language and intergroup contact: investigating the impact

HAL Id: hal-00571610https://hal.archives-ouvertes.fr/hal-00571610

Submitted on 1 Mar 2011

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Language and Intergroup Contact: Investigating theImpact of Bilingual Instruction on Children’s Intergroup

AttitudesStephen C. Wright, Linda R. Tropp

To cite this version:Stephen C. Wright, Linda R. Tropp. Language and Intergroup Contact: Investigating the Impact ofBilingual Instruction on Children’s Intergroup Attitudes. Group Processes and Intergroup Relations,SAGE Publications, 2005, 8 (3), pp.309-328. �10.1177/1368430205053945�. �hal-00571610�

https://hal.archives-ouvertes.fr/hal-00571610

https://hal.archives-ouvertes.fr

GPIR

Language and IntergroupContact: Investigating theImpact of BilingualInstruction on Children’sIntergroup Attitudes

Stephen C. Wright Simon Fraser University

Linda R. TroppBoston College

This study examined the impact of bilingual versus English-only instruction on the intergroupattitudes of White, English-speaking children in kindergarten through second grade.Replicating prior research, White children generally showed a clear preference toward theingroup in terms of positive evaluations, friendship preference, and perceived similarity to theself. However, all three effects were significantly smaller among children who were inclassrooms with a significant amount of Spanish instruction (i.e. bilingual classes). The smallerpreference for the ingroup over the outgroup found in bilingual classes resulted from higherevaluations of, greater selection of friends among, and greater perceived similarity to Latinotargets, and not from changes in preference for White ingroup targets. Furthermore,comparisons with English-only classes that had substantial Latino representation shows thatthe positive impact of bilingual instruction can be only partially explained by the greaterrepresentation of Latino children in bilingual classes. Finally, these positive patterns ofintergroup attitudes found in bilingual classes were not associated with any negative effects onWhite children’s personal self-evaluation.

keywords bilingual education, children, intergroup attitudes, intergroup contact,prejudice

Group Processes &Intergroup Relations

2005 Vol 8(3) 309–328

BILINGUAL education is under vigorous attack.In the state of California, a popular initiative‘outlawed’ most forms of bilingual educationand mandated that all non-English speakingchildren be quickly ‘mainstreamed’ into stan-dard English-only classes. Similar initiatives are

Copyright © 2005 SAGE Publications(London, Thousand Oaks, CA and New Delhi)8:3; 309–328; DOI: 10.1177/1368430205053945

Author’s noteAddress correspondence to Stephen C. Wright,Psychology Department, Simon FraserUniversity, Burnaby, BC, V5A 1S6, Canada[email: [email protected]]

08 Wright (b-ct) 27/6/05 1:54 pm Page 309

being considered or have been passed in Ari-zona, Massachusetts, and several other states,and new federal policies restrict the amount ofheritage language education children canreceive (Slavin & Cheung, 2003). These initia-tives come at a time when more than 20% of allUS students come from homes where English isnot the primary language (Van Hook & Fix,2000). Thus, just as American public schoolsare facing unprecedented linguistic and cul-tural diversity, public support for instruction inlanguages other than English appears to be inretreat.

The public debate on bilingual education hasparalleled an equally contentious academicdebate (see Hakuta, Butler, & Witt, 2000). Theargument is complicated by the wide range ofmodels of ‘bilingual education’ and the multi-tude of specific programs based on each model.However, despite critiques, most research-basedreviews conclude that the important benefits ofbilingual education outweigh the potentialcosts (see Crawford, 1989; Cummins, 1989;Genesee, 1987; Greene, 1997; Hakuta, 1986;Slavin & Cheung, 2003; Willig, 1985; Wong-Fillmore & Valdez, 1986). Indeed, in terms ofthe experiences of non-English speaking chil-dren, the data suggest that current efforts toreduce minority-language instruction are un-fortunate and misguided.

At the same time, since bilingual educationhas been viewed primarily as a compensatorymeasure for educating non-English speakers(see Brisk, 1998), most discussions have exam-ined its impact on the academic achievementand language acquisition of minority languagechildren (e.g. Edwards, 1977; Genesee, Rogers,& Holobow, 1983; Hakuta et al., 2000; Parker,1978; Slavin & Cheung, 2003; Sue & Padilla,1986; Wright, Taylor, & Macarthur, 2000). Bycontrast, the social outcomes of bilingualinstruction have been sorely understudied (seeBarker et al., 2001; Genesee & Gándara, 1999;Lambert & Cazabon, 1994; Noels, Pon, &Clément, 1996; Wright & Taylor, 1995), andlittle is known regarding how bilingual instruc-tion influences children’s attitudes towarddifferent language groups (see Aboud & Levy,2000). But as research in the tradition of this

special issue has shown, the connectionsbetween language and intergroup relations arenumerous and important. Language representsmuch more than a medium of communication,serving also as a key marker of social identityand as a determinant of intergroup behavior.Thus, more research is needed to understandhow participation in bilingual instruction mayaffect intergroup attitudes toward members ofother language groups.

Moreover, since bilingual education has beenregarded as an issue that is mostly relevant to theeducation of minority language students, verylittle attention has been paid to its impact onmajority language children (see Lambert,Genesee, Holobow, & Chartrand, 1993; Lambert& Tucker, 1972 for notable exceptions). Thepresent research focuses directly on these twounderstudied areas by considering the effects ofSpanish/English bilingual instruction versusEnglish-only instruction on the intergroupattitudes of White, English-speaking children.1

We examine these issues in the context ofchildren’s earliest school experiences—kinder-garten through Grade 2. Obviously, children atthese grade levels have had the least exposureto prior classroom contexts that could influ-ence their responses to the present classroomsetting. These are also the ages at which chil-dren tend to become aware of ethnic differ-ences (Aboud, 1988; Katz, 1983), and begin toform enduring perceptions of, and attitudestoward, members of different ethnic groups(Foster, 1994; Ramsey, 1987). Moreover, itappears that young children are well aware oflinguistic differences and their importance indetermining group membership (see Aboud,1977). Thus, it is surprising that the role of lan-guage as a marker of group membership hasremained understudied in the research litera-ture on children’s intergroup attitudes.

Generally, research on the intergroup atti-tudes of young children suggests that Whitechildren typically express positive attitudestoward their ethnic ingroup (Weiland &Coughlin, 1979; Williams & Morland, 1976),while rarely showing equally positive attitudestoward other racial or ethnic groups (Doyle &Aboud, 1995; Katz, 1983). In some instances,

Group Processes & Intergroup Relations 8(3)

310

08 Wright (b-ct) 27/6/05 1:54 pm 10

White children may even demonstrate a clearrejection of children from racial or ethnic out-groups, while maintaining positive attitudestoward members of the White ingroup (Asher &Allen, 1969; Corenblum & Wilson, 1982). Theseinitial intergroup perceptions can have anenduring impact on children’s future attitudes,and it may become increasingly difficult to influ-ence these attitudes as children grow older(Banks, 1995). Thus, children who are just form-ing intergroup attitudes may be most susceptibleto influences of the educational environment,such as the language of instruction.

Considerable evidence shows that organiz-ational structures and practices in school cansignificantly influence children’s intergroupattitudes (Damico, Bell-Nathaniel, & Green,1981; Khmelkov & Hallinan, 1999). Still, thereis virtually no published research investigatinghow bilingual instruction affects children’sintergroup attitudes.2 A possible explanationfor the lack of attention to language as a vari-able affecting intergroup attitudes might be thetraditional focus on racial differences, such asthe extensive body of research regarding theeffects of school desegregation on relationsbetween Black and White children (seeSchofield, 1989; Stephan & Rosenfield, 1978).However, when one considers relations betweenWhite and Latino (or other immigrant) chil-dren, language becomes another key marker ofgroup membership (Gudykunst & Schmidt,1987; Heller, 1987).

