Majority Children’s Evaluation of Acculturation Preferences of Immigrantand Emigrant Peers

Maykel Verkuyten, Jochem Thijs, and Jellie SierksmaUtrecht University

Using an experimental design, native majority group children (8–13 years, N = 842) evaluated acculturationstrategies (assimilation, integration, and separation) adopted by immigrant and emigrant peers. There weremedium to large effects of the perceived acculturation strategies on children’s peer evaluations. Overall,assimilation was valued most, followed by integration and separation. These effects were in part mediated byperceived national belonging. In addition, the effects were stronger for lower status compared to higher statusimmigrant groups, and for children with higher compared to lower national identification. For emigrants, sep-aration was valued most, followed by integration and assimilation. This indicates that the intergroup pro-cesses rather than migration per se are important for children’s acculturation perceptions and evaluations.

Of course they have to adapt, you always haveto do that, when we go and live in another coun-try we also have to adapt and do like they do.

This extract is from a focus group discussionamong young native adolescents in the Nether-lands. They are talking about whether immigrantshave to adapt to the Dutch culture or should beallowed to maintain their cultural heritage. In theextract, adaptation is presented as an obligation forall migrants, including Dutch emigrants.

Although there is a large literature on majoritychildren’s ethnic and racial prejudices (see Levy &Killen, 2008), very little is known about children’sunderstanding of migrants (e.g., Gieling, Thijs, &Verkuyten, 2011; Malti, Killen, & Gasser, 2012;Pfeifer et al., 2007), and their evaluation ofimmigrants’ acculturation strategies in particular(e.g., Nigbur et al., 2008). Issues of migration andimmigrants are increasingly relevant for majoritychildren and their perception of immigrant’s accul-turation strategies might have consequences fortheir attitude toward immigrants.

We examined how native Dutch children(8–13 years) evaluate migrant peers with differentacculturation strategies. Using vignettes and anexperimental questionnaire design, we consideredchildren’s evaluation of two different immigrantgroups that have a different social status in the

Netherlands, as well as the evaluation of Dutchemigrants. Our aim was to investigate whethernative children’s evaluations of immigrant peersdepend on the acculturation strategy of these peersin combination with immigrant’s group status andchildren’s national identification. In addition, wewant to examine whether the perception of immi-grants’ belonging to the host nation mediates theeffect of perceived acculturation strategy on peerevaluation. In short, this research breaks newground in examining the effects of perceived accul-turation strategies on the evaluation of peers fromtwo different immigrant groups, and in investigat-ing why such effects occur by considering the medi-ating role of perceived host national belonging. Inaddition, by making a comparison with Dutch emi-grants we examine whether the evaluations ofimmigrant peers are related to their migration back-ground per se—as suggested in the quote above—or rather to their ethnic out-group status.

Acculturation Strategies

Berry’s (1997) well-known acculturation modeldistinguishes between four acculturation strategies.When immigrants want to maintain their cultureand also adopt the host culture, a preference forintegration exists. A strategy of assimilation implies a

Correspondence concerning this article should be addressed toMaykel Verkuyten, Ercomer Department, Utrecht University,Padualaan 14, 3584 CH, Utrecht, The Netherlands. Electronicmail may be sent to m.verkuyten@uu.nl.

© 2013 The AuthorsChild Development © 2013 Society for Research in Child Development, Inc.All rights reserved. 0009-3920/2013/xxxx-xxxxDOI: 10.1111/cdev.12111

Child Development, xxxx 2013, Volume 00, Number 0, Pages 1–16

preference for host culture adoption and abandon-ing the original culture. Separation indicates a desireto maintain the heritage culture without adoptingthe host culture. Finally, when immigrants refuseboth cultural maintenance and cultural adoption,marginalization results.

