , 217–222.

Erikson, E. (1968). Identity: Youth and crisis. New York: Norton.

Gavin, L. A., & Furman, W. (19). Age difference in adoles-cents’ perceptions of their peer groups. DevelopmentalPsychology, 25, 827–834.

Greenﬁeld, P., & Cocking, R. (1994). Cross-cultural roots ofminority child development. Mahwah, NJ: Erlbaum.

Gudykunst, W., Yoon, Y., & Tsukasa, N. (1987). The inﬂu-ence of individualism–collectivism on perceptions ofcommunication in ingroup and outgroup relationships.Communication Monographs, 54, 295–306.

Hamilton, V. L., Blumenfeld, P., Akoh, H., & Miura, K.(1991). Group and gender in Japanese and Americanelementary classrooms. Journal of Cross-Cultural Psychol-ogy, 22, 317–346.

Harkness, S., Super, C., & van Tijen, N. (2000). Individual-ism and the “Western Mind” reconsidered: Americanand Dutch parents’ ethnotheories of the child. In S.Harkness, C. Raeff, & C. M. Super (Eds), New directionsfor child development: No. 87. Variability in the social con-struction of the child (pp. 23–40). San Francisco: Jossey-Bass.

Harter, S. (1998). The development of self representations.In N. Eisenberg (Ed.), W. Damon (Series Ed.), Handbook ofchild psychology: Vol. 3. Socialization (5th ed., pp. 553–617).New York: Wiley.

Harter, S. (1990). Causes, correlates, and the functionalrole of global self worth: A life span perspective. InR. J. Sternberg & J. J. Kolligian (Eds.), Competence con-sidered (pp. 67–97). New Haven, CT: Yale UniversityPress.

Helwig, C. (1995). Social context in social cognition: Psycho-logical harm and civil liberties. In M. Killen & D. Hart(Eds.), Morality in everyday life: Developmental perspectives(pp. 166–200). Cambridge, U.K.: Cambridge UniversityPress.

Helwig, C. C., Arnold, M. L., Tan, D., & Boyd, D. (2001,April). Adolescents’ judgements of democratic and authority-based decision making in the People’s Republic of China. Paperpresented at the biennial meeting of the Society for Re-search in Child Development, Minneapolis, MN.

Horn, S. (in press). Adolescents’ social reasoning about ex-clusion from social groups. Developmental Psychology.Killen, M. (1991). Social and moral development in earlychildhood. In W. M. Kurtines & J. L. Gewirtz (Eds.),Handbook of moral behavior and development (Vol. 2, pp.115–138). Hillsdale, NJ: Erlbaum.

Killen, M., McGlothlin, H., & Lee-Kim, J. (2002). Betweenindividuals and culture: Individuals’ evaluations of ex-clusion from social groups. In H. Keller, Y. Poortinga, &A. Schoelmerich (Eds.), Between biology and culture: Per-spectives on ontogenetic development (pp. 159–190). Cam-bridge, U.K.: Cambridge University Press.

Killen, M., Pisacane, K., Lee-Kim, J., & Ardila-Rey, A. (2001).Fairness or stereotypes?: Young children’s prioritieswhen evaluating group exclusion and inclusion. Devel-opmental Psychology, 37, 587–596.

Killen, M., & Stangor, C. (2001). Children’s reasoning aboutsocial inclusion and exclusion in gender and race peergroup contexts. Child Development, 72, 174–186.

Killen, M., & Wainryb, C. (2000). Independence and interde-

pendence in diverse cultural contexts. In S. Harkness &C. Raeff (Eds.), New directions for child development: No.87. Individualism and collectivism as cultural contexts for de-velopment (pp. 5–21). San Francisco: Jossey-Bass.

Lebra, T. S. (1993). Culture, self, and communication in Japanand the United States. In W. B. Gudykunst (Ed.), Commu-nication in Japan and the United States (pp. 51–87). Albany,NY: State University of New York Press.

Lewis, C. (1995). Educating hearts and minds. Cambridge,U.K.: Cambridge University Press.

Matsumoto, D. (1990). Cultural similarities and differencesin display rules. Motivation and Emotion, 14, 195–214.Markus, H., & Kitayama, S. (1991). Culture and the self: Im-plications for cognition, emotion, and motivation. Psy-chological Review, 98, 224–253.

Mukai, T. (1996). Body dissatisfaction, depressive affect,and eating problems among young adolescent girls: Aprospective study. Japanese Journal of Counseling Science,29, 37–43.

Nucci, L. P. (1996). Morality and the personal sphere of ac-tions. In E. Reed, E. Turiel, & T. Brown (Eds.), Values andknowledge. Mahwah, NJ: Erlbaum.

O’Brien, S. F., & Bierman, K. L. (1988). Conceptions and per-ceived inﬂuence of peer groups: Interviews with pread-olescents and adolescents. Child Development, 59, 1360–1365.

Opotow, S. (1990). Moral exclusion and injustice: An intro-duction. Journal of Social Issues, 46, 1–20.

Osbeck, L. M., Moghaddam, F., & Perreault, S. (1997). Simi-larity and attraction among majority and minoritygroups in a multicultural context. International Journal ofIntercultural Relations, 21, 113–123.

Peak, L. (19). Learning to become part of the group: TheJapanese child’s transition to preschool life. Journal ofJapanese Studies, 1, 93–123.

Poulin, F., Cillessen, A. H. N., Hubbard, J. A., & Coie, J. D.(1997). Children’s friends and behavioral similarity intwo social contexts. Social Development, 6, 224–236.

Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. F.(1994). Social dominance orientation: A personality vari-able predicting social and political attitudes. Journal ofPersonality and Social Psychology, 67, 741–763.

Roland, A. (1988). In search of self in India and Japan: Toward across-cultural psychology. Princeton, NJ: Princeton Uni-versity Press.

Rolandelli, D. R. (1991). Gender role portrayal analysis ofchildren’s television programming in Japan. Human Re-lations, 44, 1273–1299.

Rosenberg, N. (Ed.). (1992). Japanese sense of self. Cambridge,U.K.: Cambridge University Press.

Rubin, K. H., Bukowski, W., & Parker, J. (1998). Peer inter-actions, relationships, and groups. In N. Eisenberg (Ed.),W. Damon (Series Ed.), Handbook of child psychology:Vol. 3. Socialization (5th ed., pp. 619–700). New York:Wiley.

Rubin, K. H., Lynch, D., Coplan, R., Rose-Krasnor, L., &Booth, C. L. (1994). “Birds of a feather . . .”: Behavioralconcordances and preferential personal attraction inchildren. Child Development, 65, 1778–1785.Killen, Crystal, and Watanabe1801

Ruble, D. N., & Frey, K. S. (1991). Changing patterns of com-parative behavior as skills are acquired: A functionalmodel of self-evaluation. In J. Suls & T. A. Wills (Eds.),Social comparison: Contemporary theory and research (pp.79–113). Hillsdale, NJ: Erlbaum.

Ruble, D. N., & Martin, C. L. (1998). Gender development.In N. Eisenberg (Ed.), W. Damon (Series Ed.), Handbook ofchild psychology: Vol. 3. Socialization (pp. 933–1016). NewYork: Wiley.

Shimizu, M. (1990). Seinenki to gendai [Adolescents and thepresent day]. Tokyo: Kumondo.

Smetana, J. G. (1988). Concepts of self and convention: Ad-olescents’ and parents’ reasoning about hypothetical andactual family conﬂicts. In M. R. Gunnar & W. A. Collins(Eds.), Twenty-ﬁrst Minnesota Symposium on Child Psy-chology: Development during the transition to adolescence.Hillsdale, NJ: Erlbaum.

Smetana, J. G. (1995). Morality in context: Abstractions, ambi-guities, and applications. In R. Vasta (Ed.), Annals of child de-velopment (Vol. 10, pp. 83–130). London: Jessica Kingsley.Spiro, M. (1993). Is the Western conception of the self “pe-culiar” within the context of the world cultures? Ethos,21, 107–153.

Stretch, M. (2000). The effects of gender and group compositionon children’s evaluations of exclusion. Unpublished mas-ter’s thesis, University of Maryland, College Park.

Takeuchi, K., Roberts, R. E., & Suzuki, S. (1995). Depressivesymptoms among Japanese and American adolescents.Psychiatry Research, 53, 259–274.

Theimer, C. E., Killen, M., & Stangor, C. (2001). Young chil-dren’s evaluations of exclusion in gender-stereotypicpeer contexts. Developmental Psychology, 37, 18–27.

Tinsley, C. (1998). Models of conﬂict resolution in Japanese,German, and American cultures. Journal of Applied Psy-chology, 83, 316–323.

Tisak, M. (1995). Domains of social reasoning and beyond.In R. Vasta (Ed.), Annals of child development (Vol. 11, pp.95–130). London: Jessica Kingsley.

Triandis, H. (19). The self and social behavior in differingcultural contexts. Psychological Review, 93, 506–520.

Triandis, H. (1995). Individualism and collectivism. Boulder,CO: Westview Press.

Turiel, E. (1983). The development of social knowledge: Moralityand convention. Cambridge, U.K.: Cambridge UniversityPress.

Turiel, E. (1998). The development of morality. In N. Eisen-berg (Ed.), W. Damon (Series Ed.), Handbook of child psy-chology: Vol. 3. Socialization (5th ed., pp. 863–932). NewYork: Wiley.

Turiel, E., & Wainryb, C. (1994). Social reasoning and thevarieties of social experience in cultural contexts. InH. W. Reese (Ed.), Advances in child development and be-havior. (Vol. 25, pp. 2–326). Orlando, FL: AcademicPress.

Wainryb, C., Shaw, L., Laupa, M., & Smith, K. (2001). Chil-dren’s, adolescents’ and young adults’ thinking aboutdifferent types of agreements. Developmental Psychology,37, 373–386.

Wainryb, C., Shaw, L., & Maianu, C. (1998). Tolerance and

1802Child Development

intolerance: Children’s and adolescents’ judgments ofdissenting beliefs, speech, persons, and conduct. ChildDevelopment, 69, 1541–1555.

Wainryb, C., & Turiel, E. (1994). Dominance, subordination,and concepts of personal entitlement in cultural con-texts. Child Development, 65, 1701–1722.

Werrbach, G. B., Grotevant, H. D., & Cooper, C. R. (1990).Gender differences in adolescents’ identity develop-ment in the domain of sex role concepts. Sex Roles, 23,521–530.

Yau, J., & Smetana, J. G. (1993). Chinese-adolescents’ rea-soning about cultural conﬂicts. Journal of Adolescent Re-search, 8, 419–438.

Yau, J., & Smetana, J. G. (1996). Adolescent–parent conﬂictamong Chinese adolescents in Hong-Kong. Child Devel-opment, 67, 1262–1275.

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文