Research on intergroup contact andthe potential role of language

Like much of the research on intergroup atti-tudes in schools, our analysis of the impact ofbilingual instruction is guided by the ContactHypothesis. Allport’s (1954) formulation of thehypothesis proposes that contact between mem-bers of different groups can, under a specifiedset of conditions, lead to improved attitudestoward the outgroup as a whole. Much of theliterature on intergroup contact in schools hasprovided support for the hypothesis and hashighlighted the importance of Allport’s ‘optimalconditions’ for enhancing positive intergroup

attitudes (see Miller & Brewer, 1984; Pettigrew,1998; Pettigrew & Tropp, 2005; Schofield, 1995for reviews). Specifically, Allport proposed thatcontact should be structured so that: (a) mem-bers of the two groups hold equal status duringthe interaction; (b) it promotes cooperativeinterdependence between the members of thetwo groups; and (c) relevant authorities supportthe intergroup contact. We propose that thelanguage of instruction can significantly con-tribute to establishing these important con-ditions within the classroom context.

Equal statusGiven that language is a primary marker ofgroup membership, the perceived status ofeach group’s language will be inherently linkedto the group’s broader status (see Giles,Bourhis, & Taylor, 1977). If English is the solelanguage of instruction, children will likelygrow to believe that acquiring the valuableknowledge and skills imparted in schoolrequires English proficiency. This clearly placesthe English language and those who speak it ina higher status position than other languagesand those who speak them. The perceivedstatus of English speakers may be enhanced fur-ther if English speakers fill most of the school’shigh-status positions (e.g. teachers, principals).English-only instruction can also give English-speaking students greater access to educationalmaterials and to the teacher, from whom theycan make more sophisticated requests, andreceive more complex explanations. Thus,being an English speaker becomes associatedwith success and status, while speaking anotherlanguage is not. In sum, English-only instruc-tion suggests, either directly or subtly, thatEnglish speakers hold superior status within theclassroom relative to members of other lan-guage groups.

In contrast, bilingual instruction can providea clear affirmation of the value and status ofthe relevant minority language (Cummins,1989; Lambert & Cazabon, 1994). Bilingualinstruction also establishes a direct connectionbetween the teacher and both language groups,while equalizing their access to classroomresources (Parker, 1978).

Wright & Tropp language and intergroup contact effects

311


Cooperative interdependence Cooperative interdependence has also beenproposed as an important condition for success-ful intergroup contact, and considerableresearch supports this claim (Aronson &Patnoe, 1997; Brewer & Miller, 1984; Johnson,Johnson, & Maruyama, 1984; Kagan, 1986;Schofield, 1989; Slavin & Cooper, 1999). Thereis some evidence that the specific demands ofbilingual instruction may lead teachers to make greater use of cooperative activities,even when the content of the lesson is notlanguage (Lindholm, 1994). The degree thatcooperative strategies are used in an English-only classroom, English speakers are likely toassume the role of tutors and non-Englishspeakers are likely to become the recipients oftheir help. This dynamic not only highlightsstatus inequalities between the groups, but alsoundermines the interdependent nature ofcooperation that is essential for positive inter-group contact.

Authority supportResearch has shown that support from authori-ties can make intergroup contact more accept-able, more frequent, and more effective atimproving intergroup attitudes (see Pettigrew,1998). In the classroom, the relevant authorityis of course the teacher. The teacher’s use ofboth languages is a direct and clear statementof his/her support for a multilingual context.Thus, instructional use of both languages maybe the clearest possible statement that theauthority is sanctioning positive, equal-status,cross-group interactions.

New developments in contactresearch: Implications for bilingual instruction

While there has been considerable support forAllport’s proposed conditions (see Pettigrew,1998), a recent resurgence of interest in the con-tact hypothesis has sparked a number of otheradvances (Brewer & Brown, 1998; Gaertner &Dovidio, 2000; Hewstone & Brown, 1986; Miller& Brewer, 1984; Pettigrew, 1998; Pettigrew &Tropp, 2000). Among the new advances are two

themes that appear particularly relevant to thepresent discussion.

Cross-group friendships The first involves the kinds of contact experi-ences most likely to produce positive inter-group outcomes. Here, recent work has pointedto the particular importance of cross-groupfriendships for promoting positive changes inintergroup attitudes (Aboud, Mendelson, &Purdy, 2003; McLaughlin-Volpe, Aron, Wright,& Reis, 2000; Pettigrew, 1997; Wright, Aron,McLaughlin-Volpe, & Ropp, 1997; Wright,Aron, & Tropp, 2002; Wright, Brody, & Aron,2005; Wright & Van der Zande, 1999). We pro-pose that the classroom environment createdby bilingual instruction may be more likelythan English-only instruction to facilitate cross-group friendships. Instruction in a secondlanguage may lead children to perceive greatersimilarity between themselves and members ofthe other language group (Genesee et al.,1983). There is ample evidence to suggest thatperceptions of similarity are associated withgreater propensities to form friendships (seeFehr, 1996). Thus, by learning a secondlanguage, White, English-speaking childrenmay perceive greater similarity to their non-English-speaking classmates, and this greaterperceived similarity should increase children’swillingness to develop friendships acrossethnolinguistic boundaries (see Lambert &Cazabon, 1994).

Generalization of contact effectsA second theme concerns how contact with anindividual outgroup member generalizes toattitudes about the outgroup as a whole. Basedon the social identity approach (Tajfel, 1978;Turner, Hogg, Oakes, Reicher, & Wetherell,1987), Hewstone & Brown (1986) have arguedthat group memberships must be salient duringcontact for positive feelings toward individualoutgroup members to generalize to the entireoutgroup. There is now a growing consensus inthe literature that, at some point, group mem-berships must become salient for interactionsbetween individual group members to affectintergroup attitudes (see also Brewer & Brown,


312


1998; Brown & Hewstone, 2005; Pettigrew,1998; Rothbart & John, 1985). Bilingualinstruction may be able to facilitate positiveinteractions across group lines and increasefeelings of interpersonal similarity while at thesame time maintaining awareness of theethnolinguistic distinctiveness of two languagegroups.

The possible confound of classroomethnic composition

The numerical representation of White andnon-White children in a classroom may alsohave a profound impact on the development ofintergroup attitudes. Indeed, other worksuggests that White children in more ethnicallybalanced classrooms tend to be more positive intheir assessment of non-White children thanthose in White-only or predominantly Whiteclasses (see Aboud, 1988; Cohen, 1980; Kistner,Metzler, Gatlin, & Risi, 1993; Schofield & Sagar,1977).

Typically, bilingual classes have more equalrepresentations of White and non-White chil-dren than English-only classes. Thus, there isthe possibility of confounding language ofinstruction with ethnic mix in the classroom. Toaddress this issue, the present research includestwo groups of English-only classrooms: thosewith balanced representations of White andLatino children and those with token or norepresentation of Latino children. Thus, we canconsider the impact of interethnic contactalone versus contact supported by the use ofthe minority language.

Language of instruction and self-evaluation

Finally, we must also consider the potentialeffects of bilingual instruction on children’spsychological well-being. The limited researchon the impact of bilingual instruction on major-ity language children has shown no ill effectsfor these children (Lambert & Cazabon, 1994;Lambert et al., 1993). However, some publicopposition to bilingual education has focusedon possible negative outcomes for White

children, and specifically, on the possibility ofpoorer psychological adjustment. One import-ant and well-studied aspect of adjustment in theschool context concerns children’s evaluationsof themselves (Covington, 1989; Harter, 1986).The procedures used in the present studyinclude a measure of children’s self-evaluations,allowing for an investigation of the effects oflanguage of instruction and classroom repre-sentation of Latino children on White chil-dren’s views of themselves.

Summary of research hypotheses

Past research has shown that White children aregenerally more positive in their perceptions ofingroup targets than minority outgroup targets(e.g. Doyle & Aboud, 1995). Thus, we hypothe-sized that White children would show generalingroup preference on measures of intergroupattitudes and perceived similarity. Specifically,we propose that White children will show ingroupbias by selecting more White than Latino targets inresponse to measures of (a) positive evaluations, (b)friendship preferences, and (c) perceived similarity(Hypothesis 1).

However, the extensive literature on inter-group contact would predict that White chil-dren in classrooms with a significant percentageof Latino students should hold more positiveattitudes about Latinos. Therefore, we predictthat White children in ethnically mixed classes (bothbilingual and English-only) will show less bias infavor of White over Latino targets on measures of (a)positive evaluations, (b) friendship preferences, and(c) perceived similarity, relative to White children inEnglish-only classes with few or no Latino children(Hypothesis 2).