From a dynamic intergroup perspective (Brown& Zagefka, 2011), it can be argued that the waynative children react to and evaluate immigrantpeers depends on the acculturation of those peers.More specifically, immigrant peers who preferassimilation or integration can be expected to beevaluated more positively than separating peers.One reason for this is that assimilation and integra-tion involve the adoption of the host culture,whereas separation does not. The former two strate-gies indicate that the host culture is valued to theextent that immigrants want to adopt it. Whenimmigrants value the host culture and identity,native children feel valued by them and this canlead to a more favorable attitude toward immigrantpeers (Zagefka, Tip, Gonzalez, Brown, & Cinnirella,2012). In general, research has shown that whenanother person is perceived to have a positive atti-tude toward the social self this results in more posi-tive attitudes toward this person (Curtis & Miller,1986).

In addition, the adoption of the host cultureimplies increased similarity between native andimmigrant peers, and perception of similarity is onefactor that leads to a sense of shared identity(Dovidio, Gaertner, & Saguy, 2007). Thus,immigrant peers who adopt the host culture are rel-atively more similar and therefore more easilyincluded in the national category. According to thecommon in-group identity model, inclusion in ashared category improves attitudes towardout-group members through the general tendencyto favor members who belong to the same categoryas oneself (Gaertner & Dovidio, 2000). This has alsobeen found among children and adolescents(e.g., Gaertner et al., 2008; Pfeifer et al., 2007). Thisreasoning implies that we can expect that perceivednational belonging mediates the relation betweenacculturation strategy and peer evaluation.Immigrant peers who adopt the host culture will beperceived more strongly as Dutch nationals andthis perception, in turn, leads to a more positiveevaluation of these peers. The current study willtest this mediation hypothesis that shed light onwhy different acculturation strategies might affectthe evaluation of immigrant peers differently.

In addition, previous research in Europe hasfound that majority members prefer assimilation

and then integration of immigrants, rather thanseparation or marginalization (Van Oudenhoven,Prins, & Buunk, 1998; Verkuyten, 2005). Thus, itcan be expected that assimilating immigrant peerswho adopt the host culture without wanting tomaintain their minority culture are evaluated morepositively than integrating peers. Assimilationimplies that immigrants shed their previous markersof group identity and adopt those of the host society(Hartmann & Gerteis, 2005). This means that in com-parison to integration, they can be perceived as moreexclusively valuing the host culture and as becomingmore similar to “us” and thereby “one of us.”

Intergroup Context

Studies in Canada have shown that natives pre-fer integration strategies for valued immigrants andseparation or assimilation for nonvalued immi-grants (Montreuil & Bourhis, 2001). Research inFrance (Maisonneuve & Teste, 2007) and in theNetherlands (Van Oudenhoven et al., 1998) foundno differences in the evaluation of acculturationstrategies of different immigrant groups. However,this research did not examine children and did notfocus on groups that clearly differ in status posi-tion. For example, Van Oudenhoven et al. (1998)examined the evaluation of Moroccan and TurkishMuslim immigrants who are evaluated quite simi-larly and negatively in the Netherlands, also by(early) adolescents (Hagendoorn, 1995; Verkuyten& Kinket, 2000). In the current study, we focus onimmigrants of Turkish and Chinese origin thatclearly differ in status. The Turks are the largestminority group in the Netherlands and have one ofthe worst socioeconomic positions. The great major-ity of them are Muslim and Turkish children facerelatively high levels of peer discrimination(Verkuyten & Thijs, 2002). In contrast, the Chineseare a relatively small group with a much bettersocioeconomic position (Gijsberts, Huijnk, & Vogels,2011). They are typically not discussed in the strongand rather negative Dutch integration debate thatfocuses on the alleged threats that Islam and Mus-lims pose to the Dutch identity and culture.Research among Dutch native (early) adolescentshas found relatively high levels of perceived iden-tity threats by Muslim immigrants (VelascoGonz�alez, Verkuyten, Weesie, & Poppe, 2008).Given this societal context and the importance ofstatus differences and out-group threats for chil-dren’s group attitudes (e.g., Bigler, Spears Brown,& Markell, 2001; Nesdale, Maass, Durkin, &Griffiths, 2005), we assumed that compared to

2 Verkuyten, Thijs, and Sierksma

Chinese peers, the perception that Turkish peersmaintain their culture poses an element of identitythreat for the native Dutch. This would mean thatexclusive Dutch culture adoption is consideredmore important for the evaluation of Turkish com-pared to Chinese peers. Thus, the difference in eval-uation between immigrant peers who assimilateversus those who prefer integration or separationwas expected to be stronger for Turkish comparedto Chinese peers.