More importantly, however, we propose thatbilingual instruction should create a more opti-mal contact setting and thus produce evenmore positive intergroup attitudes and greaterperceptions of similarity to Latino targets. Thus,we further propose that White children in bi-lingual classes will show less bias in favor of Whiteover Latino targets on measures of (a) positive evalu-ations, (b) friendship preferences, and (c) perceivedsimilarity, relative to White children in ethnicallymixed, English-only classes (Hypothesis 3).


313


Finally, we expect that bilingual instructionwould carry no cost to White children’s evalu-ation of themselves. Therefore, we predict thatWhite children in bilingual classes will report self-evaluations that are equally positive as those ofchildren in English-only classes (Hypothesis 4).

Method

Schools and sample Five Central California schools were selectedbecause the student populations included pri-marily Latino and White children. The schoolsalso varied in terms of socioeconomic status,geographic location (urban vs. rural), and therelative proportions of White and Latino stu-dents. Information about school populationdemographics was obtained in discussions withprincipals and teachers. Thus, these descrip-tions represent broad generalizations abouteach school.

Two schools were predominantly White(approximately 70% of the student population).These schools were located in upper middleclass neighborhoods, and most Latino studentswere ‘bussed’ from poorer neighborhoods. Twoother schools had more balanced enrollmentsof White and Latino students and were locatedin neighborhoods that were economically andethnically mixed, with most children comingfrom working class and lower middle classfamilies. The final school was predominantlyLatino (over 75% of the student population).This school was in a farming community, andmost of the White children were working class,while most of their Latino classmates were thechildren of farm workers (most with relativelypermanent jobs in the area).

School principals approved access to kinder-garten through second grade classrooms,pending approval of classroom teachers. Onlyone teacher declined. A total of 38 classroomshad adequate numbers of White children foruse in the study. The response rate for parentalapproval ranged from 54% to 100% acrossclassrooms. Non-participation resulted almostexclusively from children failing to returna parental permission form (less than 4% of parents who returned a form refused

participation). The present study includes onlyresponses from White children.3

The final sample of participants consisted of351 White students (167 girls and 184 boys)from a total of 38 kindergarten (n = 130, age5–7 years), first grade (n = 110, age 6–8 years),and second grade (n = 101, age 7–9 years)classrooms. The children’s mean age was 6.55years.

Classroom typesTeachers of participating classes completed ashort survey, in which they reported the amountof Spanish and English used in classroominstruction and the ethnic/linguistic heritage ofeach child in the class. Classrooms were thendivided into three categories: (a) Ethnically-Mixed Bilingual Instruction (25–88% Latinostudents; 35–60% instruction in Spanish), (b)Ethnically-Mixed English-Only Instruction(25–88% Latino students; less than 10%instruction in Spanish), and (c) Predominantly-White English-Only Instruction (less than 7%Latino students; less than 10% instruction inSpanish). There were 109 students in 16Ethnically-Mixed, Bilingual classes (Bilingual),80 students in 12 Ethnically-Mixed, English-Only classes (Mixed E-O), and 162 students in10 Predominantly-White, English-Only classes(White E-O).

While the percentage of students in each typeof classroom does vary somewhat from school toschool, we have done everything possible tocheck that any effects of type of classroomcannot be accounted for by school. First, threeof the five schools had all three types of classes—both upper middle class schools and onemixed neighborhood school. The remainingtwo schools had Mixed E-O and Bilingualclasses (not White E-O classes). Thus, compari-sons between bilingual instruction and English-only instruction can be made in all five schools.

It is also important to note that students arenested within classrooms. Thus, observations atthe level of individual children may representdata points that are not entirely independent.The problem of non-independence can beaddressed by using classroom as the unit ofanalysis. This, of course, greatly reduces the


314


power of the test—reducing the cases from 351to 38. However, to demonstrate consistency inpatterns of effects across levels of analysis, wewill perform key analyses twice, first with childand then again with classroom as the unit ofanalysis.

MaterialsMaterials and procedures used here wereadapted from Wright and Taylor (1995). Over200 head-and-shoulders Polaroid photos of5–7 year old Latino and White children werepretested using three Latino and three Whiteadults. These six raters indicated the gender andethnicity of the child in the photograph, andrated the photograph on three criteria, using5-point Likert-type scales: (a) the clarity of thephotograph, (b) the positivity of the child’sfacial expression, and (c) the child’s physicalattractiveness. The photographs were firstscreened such that only those photographs onwhich there was 100% agreement about genderand ethnicity were considered. Photographswere then matched on the remaining three cri-teria, to create same-gender Latino/White pairsthat were near equivalent on each of the threeratings. Four (2 male and 2 female) Latino/White matched pairs were combined to create apacket containing eight photographs. Six sepa-rate eight -picture testing packets were created.

ProceduresWe ensured that children did not know anychildren in the photographs.4 Testing packetswere randomly distributed to testers beforeeach session, and each packet was used approx-imately equally in each school. Children weretested individually by a White female tester.Children were taken from their regular classesto a quiet place in the school. At the beginningof the session, the tester took a Polaroid photo-graph of the child and explained the activity asthe photograph developed. The child’s photo-graph was then added to the testing packet. Allnine photographs (the child’s and the eighttarget photographs) were shuffled and placedin random order in front of the child. Thetester then asked the child to sort the ninephotographs on a series of trials, using a

standard request: ‘Pick all the children who are__________ and put them in this box, and leaveall the children who are not ___________ on thetable’. On each trial, the child was free to selectall, some, or none of the nine photographs.Before each sorting trial, the tester would againshuffle the photographs and place them in arandom order in front of the child.

The child was first asked to pick (1) the ‘girls’and (2) the ‘boys’. As children readily identifygender by this age (Cole & Cole, 1993), theserequests were used to ensure that the childunderstood the nature of the task.

Dependent measuresThe child’s selection and rejection of Latinoand White targets in the subsequent sortingrequests were used to assess: (a) Ingroup/Out-group Evaluations, (b) Friendship Preferences,and (c) Perceived Similarity. In addition, thechild’s selection of his or her own picture inresponse to the six evaluation items was used asa measure of Personal Self-Evaluation.

Ingroup/outgroup evaluations The child wasasked to sort the photographs in response to six evaluative items—selecting those: (1) whoare good at lots of things; (2) who are smart;(3) who have lots of friends; (4) who are nice;(5) who are happy; and (6) who like school. Foreach trial, the number of Latino and thenumber of White targets were tabulated,creating separate ratings of ingroup evaluationand outgroup evaluation. In order to clarify thepresentation of the results, the number oftargets was converted into percentages of thepossible total, such that the child’s score oneach trial could be 0% (no targets from thatethnic group selected), 25% (one of the fourtargets), 50% (two of the four), 75% (three ofthe four), or 100% (all targets from that ethnicgroup selected). Responses to the six sortingtrials were combined to produce measures ofIngroup Evaluation (� = .80) and OutgroupEvaluation (� = .82).

Friendship preferences On a single sortingtrial, children were asked to pick the targetsthey would like as best friends. Children’s


315


choices were tabulated separately for Latino andWhite targets and converted to percentages,creating two ratings that ranged from 0% (notargets from an ethnic group selected) to 100%(all four targets from an ethnic group selected).

Perceived similarity On a single sorting trial,children were asked to pick all the targets whoare most like them. Children’s choices weretabulated separately for Latino and Whitetargets and converted to percentages, creatingtwo ratings that ranged from 0% (no targetsfrom an ethnic group selected) to 100% (alltargets from an ethnic group selected).

Personal self-evaluation Since the packet oftarget photographs included the child’s ownpicture, the child’s self-evaluation could becalculated using the number of times he/sheselected his/her own picture on the six evalu-ation sorting trials (good at lots of things,smart, lots of friends, nice, happy, and likeschool). These scores ranged from 0% (neverselected his or her own picture) to 100%(selected his or her picture on all six trials).

Supplementary measuresTeacher survey After all the children in a classwere tested, the teacher completed a briefsurvey. Teachers first reported the amount ofEnglish and Spanish used in classroom instruc-tion. They then indicated each child’s ethnic/linguistic background.

Principal survey After testing in a school wascompleted, the principal was contacted byphone and asked about his/her policy regard-ing the assignment of students to classes inKindergarten, First and Second Grade class-rooms. Questions included: (a) the percentageof White/Anglo children assigned to class-rooms on the basis of preferences expressed bytheir parents; (b) the percentage of Spanish-speaking children assigned to classrooms onthe basis of preferences expressed by their par-ents; and (c) the degree to which these ‘special’assignments based on parental preference hadto do with preference for bilingual versusEnglish-only instruction.

Responses to this survey allowed us to evalu-ate the degree to which parental preferencemight represent a meaningful confound. Acrossthe five schools, principals’ estimates of White/Anglo children placed in classrooms on thebasis of parental preference ranged from 0% to30%. In addition, principals who did makesome special placements indicated thatbetween 30% and 50% of these were for reasonsother than language of instruction. Thus, evenin the school with the highest degree ofparental influence, only approximately twochildren per classroom were there because ofparental preference for a particular languageprogram. Thus, parental selection appears toplay a minimal role in the distribution of chil-dren across classroom type.5

Results

Analytic strategyMuch of the interest in understanding chil-dren’s intergroup attitudes involves attitudes atthe negative end of the spectrum—that is, atti-tudes that can be considered ‘prejudice’. Thisraises the question: ‘At what point can we call achild’s pattern of preferences “prejudice”?’ (seeAboud, 1987; Nesdale, 2001). One response hasbeen to compare children’s appraisals of andpreferences for ingroup members versus out-group members. The larger the differencebetween these two appraisals/preferences, themore negative the intergroup attitude.

However, this approach presents three prob-lems. First, under some circumstances, somepreferential evaluation of the ingroup may beadaptive and appropriate (see Wright & Taylor,1995). Second, difference scores ignore theabsolute level of each evaluation. For example,consider Child A, who describes ingroup targetsas positive in six of six trials and outgroup tar-gets as positive in three of six trials. This childhas a difference score of 3, a relatively largeingroup preference. Child B, who describesingroup targets as positive in three of six trialsand outgroup targets as positive in one of sixtrials, has a difference score of 2. Using onlydifference scores, Child B is seen to have morepositive intergroup attitudes than is Child A.


316


However, in absolute terms Child B has clearlyrated the outgroup lower than Child A. Thequestion of which child is more prejudiced (orwhether either should be labeled as preju-diced) is not easily answered. Third, there aregood reasons to believe that positive evaluationof the ingroup and negative evaluations of theoutgroup may represent related but distinctpsychological processes (see Brewer, 1999).

Our analytic strategy attempts to addressthese issues by reporting both absolute valuesand comparisons of ingroup and outgroupevaluations and preference using a repeatedmeasures design. Additional analyses are alsoperformed to isolate subgroups of children whodemonstrate complete rejection of all outgroupmembers (see Katz, Sohn, & Zalk, 1975) and/ordemonstrate no bias in evaluations of and pref-erences for the ingroup and the outgroup.Complete rejection of the outgroup might bedemonstrative of ‘prejudice’, while equal pref-erences for ingroup and outgroup membersseems to represent a ‘non-prejudice’ pattern ofresponse. Comparing the size of these two sub-groups across different classroom contexts pro-vides an additional test of the relative impact ofbilingual and English-only instruction on inter-group attitudes.

Preliminary analysesAccuracy in identifying targets To determinewhether the children understood the task andwere able to follow the tester’s instructions, weexamined their accuracy in sorting the targetphotos on the basis of gender. Of 351 children,28 failed to identify all eight target photoscorrectly by gender. Of these, 27 children mis-labeled one photo, and one child mislabeledtwo photos. Thus, it appears that children gen-erally understood the task and were able tofollow the directions. The children makingerrors were relatively equally distributed acrossthe three types of classes (13 in White E-Oclasses, 8 in the Mixed E-O classes, 7 in theBilingual classes).

Gender effects Children’s scores on ingroup/outgroup evaluation, friendship preferences,and perceived similarity were analyzed using 2

(Participant Gender) � 3 (Type of Classroom)� 2 (Ethnicity of Target) mixed analyses of vari-ance. No significant main or interaction effectsinvolving gender emerged, and thus gender wasdropped from subsequent analyses.

Grade effects Children’s scores on ingroup/outgroup evaluation, friendship preferences,and perceived similarity were initially examinedusing 3 (Grade) � 3 (Type of Classroom) � 2(Ethnicity of Target) mixed analysis of variance(ANOVA). The only significant effects involvinggrade were main effects of grade for friendshippreference and perceived similarity, such thatkindergarteners selected more friends and sawmore targets as similar to them than did chil-dren in Grades 1 and 2—who do not differsignificantly on either of these measures. Nosignificant two-way or three-way interactionsemerged and grade was dropped from subse-quent analyses.

Evaluations, friendship preferences, andperceived similarity to the selfThe 3 (Type of Classroom) � 2 (Ethnicity ofTarget) mixed ANOVAS were performed twice,first with child as the unit of analysis and thenwith classroom as the unit of analysis. Type ofClassroom (White E-O, Mixed E-O, Bilingual)was a between-subjects factor, and Ethnicity ofTarget (White, Latino) was a within-subjectsfactor.

Child as unit of analysis The analysis forEvaluations supported Hypothesis 1, yielding asignificant main effect of Ethnicity of Target(F(1, 348) = 48.58, p < .001, �2 = .12), indicatinga general bias toward White (71%) over Latino(63%) targets. A significant two-way interactionalso emerged (F(2, 348) = 6.35, p = .002, �2 =.03) (see Figure 1). Simple effects tests indi-cated no significant differences in evaluationsof White targets among the three types of class-rooms (F(2, 348) = 0.64, p = .53, �2 < .001), buta significant difference in evaluations of Latinotargets among the three types of classrooms(F(2, 348) = 5.08, p = .007, �2 = .03). Consistentwith Hypothesis 2a, pairwise comparisonsrevealed that children in both Bilingual classes


317


(F(2, 348) = 8.54, p = .004, �2 = .03) and MixedE-O classes (F(2, 348) = 4.93, p = .03, �2 = .02)evaluated Latinos significantly more positivelythan children in the White E-O classes. How-ever, the differences in evaluations of Latinotargets between children in Bilingual andMixed E-O classes was not significant, thusfailing to provide support of Hypothesis 3a.

The analysis for Friendship Preference also sup-ported Hypothesis 1, yielding a significant maineffect of Ethnicity of Target (F(1, 348) = 85.94,p < .001, �2 = .20), indicating a general prefer-ence for Whites (46%) over Latinos (30%) asfriends. The main effect of Type of Classroom(F(2, 348) = 4.51, p = .01, �2 = .02) and the two-way interaction (F(2, 348) = 7.50, p = .001, �2 =.04) were also significant (see Figure 2). Simpleeffects tests indicated no significant differencesamong the three types of classrooms for selec-tion of Whites as friends (F(2, 348) = 1.87, p =.16, �2 < .01) but significant differences amongthe types of classrooms for selection of Latinos(F(2, 348) = 8.23, p < .001, �2 = .04). Pairwise

comparisons revealed partial support forHypothesis 2b, as children in Bilingual classeswere more likely to select Latino friends thanthose in White E-O classes(F(2, 348) = 14.88,p > .001, �2 = .05), yet children in Mixed E-Oand White E-O classes did not differ (F < 1.0).Additionally, Hypothesis 3b was supported, withchildren in Bilingual classes selecting moreLatino targets as friends than those in MixedE-O classes (F(2, 348) = 5.54, p = .02, �2 = .03).

The analysis for Perceived Similarity also sup-ported Hypothesis 1, with a significant maineffect of Ethnicity of Target (F(1, 348) = 142.48,p < .001, �2 = .31). White children generallyselected more Whites (38%) over Latinos(17%) as similar to the self. The main effectof Type of Classroom (F(2, 348) = 22.88,p < .001, �2 = .12) was also significant. Post hoccomparisons6 revealed that children in Bilingualclasses selected significantly more targets as likeself (38%) than children in either Mixed E-Oclasses (21%) or White E-O classes (23%). Theinteraction approached statistical significance


318

Figure 1. Evaluations of Latino and White targets by White children in three different types of classrooms.


(F(2, 348) = 2.46, p = .10, �2 = .02) (see Figure3), and the pattern of results is consistent withthose found for Friendship Preferences. Par-tially supporting Hypothesis 2c and fully support-ing Hypothesis 3c, children in Bilingual classesselected more Latino targets as similar to them-selves than children in either the Mixed E-Oclasses (F(2, 348) = 17.03, p < .001, �2 = .08) orchildren in the White E-O classes (F(2, 348) =28.37, p <.001, �2 = .09). All other comparisonswere not significant.