However, it is likely that concerns about Dutchculture and identity depend on the level of groupidentification. According to social identity theory,highly identified group members are motivated tothink and behave as in-group members becausethey view their group as a reflection of the self(Tajfel & Turner, 1979). For example, compared tolower identifiers, children with higher in-groupidentification are more likely to be concerned aboutout-group threats and the continuity and value oftheir group (e.g., Nesdale et al., 2005; Pfeifer et al.,2007). This means that especially higher identifierswill respond to the acculturation strategies ofthreatening out-groups. Thus, higher Dutch identifi-cation can be expected to be associated with a morenegative evaluation of Turkish immigrant peers butless so with the evaluation of nonthreatening Chi-nese peers. Furthermore, higher compared to loweridentifiers should evaluate the assimilating Turkishpeer more positively compared to integrating orseparating Turkish peers.

Emigrants

In the quote at the beginning of this article, itis argued that immigrants have to adopt the hostculture because that is what migrants are supposedto do, including native Dutch who emigrate fromthe Netherlands. This suggests that the evaluationof minority groups is related to their migrationbackground and not to the fact that they are con-sidered an ethnic out-group. Rather than intergroupprocesses, a perceived general obligation andresponsibility to adopt the host culture mightunderlie native children’s evaluation of immigrants(Gieling et al., 2011). One way to examine this pos-sibility is by making a comparison between immi-grants and emigrants. We focused on immigrantscoming to the Netherlands and on Dutch emigrantpeers that have left the country. A similar effect ofperceived acculturation strategies on peer evalua-tions of immigrants and emigrants would providesupport for the proposition that the acculturationstrategy of migrants per se has an impact on peer

evaluations. In contrast, from an intergroup per-spective, it can be expected that assimilating immi-grants are evaluated more positively than otherimmigrants, whereas assimilating emigrants areevaluated more negatively than those who maintaintheir heritage culture. Not maintaining Dutch cul-ture suggests that this culture is less valued andcan also be considered in-group threatening.Research on the so-called black sheep effect (Mar-ques & P�aez, 1994) and the developmental subjec-tive group dynamics model (Abrams & Rutland,2008) has shown that in-group peers who are dis-loyal or do not support the continuation of the in-group are evaluated rather negatively. More nega-tive evaluation is especially likely for higher in-group identifiers. The more children identify withtheir in-group, the more likely they are to be con-cerned about disloyal in-group members (Abrams,Rutland, & Cameron, 2003). Thus, higher comparedto lower national identifiers might sympathize morewith Dutch emigrants who remain loyal to theDutch culture and do not assimilate. Therefore,higher compared to lower identifiers can beexpected to be more positive about emigrant peerswho maintain the Dutch culture without adoptingthe host culture (separation).

Age-Related Changes

Developmental intergroup research has shownthat with age a more complex understanding ofgroups and group relations develops (Killen &Rutland, 2011). Children increasingly understandthe importance of group functioning, group status,group commitment, and the maintenance andstability of the in-group. On one hand, this canmean that immigrant peers wanting to maintaintheir culture are evaluated more negatively by older(11–12) than younger (8–9) native children.Research has demonstrated that (early) adolescentsconsider group-specific norms and the preservationof these norms more important than children(Killen, Rutland, Abrams, Mulvey, & Hitti, 2012).Furthermore, research on social cliques has demon-strated that the importance of group status becomesstronger around 13–14 years and that group func-tioning becomes an important reason to justifysocial exclusion (Horn, 2003, 2006). This increasedfocus on status and group identity might extend tothe domain of nationality and immigration. In thecontext of Switzerland (Malti et al., 2012), it hasbeen found that middle compared to young adoles-cents judge exclusion based on nationality morelegitimate and attribute more positive emotions to

Acculturation Preferences and Peer Evaluation 3

excluder peers. And, with age, native Dutch adoles-cents tend to endorse assimilation of immigrantsmore strongly because they are more concernedabout the functioning of society and maintenance ofDutch culture (Gieling et al., 2011). This leads tothe expectation that older compared to youngerchildren evaluate integrating immigrant peers morenegatively and assimilating peers more positively.

On the other hand, children between the ages of8 and 12 increasingly develop the ability to under-stand different social perspectives and how thesediffer from their own (Abrams, Rutland, Pelletier, &Ferrell, 2009). Furthermore, older children’s under-standing of the importance of group identity andautonomy might imply that they recognize why animmigrant peer would want to hold on to his orher own culture. This could mean that older com-pared to younger children are more positive towardintegrating immigrant peers and more negativetoward assimilating peers.

In Summary

To our knowledge, this is the first study thatexamines native majority children’s evaluation ofacculturation strategies of immigrant peers. Using arandomized experimental design, we first expectedthat assimilating peers would be evaluated mostpositively, followed by integrating peers, and thatpeers who do not adopt the host culture (separa-tion) would be evaluated least positively. Further-more, we expected these effects of acculturationstrategies on peer evaluations to be mediated byperceptions of host national belonging. Followingthe common in-group identity model (Gaertner &Dovidio, 2000), assimilating peers were expected tobe evaluated more positively because they areconsidered more strongly as national in-groupmembers. These predictions were examined inrelation to two immigrant groups that clearly differin status position in the Netherlands. Thehypothesized positive effect of assimilation on peerevaluation was expected to be stronger for thelower status group peers (Turkish origin) comparedto the higher status group peers (Chinese origin), inparticular for children with relatively strongnational identification. Importantly, by making acomparison with Dutch emigrant peers, we alsoexamined whether migration per se or rather theintergroup context underlies children’s peer evalua-tions.

For testing these hypotheses and following otherstudies (Maisonneuve & Teste, 2007; Van Acker &Vanbeselaere, 2012; Van Oudenhoven et al., 1998;

Zagefka et al., 2012), we used vignettes presentingpeers as adopting a separation, integration, orassimilation acculturation strategy. Compared toacculturation scales, the vignette method has theadvantage of being more realistic and understand-able for children and less sensitive to social desir-able responding (Maisonneuve & Teste, 2007). Thefourth strategy of marginalization was not studiedbecause in Europe this strategy is rare and not veryrealistic for children (Verkuyten, 2005; Zagefka &Brown, 2002).

Method

Participants and Procedure

Participants came from 22 regular, middle-classelementary schools in different parts of the Nether-lands. All children who identified both themselvesand their parents as native Dutch (N = 842) and whohad no missing values on the outcome measures ofour experimental manipulation (96.8% of the nativesample) were considered in the present analysis.Most of their classmates self-identified as nativeDutch as well (M%Dutch = 80, SD = 0.22), which indi-cates that most participants came from relativelyWhite schools. The children had a mean age of10.73 years (SD = 0.99, range = 8–13, Mdn = 11) and53.7% (N = 452) of them were girls. Participation inthe study was voluntary and parental permissionwas obtained. The children anonymously filled out aquestionnaire, which was randomly divided in eachclassroom.

Design and Measures

Acculturation judgments. The participants werepresented with three vignettes that portrayed,respectively, a Turkish and a Chinese girl living inthe Netherlands, and a native Dutch girl living inTurkey. We focused on emigration to Turkeybecause we wanted to have a similar culturaldistance for the low-status Turkish immigrant peerand the Dutch emigrant peer. Similar to otherresearch among children (e.g., Malti et al., 2012;Verkuyten & Slooter, 2008), and because the judg-ments for the vignettes were expected to be rela-tively independent, the vignette order was heldconstant with the Dutch emigrant peer presentedlast and the Turkish peer presented first. The Dutchpeer was presented last to make the intergroupcomparative context salient when evaluating thispeer. This is important because participants tend toundertake within-group comparisons, rather than

4 Verkuyten, Thijs, and Sierksma

between-group comparisons, if their in-group ispresented first without explicit reference to an out-group (Oakes, Haslam, & Turner, 1994).