Classroom as unit of analysis The analysis forEvaluations supported Hypothesis 1, yielding asignificant main effect of Ethnicity of Target(F(1, 35) = 18.37, p < .001, �2 = .34). Whitechildren demonstrated a general bias towardWhite (71%) over Latino (63%) targets. A sig-nificant two-way interaction also emerged (F(2,35) = 4.00, p = .03, �2 = .18). Simple effects testsindicated that the pattern of means for thisinteraction were identical to those found in the

analyses using children as the unit of analysis.There were no significant differences amongthe three types of classes for evaluations ofWhites (F(2, 35) = 0.002, p = .99). However,pairwise comparisons showed that children inBilingual classes evaluated Latino targets signifi-cantly more positively (70%) than children inWhite E-O classes (61%) (F(2, 35) = 6.19, p =.021, �2 = .20) and the difference betweenchildren in Bilingual (70%) and Mixed E-Oclasses (64%) approached significance (F(2,35) = 2.85, p = .10, �2 = .05).

The analysis for Friendship Preference also sup-ported Hypothesis 1, yielding a significant maineffect of Ethnicity of Target (F(1, 35) = 45.59,p < .001, �2 = .56) such that the childrenshowed a general preference for Whites (47%)over Latinos (35%). The main effect for Type ofClassroom was also significant (F(2, 35) = 5.07,p = .01, �2 = .22) and although the two-wayinteraction did not reach traditional signifi-cance levels (F(2, 35) = 2.11, p = .12, �2 = .09),


319

Figure 2. Friendship preferences for Latino and White targets by White children in three different types ofclassrooms.

Latino Targets as Friends

White Targets as Friends


simple effects tests and pairwise comparisonsdemonstrated a pattern of responses consistentwith those found using child as the unit ofanalysis. There were no differences across thethree classroom types in the selection of Whitefriends (F(2, 35) = 2.24, p = .12), while therewere significant differences in the selection ofLatino friends (F(2, 35) = 5.85, p < .01, �2 = 25).Consistent with Hypotheses 2b and 3b, pairwisecomparisons revealed those in Bilingual classes(43%) selected significantly more Latino friendsthan those in Mixed E-O classes (31%) (F(2, 35)= 5.98, p = .02, �2 = .18) and White E-O classes(27%)(F(2, 35) = 8.86, p = .007, �2 = .27).

The analysis for Perceived Similarity yieldedsupport for Hypothesis 1, with a significant maineffect of Ethnicity of Target (F(1, 35) = 53.56,p < .001, �2 = .60). Children generally perceivedmore White targets as like themselves (M =37%) than Latino targets (M = 20%). The maineffect of Type of Classroom was also significant(F(2, 35) = 6.63, p = .004, �2 = .27), such that

children in Bilingual classes (M = 39%) selectedmore targets from both groups as similar to selfthan children in either Mixed E-O classes (M =24%) or White E-O classes (M = 23%). Theinteraction effect was not significant (F(1, 35) =0.02, p = .95, �2 < .001), thus offering no directsupport for Hypotheses 2c or 3c in this analysis.

Examining subgroups of children: No ingrouppreference and complete outgroup rejectionAdditional analyses were conducted on twosubgroups of particular interest: (a) childrendemonstrating complete rejection of Latinotargets, and (b) those who demonstrated nopreference for Whites over Latinos. Friendshippreferences and perceived similarity measureswere used to define these subgroups, with onesubgroup including children who selected onlyWhites as potential friends (or as similar toself), while the other subgroup included chil-dren who selected an equal number of Whitesand Latinos as friends (or as similar to self).


320

Figure 3. Perceived similarity of Latino and White targets to self by White children in three different types ofclassrooms.


The distributions of these subgroups across theclassroom types were compared using child asthe unit of analysis.7

Friendship preferences Table 1 presents thepercentages of children in each type of class-room who: (a) chose only Whites as friends; or(b) chose equal numbers of Whites and Latinosas friends. The 2 � 3 chi-square was significant(�2(2, 230) = 17.29, p < .001). Consistent withthe essence of Hypotheses 2b and 3b, the percent-age of children selecting only Whites was morethan twice as high in White E-O and Mixed E-Oclasses than in Bilingual classes. Also, nearlyhalf of the children in Bilingual classes choseequal numbers of Whites and Latinos, whileonly one third in Mixed E-O and slightly morethan one quarter in White E-O classes choseequal numbers of Whites and Latinos.

Perceived similarity Table 2 presents the per-centages of children in each type of classroomwho: (a) chose only Whites as similar to self; or(b) chose equal numbers of Whites and Latinosas similar to self. The 2 � 3 chi-square wassignificant (�2(2, 265) = 8.53, p < .01). Consist-ent with the essence of Hypotheses 2c, and 3c, the

percentage of children selecting only Whites assimilar to self was higher in White E-O andMixed E-O classes than in Bilingual classes.Differences between classroom types weresmaller for the percentage of children selectingequal numbers of Latinos and Whites, but thepattern was in the predicted direction.

Self-evaluation and type of classroomenvironmentUsing child as the unit of analysis, a one-wayANOVA compared self-evaluations for childrenin each of the three types of classrooms. Thisanalysis yielded no significant differences (F(2,353) = 0.95, p = .39, �2 = .005), with most of thechildren in all three groups selecting their ownphotograph in response to all or most of thepositive evaluation items (94% for children inWhite E-O classes, 91% for children in MixedE-O classes, and 92% for children in Bilingualclasses).

Discussion

Consistent with Hypothesis 1 and previous find-ings (see Aboud, 1988), White children showed apreferential bias toward ingroup over outgroup


321

Table 1. Percentage and numbers of White children who chose only Whites or who chose equal numbers ofWhites and Latinos as potential friends by type of classroom

Type of classroom Only White targets Equal numbers of White/Latino targets

White E-O 38.7% 28.2%(n = 163) (n = 63) (n = 45)Mixed E-O 36.3% 33.8%(n = 80) (n = 29) (n = 27)Bilingual 16.8% 46.0%(n = 113) (n = 19) (n = 52)

Table 2. Percentages of White children who chose only Whites or equal numbers of Whites and Latinos assimilar to themselves by type of classroom

Type of classroom Only White targets Equal numbers of White/Latino targets

White E-O 58.3% 23.3%(n = 163) (n = 95) (n = 38)Mixed E-O 47.5% 28.8%(n = 80) (n = 38) (n = 23)Bilingual 34.5% 32.7%(n = 113) (n = 39) (n = 37)


targets. They evaluated White targets more pos-itively than Latino targets, selected more Whitethan Latino targets as potential friends, andwere more likely to see Whites than Latinos assimilar to themselves. However, as predicted,classroom environment moderated the strengthof these biases. Overall, findings across thethree measures (evaluation, friendship prefer-ence, and perceived similarity) support the pre-diction that bilingual instruction can contributeto more positive perceptions of the ethnolin-guistic outgroup than English-only instruction.In addition, this greater positivity toward theoutgroup occurs with no loss in positivitytoward the White ingroup.

Moreover, these results clearly suggest thatthe positive effect of bilingual instruction is notsimply the result of a more integrated class-room. Greater proportions of Latino classmatesin English-only classes corresponded with morepositive evaluations of Latinos, but not withgreater preferences for Latinos as friends orperceiving more Latinos as similar to self.Instead, only bilingual instruction was associ-ated with a greater tendency to select Latinotargets as friends and greater perceptions ofLatino targets as similar to the self. Thus, whileethnically balanced classes can contribute topositive intergroup attitudes, it appears thatlanguage of instruction has an additional posi-tive impact on children’s orientations towardmembers of an ethnolinguistic outgroup.

Also noteworthy is that the children arereporting evaluations, friendship preferences,and perceived similarity in response to photosof children that they have never seen before.Thus, we can conclude that bilingual instruc-tion is associated with more positive orienta-tions toward Latino children generally, and notjust the specific Latino children with whom theyhave contact. Most contact research in edu-cational settings has focused on attitudestoward one’s classmates, rather than general-ized attitudes (see Schofield, 1995). Whileachieving harmonious intergroup relationswithin the immediate contact situation is a valu-able goal, only generalized positive attitudeswill improve interactions with new outgroupmembers in new contact situations (see Brown

& Hewstone, 2005; Hewstone & Brown, 1986;Rothbart & John, 1985). Thus, the presentfindings are particularly important, in that theyshow that the positive outgroup orientationlearned in bilingual classes can generalize tothe outgroup as a whole.