Due to practical constraints we could notconsider all variations in the design and werenot able to systematically vary the gender of thepeers in the vignettes or to match the gender ofthe peers with the gender of the participants. There-fore, we decided to hold gender constant and tofocus on female peers only. This means that wepresent a more conservative test because there aremore negative stereotypes about exclusion regard-ing boys than girls (Schneider, 2004), and in generalgirls are evaluated more positively than boys,which has also been found for Dutch adolescents’evaluations of Muslim immigrant peers (Poppe &Verkuyten, 2012). The three (hypothetical) peershad one of three acculturation strategies that weremanipulated in a between-subjects design. Thus,each participating child judged for each of the threetarget peers similar strategies. Following previousresearch (Matera, Stefanile, & Brown, 2011; Van Ac-ker & Vanbeselaere, 2012), the peers were describedas favoring either integration (N = 300), separation(N = 275), or assimilation (N = 267). The vignetteswere developed on the basis of informal discussionswith four groups of native Dutch children (see alsoVerkuyten & Steenhuis, 2005). To give an example,for the Turkish peer all vignettes started with: “Fat-ma comes from Turkey and has been living in theNetherlands for 6 years already,” but then theassimilation vignette stated: “She has many Dutchfriends and she loves Dutch food and Dutch music.She wants to stay in the Netherlands for the rest ofher life. She doesn’t feel Turkish anymore. She pre-fers not to speak Turkish, and she definitely doesn’twant to wear a headscarf in the future.” The inte-gration vignette continued with “She has manyDutch friends and she loves Dutch food and Dutchmusic. She wants to stay in the Netherlands for therest of her life. But she also feels Turkish. She wantsto speak Turkish and wear a headscarf in thefuture.” The separation vignette continued with “Shehas no Dutch friends and she doesn’t like Dutchfood or Dutch music. She feels very much Turkishand she want to stay Turkish as much as possible.She always likes to speak Turkish, and she wantsto wear a headscarf in the future.”

Perceived host national belonging. Perceived hostnational belonging of the Turkish and Chinese peerwas assessed by the question “Do you think that[name] is Dutch?” For the Dutch emigrant peer thequestion was “Do you think that [name] isTurkish?” Responses were rated on a Likert-type

scale ranging from no, certainly not (which wascoded as �2) to yes, certainly (coded as +2).

Peer liking. Children’s peer liking was measuredby the question “What do you think of [name]?”which had a Likert-type response format consistingof seven faces. These faces ranged from very happy(big smile, which was coded as +3) to very sad (bigfrown, coded as �3), and there was a neutral mid-point (straight face, coded as 0). The “seven faces” for-mat (Yee & Brown, 1992) has been successfully usedto examine group attitudes among children and earlyadolescents (e.g., Verkuyten & Thijs, 2001).

National identification. Prior to assessing chil-dren’s acculturation judgments, we measured theirDutch national identification with four questions.These questions focused on children’s belonging tothe Netherlands as a country rather than on feelingDutch (Barrett, 2007). The reason is that Dutchimmigration debates are typically framed in termsof what immigration means for the country andfunctioning of the nation. Furthermore, in theDutch language the name of the country and ofthe people are similar (Nederland and Nederlanders).The items were as follows: “Do you feel at home inthe Netherlands?” “Are you proud of the Nether-lands?” “Do you ever think, the Netherlands isreally my country?” and “Do you like it in theNetherlands?” Responses were rated on Likert-typescale ranging from 1 (no, certainly not) to 5 (yes,certainly). The four questions yielded a Cronbach’salpha of .77. To examine whether they loaded on asingle factor we conducted confirmatory factoranalysis. With the exception of the root meansquare error of approximation (RMSEA) value, thefit indices were sufficient for the one-factormodel we specified: v2(2) = 16.115, comparative fitindex (CFI) = .975, RMSEA = .120, standardizedroot mean square residual (SRMR) = .031 (see Hu& Bentler, 1999). Moreover, after a correlation wasallowed between error terms of “Are you proud ofthe Netherlands?” and “Do you ever think, theNetherlands are really my country?” model fit wassatisfactory according to Hu and Bentler’s (1999)criteria, that is, CFI > .95, RMSEA < .06, andSRMR < .08.