Moreover, the more positive outgroup orien-tation found among White children in bilingualclasses did not come at the expense of the chil-dren’s evaluations of themselves. Children gavevery positive self-evaluations, which did not varyin relation to the language of instruction orrepresentations of Latino children in the class-room. Thus, the more positive intergroup atti-tudes associated with bilingual instructioncomes without any associated negative impacton children’s evaluations of themselves or theirown group.

Our findings also complement a number ofcentral themes that have been stressed in recenttheory and research on intergroup contact.Specifically, the results suggest that bilingualinstruction may foster the type of equal status,authority supported, cooperative contact thatencourages positive intergroup attitudes (seeAllport, 1954; Pettigrew & Tropp, 2000), andthis is accomplished in a context where ethnicgroup membership remains salient (see Brown& Hewstone, 2005). The results concerningchildren’s friendship preferences also providemore direct support for the potential role offriendship in these intergroup contact effects(Wright et al., 2005; Pettigrew, 1997; Wrightet al., 1997). Indeed, it may be that children inbilingual classes have been particularly success-ful in developing friendly relationships withLatino classmates, such that these friendshipexperiences account for the more general posi-tive attitudes toward Latino targets.

Additionally, though we have focused on theexperiences of White children, we might alsoconsider the significance of our findings for theeducational experiences of minority languagestudents. Being the target of prejudice and biascan have a number of negative effects (see Swim& Stangor, 1998), and minority language chil-dren will likely benefit from an environment inwhich their classmates hold them and theirgroups in high esteem (Berry & Williams,


322


2004). Moreover, the positive intergroup atti-tudes of their White classmates might have animpact on minority language speakers’ acquisi-tion of English. There is evidence that motiva-tion to learn a second language, success inlearning that language, and involvement in theeducational context are all influenced by thequality of one’s interactions with members ofthe target language group (see Gardner &Clément, 1990). Furthermore, Genesee et al.(1983) have shown that students’ expectationsof support from the target language group weresignificantly related to various measures ofsecond language use, proficiency, and willing-ness to affiliate with members of the target lan-guage group (see also Clément & Kruidenier,1985). Finally, there is also evidence that havingthe opportunity to develop friendships withWhites can be an important predictor ofsubsequent success for minority students(Braddock, 1985). Thus, improving the inter-group attitudes of White students may in turnhave important implications for the classroomexperiences and success of minority languagechildren.

Conclusions

The extensive debates over bilingual educationhave generally overlooked how the language ofinstruction can impact intergroup attitudes. Atthe same time, research on intergroup attitudesand contact between groups has largely ignoredthe role of language use in these intergroupprocesses. This research links these areas ofinvestigation and extends discussions of bilin-gual instruction beyond issues of academicachievement and language proficiency, byfocusing on the intergroup implications ofclassroom language use.

Language is used to communicate and sharing alanguage certainly aids in the development ofcross-group relationships, but the use of a lan-guage communicates important cues about who isvalued and thus how we should expect cross-group interactions to progress. Thus, while useof a minority language in the classroom mightenhance the possibility that White English-speaking children would grow to learn the

language of their minority classmates, the moreimportant consequence, perhaps, is thatinstruction in the minority language structuresthe classroom context to improve the quality ofthe intergroup contact. Indeed, language isnot only a marker of group differences but alsois a tool that can be used to diminish (orstrengthen) status inequalities, enhance (orundermine) interdependence, and demon-strate authority support (or lack thereof) forintergroup contact.

Results from this study show important differ-ences in the attitudes of White children inEnglish-only classes and those receiving bi-lingual instruction. While part of these differ-ences may be accounted for by the greaterrepresentation of Latino children in bilingualclasses, contact alone does not produce thepattern of positive attitudes found among Whitechildren who also receive consistent exposureto instruction in the Spanish language. Overall,these results suggest that bilingual instructionmay have positive effects on the generalizedintergroup attitudes of White children, support-ing the claim that recent efforts to dismantlebilingual education may represent a seriousbackward step in the struggle to improveinterethnic relations in our schools and oursociety.

Notes1. To avoid the many difficulties associated with

defining ‘bilingual education’ and attempting tomatch actual classroom practices with variedmodels of bilingual education (see Brisk, 1998),we will use the term ‘bilingual instruction’. Here,bilingual instruction is meant to describe a moregeneral classroom context in which a significantamount of instruction is provided in a languageother than English. We compare this to‘English-only instruction’ where English is theexclusive (or near exclusive) language ofinstruction. In the present research, bilingualinstruction is operationalized as a classroom inwhich the teacher reports that between 30% and60% of the classroom instruction is provided inSpanish.

2. Lambert, Genesee, and their colleagues haveinvestigated the ethnic attitudes of


323


English-speaking Canadian children in Frenchimmersion programs and found fairly consistentlythat French immersion education has a positiveimpact on attitudes toward Francophones(e.g. Genesee, 1987; Lambert & Tucker, 1972).However, these programs involve classrooms andschools that are almost entirely made of nativeEnglish speakers who are instructed entirely inFrench. This context differs substantially from thecontext being considered here in that the Frenchimmersion context often does not involve directintergroup contact across linguistic groups andinstruction in both languages.

3. More than 400 Latino children were also tested.Data from these children are presented elsewhere(see Tropp & Wright, 2003).

4. Approximately three-quarters of the photographswere taken at schools other than those used inthe present study. The remaining photographswere taken at the rural school and the twopackets (of six) containing these pictures werenot used at that school.

5. In addition, we performed a number of analysescomparing patterns of results across the fivedifferent schools. These analyses showedsurprising consistency across schools. Further, theschool at which the principal indicated thatparental preference played no part in theassignment of students to classes showed one ofthe strongest and most consistent effects acrossthe three dependent measures. The school atwhich the principal indicated the largest amountof parental influence on children’s classroomplacement (30%) fell in the middle of the groupin terms of the consistency and strength of thepredicted effects.

6. All post hoc comparisons use Newman-Kuels, � = .05.

7. This analysis can only be done using child as theunit of analysis. However, given that most of theprevious analyses showed that, despite thesubstantial loss in power associated with analysesat the level of classroom, most of the key effectsremained significant at this level of analysis. Thisconsistency should provide some additionalconfidence regarding the analyses in which childwas used as the unit of analysis.

AcknowledgmentsThe authors would like to thank the principals andteachers who supported this work by allowing usaccess to their students. Also, we thank the many

undergraduate students at the University ofCalifornia, Santa Cruz, who provided invaluableassistance in the collection of these data: AnabelKnoche, Virginia Martinez, Nicole Seno, LennaGochoel, Brenda Abrego, Olona Dientz, KristenMartinez-Pedersen, Ryah Lindemuth, Ana Quiran,Sonia Ramos, Pam Barnett, Carolina Camarena,Monica Curtis, and Roberta Romero. This researchwas supported in part by a research grant from theSocial Sciences Division, University of CaliforniaSanta Cruz and by a Regents Faculty Research Grant,University of California Santa Cruz.

ReferencesAboud, F. E. (1977). Interest in ethnic information:

A cross-cultural developmental study. CanadianJournal of Behavioural Science, 9, 134–146.

Aboud, F. E. (1987). The development of ethnic self-identification and attitudes. In J. S. Phinney &M. J. Rotheram (Eds.), Children’s ethnic socialization:Pluralism and development (pp. 32–55). NewburyPark, CA: Sage.

Aboud, F. E. (1988). Children and prejudice. New York:Blackwell.

Aboud, F. E., & Levy, S. R. (2000). Interventions toreduce prejudice and discrimination in childrenand adolescents. In S. Oskamp (Ed.), Reducingprejudice and discrimination (pp. 269–293). Mahwah,NJ: Erlbaum.

Aboud, F. E., Mendelson, M. J., & Purdy, K. T. (2003).Cross-race peer relations and friendship quality.International Journal of Behavioral Development, 27,165–173.

Allport, G. W. (1954). The nature of prejudice.Cambridge, MA: Addison-Wesley.

Aronson, E., & Patnoe, S. (1997). The jigsaw classroom:Building cooperation in the classroom. New York:Longman.

Asher, S. R., & Allen, V. L. (1969). Racial preferenceand social comparison processes. Journal of SocialIssues, 25, 157–167.

Banks, J. A. (1995). Multicultural education and themodification of students’ racial attitudes. In W. D. Hawley & A. W. Jackson (Eds.), Towards acommon destiny: Improving race and ethnic relations inAmerica (pp. 315–339). San Francisco: Jossey-Bass.

Barker, V., Giles, H., Noels, K. A., Duck, J., Hecht, M.,& Clément, R. (2001). The English-onlymovement: A communication analysis of changingperceptions of language vitality. Journal ofCommunication, 51, 3–37.

Berry, R. S. Y., & Williams, M. (2004). In at the deep


324


end: Difficulties experienced by Hong KongChinese ESL learners at an independent school inthe United Kingdom. Journal of Language and SocialPsychology, 23, 118–134.

Braddock, J. H. (1985). School desegregation andblack assimilation. Journal of Social Issues, 41, 9–22.

Brewer, M. B. (1999). The psychology of prejudice:Ingroup love or outgroup hate? Journal of SocialIssues, 55, 429–444.

Brewer, M. B., & Brown, R. J. (1998). Intergrouprelations. In D. T. Gilbert, S. T. Fiske, & G. Lindzey(Eds.), Handbook of social psychology (4th ed.,pp. 554–594). Boston: McGraw-Hill.

Brewer, M. B., & Miller, N. (1984). Beyond thecontact hypothesis: Theoretical perspectives ondesegregation. In N. Miller & M. B. Brewer (Eds.),Groups in contact: The psychology of desegregation(pp. 281–302). New York: Academic.

Brown, R., & Hewstone, M. (2005). An integrativetheory of intergroup contact. In M. P. Zanna (Ed.)Advances in Experimental Social Psychology (Vol. 37,pp. 255–343). San Diego, CA: Elsevier AcademicPress.

Brisk, M. (1998). Bilingual education: From compensatoryto quality schooling. Mahwah, NJ: Erlbaum.

Clément, R., & Kruidenier, B. G. (1985). Aptitude,attitude and motivation in second languageproficiency: A test of Clément’s model. Journal ofLanguage and Social Psychology, 4, 21–37.

Cohen, E. G. (1980). Design and redesign of thedesegregated school: Problems of status, power,and conflict. In W. G. Stephan & J. R. Feagin(Eds.), School desegregation: Past, present and future(pp. 251–280). New York: Plenum.

Cole, M., & Cole, S. R. (1993). The development ofchildren (2nd ed.) New York: Freeman.

Corenblum, B., & Wilson, A. E. (1982). Ethnicpreference and identification among CanadianIndian and White children: Replication andextension. Canadian Journal of Behavioural Science,14, 50–59.

Covington, M. V. (1989). Self-esteem and failure atschool: Analysis and policy implications. In A. M. Mecca, N. J. Smelser, & J. Vasconcellos(Eds.), The social importance of self-esteem(pp. 72–124). Berkeley, CA: University ofCalifornia Press.

Crawford, J. (1989). Bilingual education: History,policies, theory, and practice. Trenton, NJ: Crane.

Cummins, J. (1989). Empowering minority students.Sacramento, CA: California Association forBilingual Education.

Damico, S. B., Bell-Nathaniel, A., & Green, C. (1981).Effects of school organizational structure on

interracial friendships in middle schools. Journal of Educational Research, 74, 388–393.

Doyle, A. B., & Aboud, F. E. (1995). A longitudinalstudy of White children’s racial prejudice as asocial–cognitive development. Merrill-PalmerQuarterly, 41, 209–228.

Edwards, J. R. (1977). Ethnic identity and bilingualeducation. In H. Giles (Ed.), Language, ethnicity,and intergroup relations (pp. 253–282). London:Academic Press.

Fehr, B. (1996). Friendship processes. Thousand Oaks,CA: Sage.

Foster, D. (1994). Racism and children’s intergrouporientations: Their development and the questionof psychological effects on minority groupchildren. In A. Dawes & D. Donald (Eds.),Childhood and adversity: Psychological perspectives fromSouth African research (pp. 220–239). Cape Town:David Philip.

Gaertner S. L., & Dovidio, J. F. (2000). Reducingintergroup bias: The common ingroup identity model.Philadelphia: Psychology Press.

Gardner, R. C., & Clément, R. (1990). Socialpsychological perspectives on second languageacquisition. In H. Giles & W. P. Robinson (Eds.),Handbook of language and social psychology(pp. 495–517). Chichester, UK: Wiley.

Genesee, F. (1987). Learning through two languages:Studies of immersion and bilingual education.Cambridge, MA: Newbury House.

Genesee, F., & Gándara, P. (1999). Bilingualeducation programs: A cross-national perspective.Journal of Social Issues, 55, 665–685.

Genesee, F., Rogers, P., & Holobow, N. (1983). Thesocial psychology of second language learning:Another point of view. Language Learning, 33,209–224.

Giles, H., Bourhis, R. Y., & Taylor, D. M. (1977).Towards a theory of language in intergrouprelations. In H. Giles (Ed.), Language, ethnicity, andintergroup relation (pp. 307–348). London:Academic Press.

Greene, J. P. (1997). A meta-analysis of the Rosselland Baker review of bilingual education research.Bilingual Research Journal, 21, 103–121.

Gudykunst, W. B., & Schmidt, K. L. (1987).Language and ethnic identity: An overview andprologue. Journal of Language and Social Psychology,6, 157–170.

Hakuta, K. (1986). Mirror of language: The debate onbilingualism. New York: Basic Books.

Hakuta, K., Butler, Y. G., & Witt, D. (2000). How longdoes it take English learners to attain proficiency?(Policy Report 2000–1). Santa Barbara, CA: The


325


University of California Linguistic MinorityResearch Institute.

Harter, S. (1986). Processes underlying theconstruction, maintenance, and enhancement ofthe self-concept of children. In J. Suls & A. Greenwald (Eds.), Psychological perspectives on theself (pp. 136–182). Hillsdale, NJ: Erlbaum.

Heller, M. (1987). The role of language in theformation of ethnic identity. In J. S. Phinney &M. J. Rotheram (Eds.), Children’s ethnic socialization:Pluralism and development (pp. 180–200). NewburyPark, CA: Sage.

Hewstone, M., & Brown, R. (1986). Contact is notenough: An intergroup perspective on the‘Contact Hypothesis’. In M. Hewstone & R. Brown(Eds.), Contact and conflict in intergroup encounters(pp. 1–44). Oxford, UK: Basil Blackwell.

Johnson, D. W., Johnson, R., & Maruyama, G. (1984).Goal interdependence and interpersonalattraction in heterogeneous classrooms: A meta-analysis. In N. Miller & M. B. Brewer (Eds.),Groups in contact: The psychology of desegregation(pp. 187–212). New York: Academic Press.

Kagan, S. (1986). Cooperative learning andsociocultural factors in schooling. In BilingualEducation Office (Ed.), Beyond language: Social andcultural factors in schooling language minority students(pp. 231–298). Sacramento, CA: California StateDepartment of Education.

Katz, P. A. (1983). Developmental foundations ofgender and racial attitudes. In R. L. Leahy (Ed.),The child’s construction of social inequality(pp. 41–78). New York: Academic Press.

Katz, P. A., Sohn, M., & Zalk, S. R. (1975). Perceptualconcomitants of racial attitudes in urban grade-school children. Developmental Psychology, 11,135–144.

Khmelkov, V. T., & Hallinan, M. T. (1999).Organizational effects on race relations in schools.Journal of Social Issues, 55, 627–645.

Kistner, J., Metzler, A., Gatlin, D., & Risi, S. (1993).Classroom racial proportions and children’s peerrelations: Race and gender effects. Journal ofEducational Psychology, 85, 446–452.

Lambert, W. E., & Cazabon, M. (1994). Students’ viewsof the Amigos program (Research Report No. 11).Santa Cruz, CA: University of California, NationalCenter for Research on Cultural Diversity andSecond Language Learning.

Lambert, W. E., Genesee, F., Holobow, N. E., &Chartrand, L. (1993). Bilingual education formajority English-speaking children. EuropeanJournal of Psychology of Education, 8, 3–22.