Data Analytic Strategy

To examine and compare children’s responses tothe different peers (Turkish, Chinese, and Dutch)we conducted multivariate multilevel regressionanalyses in MLWin (Rasbash, Browne, Healy, Cam-eron, & Charlton, 2004). In the multivariate multi-level model, different responses (within subjects)

Acculturation Preferences and Peer Evaluation 5

can be examined and compared by treating them asobservations nested within individuals (see Gold-stein, 1995; Snijders & Bosker, 1999). We used mul-tilevel modeling rather than more conventionalrepeated measures analysis (generalized linearmodel [GLM]) because it allowed us to directlyexamine the mediating role of perceived nationalbelonging. Perceived belonging was no simplewithin-subjects factor as it had both within-subjectand between-subjects variation (due to the experi-mental manipulation and the individual differencesvariables). Hence, its mediating role could not bedirectly examined by including it as a predictor inGLM.

In this study, we analyzed a two-level data struc-ture for each acculturation judgment (i.e., liking ofthe peers, and perceptions of their host nationalbelonging). Level 1 was the within-subject levelinvolving each participant’s responses toward thethree peers (n = 2526). Note that we specified a mul-tivariate rather than a univariate multilevel model.Hence, Level 1 was only included to define the mul-tivariate structure and there was no variation at thislevel (see Goldstein, 1995; Snijders & Bosker, 1999).Level 2 was the between-subject level involving theseparate means and variances of each of the threeresponse variables across participants (n = 842). Theethnicity of the target peers was examined as aLevel 1 variable and all other variables includingthe peers’ (manipulated) acculturation strategiesand children’s national identification were examinedas Level 2 variables. There was also a third levelbecause the participating children were nestedwithin their schools (n = 22), but we did not includethis level in our analyses. The proportions of Level 3variance (i.e., the intraclass correlations; see Snijders& Bosker, 1999) were very low (0.0%–3.7%), indicat-ing relatively small differences between schools.Moreover, preliminary analyses showed that thesedifferences could not be explained by ethnic schoolcomposition (percentage of native Dutch students),and that controlling for them yielded similar resultsas the simpler two-level models. Hence, the latterare presented in this study.

All models were estimated using the Iterativegeneralized least squares algorithm, and relativemodel improvement was assessed by comparingthe fit (deviance) of nested models. Differencesbetween these statistics follow a chi-square distribu-tion, and degrees of freedom are given by thedifferences in numbers of parameters (Snijders &Bosker, 1999). In the multivariate model, one cantest whether the effects of a predictor are statisti-cally different for two or more dependent variables

by comparing the fit of models with commonregression coefficients versus separate regressioncoefficients.

Differences in responses to the acculturationvignettes (between subjects) were examined withdummy variables. In our analyses, we first used adummy that equals 1 for assimilation, and 0 for theother strategies, and a dummy that equals 1 forintegration, and 0 for the other strategies using theseparation strategy as a reference category. Whenincluded together, these dummies compare, respec-tively, the assimilation vignettes and the integrationvignettes to the separation vignettes. In addition,we also changed the reference category to examinedifferences between the assimilation and the inte-gration strategies. This is common practice indummy regression analysis as the choice of thereference group is arbitrary (Fox, 1997).

Results

Descriptive Findings

Table 1 provides an overview of the means forperceived national belonging and liking both withinand across the three acculturation conditions. Toexamine those means and to test our hypotheses,we conducted four sets of multilevel analyses, firstfor liking as the dependent variable and then forperceived host national belonging as the expectedmediator. Subsequently, we tested whether hostnational belonging is indeed a mediator betweenperceived acculturation strategies and liking.