Lambert, W. E., & Tucker, G. R. (1972). Bilingual

education of children: The St. Lambert experiment.Rowley, MA: Newbury House.

Lindholm, K. J. (1994). Promoting positive cross-cultural attitudes and perceived competencein culturally and linguistically diverse classrooms.In R. A. Devillar, C. J. Faltis, & J. P. Cummins(Eds.), Cultural diversity in schools: From rhetoric topractice (pp. 189–206). Albany, NY: State Universityof New York Press.

McLaughlin-Volpe, T., Aron, A., Wright, S. C., & Reis, H. T. (2000). Intergroup social interactions andintergroup prejudice: Quantity versus quality.Manuscript under review.

Miller, N., & Brewer, M. B. (1984). Groups in contact:The psychology of desegregation. New York: AcademicPress.

Nesdale, D. (2001). Language and the developmentof children’s ethnic prejudice. Journal of Languageand Social Psychology, 20, 90–110.

Noels, K. A., Pon, G., & Clément, R. (1996).Language, identity, and adjustment: The role oflinguistic self-confidence in the acculturationprocess. Journal of Language and Social Psychology,15, 246–264.

Parker, L. L. (1978). Current perspectives. In Centerfor Applied Linguistics (Ed.), Bilingual education:Current perspectives (pp. 1–62). Arlington, VA:Center for Applied Linguistics.

Pettigrew, T. F. (1997). Generalized intergroupcontact effects on prejudice. Personality and SocialPsychology Bulletin, 23, 173–185.

Pettigrew, T. F. (1998). Intergroup contact theory.Annual Review of Psychology, 49, 65–85.

Pettigrew, T. F., & Tropp, L. (2000). Does intergroupcontact reduce prejudice? Recent meta-analyticfindings. In S. Oskamp (Ed.), Reducing prejudiceand discrimination: Social psychological perspectives(pp. 93–114). Mahwah, NJ: Erlbaum.

Pettigrew, T. F., & Tropp, L. R. (in press). A meta-analytic test of intergroup contact theory.Manuscript under review. Journal of Personality andSocial Psychology.

Ramsey, P. G. (1987). Young children’s thinkingabout ethnic differences. In J. S. Phinney & M. J. Rotheram (Eds.), Children’s ethnic socialization:Pluralism and development (pp. 56–72). NewburyPark, CA: Sage.

Rothbart, M., & John, O. P. (1985). Socialcategorization and behavioral episodes: Acognitive analysis of the effects of intergroupcontact. Journal of Social Issues, 41, 81–104.

Schofield, J. W. (1989). Black and White in school:Trust, tension, or tolerance? New York: TeachersCollege Press.


326


Schofield, J. W. (1995). Promoting positiveintergroup relations in school settings. In W. D. Hawley & A. W. Jackson (Eds.), Towards acommon destiny: Improving race and ethnic relations inAmerica (pp. 257–289). San Francisco: Jossey-Bass.

Schofield, J. W., & Sagar, H. A. (1977). Peerinteraction patterns in an integrated middleschool. Sociometry, 40, 130–138.

Slavin, R. E., & Cheung, A. (1999). Effective readingprograms for English language learners: A best evidencesynthesis (Report No. 66). Baltimore, MD: JohnsHopkins University, Center for Research on theEducation of Students Placed At Risk.

Slavin, R. E., & Cooper, R. (1999). Improvingintergroup relations: Lessons learned fromcooperative learning programs. Journal of SocialIssues, 55, 647–663.

Stephan, W. G., & Rosenfield, D. (1978). Effects ofdesegregation on racial attitudes. Journal ofPersonality and Social Psychology, 36, 795–804.

Sue, S., & Padilla, A. (1986). Ethnic minority issuesin the United States: Challenges for theeducational system. In Bilingual Education Office(Ed.), Beyond language: Social and cultural factors inschooling language minority students (pp. 35–72).Sacramento, CA: California State Department ofEducation.

Swim, J. K., & Stangor, C. (Eds.). (1998). Prejudice:The target’s perspectives. San Diego, CA: AcademicPress.

Tajfel, H. (1978). Differentiation between social groups:Studies in the social psychology of intergroup relations.London: Academic Press.

Tropp, L. R., & Wright, S. C. (2003). Examiningegocentrism and ethnocentrism among youngLatino and White children. Self and Identity, 2,203–221.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D.,& Wetherell, M. S. (1987). Rediscovering the socialgroup: A self-categorization theory. New York: BasilBlackwell.

Van Hook, J., & Fix, M. (2000). A profile of theimmigrant student population. In J. Ruiz deVelasco, M. Fix, & B. Chu Clewell (Eds.), Overlooked and underserved: Immigrant students inU.S. secondary schools. Washington, DC: UrbanInstitute.

Weiland, A., & Coughlin, R. (1979). Self-identificationand preferences: A comparison of White andMexican-American first- and third-graders. Journalof Cross-Cultural Psychology, 10, 356–365.

Williams, J., & Morland, J. (1976). Race, color, and theyoung child. Chapel Hill, NC: University of NorthCarolina Press.

Willig, A. C. (1985). A meta-analysis of selectedstudies on the effectiveness of bilingual education.Review of Educational Research, 55, 269–317.

Wong-Fillmore, L., & Valdez, C. (1986). Teachingbilingual learners. In M. C. Wittrock (Ed.),Handbook of research on teaching (3rd ed.,pp. 648–685) New York: Macmillan.

Wright, S. C., Aron, A., McLaughlin-Volpe, T., &Ropp, S. A. (1997). The extended contact effect:Knowledge of cross-group friendships andprejudice. Journal of Personality and Social Psychology,73, 73–90.

Wright, S. C., Aron, A., & Tropp, L. R. (2002).Including others (and groups) in the self: Self-expansion and intergroup relations. In J. P. Forgas & K. D. Williams (Eds.), The social self:Cognitive, interpersonal and intergroup perspectives(pp. 343–363). Philadelphia: Psychology Press.

Wright, S. C., Brody, S. A., & Aron, A. (2005).Intergroup contact: Still our best hope forimproving intergroup relations. In C. Crandall &M. Schaller (Eds.), Social psychology of prejudice:Historical perspectives (pp. 115–142). Seattle, WA;Lewinian Press.

Wright, S. C., & Taylor, D. M. (1995). Identity andlanguage of the classroom: Investigating theimpact of heritage versus second languageinstruction on personal and collective self-esteem.Journal of Educational Psychology, 87, 241–252.

Wright, S. C., Taylor, D. M., & Macarthur J. (2000).Subtractive bilingualism and the survival of theInuit language: Heritage versus second languageeducation. Journal of Educational Psychology, 92,63–84.

Wright, S. C., & Van Der Zande, C. C. (1999). Bicultural friends: When cross-group friendships causeimproved intergroup attitudes. Paper presented at theSociety for Experimental Social Psychology, St.Louis, MO.

Paper received 14 February 2004; revised version accepted17 January 2005.

Biographical notesSTEPHEN C. WRIGHT is currently serving as associate

professor of psychology and Canada ResearchChair in Social Psychology at Simon FraserUniversity. He is also associate professor ofpsychology at the University of California, SantaCruz. He received his PhD from McGill Universityin 1991. His research interests focus on intergrouprelations, collective action, prejudice, prejudicereduction, and issues of minority language and


327


culture. He has served as associate editor ofPersonality and Social Psychology Bulletin, and on theeditorial boards of numerous scholarly journals.His work has been published in numerous bookchapters and in articles in major social,educational, and cross-cultural psychologyjournals, and he is the co-author of SocialPsychology in Cross-Cultural Perspective.

LINDA R. TROPP received her BA from WellesleyCollege and her PhD in social psychology fromthe University of California, Santa Cruz. She iscurrently an assistant professor of psychology at

Boston College and a member of the GoverningCouncil of the Society for the Psychological Studyof Social Issues. Her main research programsconcern experiences with intergroup contactamong minority and majority status groups,membership in and identification with socialgroups, and responses to prejudice anddisadvantage among socially devalued groups. Her work has been published in numerous journalarticles and book chapters, and she received the2003 Gordon W. Allport Intergroup RelationsPrize for her research on intergroup contact (in collaboration with Thomas F. Pettigrew).


328


language and intergroup contact: investigating the impact

Documents