Liking

In Step 1, we tested a so-called intercept-onlymodel to examine the average liking of the Turkish,Chinese, and Dutch peers ignoring the between-subjects experimental manipulation. This modelyields the exact overall mean score shown in thebottom row of Table 1, together with their vari-ances and covariances. In addition, the model yieldsa deviance statistic (loglike), which was 9427.040.This statistic was used to evaluate the relative fit ofthe subsequent models. The mean ratings for likingwere significantly above the midpoint of the scale,ps < .01. Thus, overall, the peers were evaluatedpositively. In addition, there were significant differ-ences between the three target peers. The childrenliked the Dutch peer and the Chinese peer morethan the Turkish peer, respectively, p < .01 andp < .05, and they did not make an evaluative dis-tinction between the former two.

6 Verkuyten, Thijs, and Sierksma

In Step 2, we entered the two dummy variablesfor the acculturation strategies of the peers(Table 2). This led to a significant improvement inmodel fit, v2(6) = 278.72, p < .01. As expected, theassimilating immigrant peers were evaluated mostpositively, followed by the integrating peers andthen the separating ones. The effect size of assimila-tion versus separation was medium to large for thetwo immigrant peers, g2

partial ranging from 0.12 to0.15 (see Cohen, 1988), but, as expected, the effectwas significantly stronger for the Turkish versusthe Chinese peer, v2(1) = 6.97, p < .01. Moreover,the effect size of the integration strategy was betweensmall and medium for the Turkish peer, g2 = 0.04,and medium for the Chinese peer, g2

partial = 0.09,and this difference was significant, v2(1) = 6.95,p < .01. In addition, further analysis showed thatthere was a medium-sized positive effect of assimi-lation versus integration on children’s liking of theTurkish peer, b = 0.96, p < .01, g2

partial = 0.05, butthis effect was lower, v2(1) = 27.86, p < .01,and nonsignificant for the evaluation of theChinese peer, b = 0.23, p > .05, g2

partial = .00. Taken

together, and as shown in Figure 1, these findingssupport our prediction that the lower status immi-grant peer (Turkish) is liked most if she adopts thehost national culture without maintaining her heri-tage culture (assimilation). The liking of the higherstatus immigrant peer (Chinese) depends on heradoption of the Dutch host culture with or without

Table 1Means for Perceived National Belonging and Liking of Targets

Perceived national belonging Liking

Turkish M (SD) Chinese M (SD) Dutch M (SD) Turkish M (SD) Chinese M (SD) Dutch M (SD)

ConditionAssimilation 0.48 (1.13) 0.56 (1.14) 0.15 (1.24) 1.90 (1.37) 1.76 (1.27) 0.79 (1.68)Integration �0.23 (1.14) 0.08 (1.07) �0.19 (1.10) 0.94 (1.80) 1.53 (1.51) 1.11 (1.54)Separation �1.10 (1.02) �0.91 (1.02) �1.12 (0.97) 0.15 (1.88) 0.39 (1.64) 1.53 (1.45)

Overall �0.22 (1.28) �0.09 (1.24) �0.39 (1.23) 0.98 (1.84) 1.23 (1.60) 1.15 (1.59)

Note. Perceived national belonging was measured on a scale from �2 to +2. Evaluation was measured on a scale from �3 to +3.

Table 2Effects of Acculturation Strategies on Perceived National Belonging and Liking

Perceived national belonging Liking

Turkish Chinese Dutch Turkish Chinese Dutch

Constant (= separation) �1.10** �0.91** �1.12** 0.15 0.39** 1.54**Assimilation (vs. separation) 1.58**a 1.47**a,b 1.27**b 1.75**a 1.38**b �0.75**cIntegration (vs. separation) 1.08**a 0.99**a 0.93**a 0.79**a 1.15**b �0.43**c(Co)varianceTurkish peer 1.204 2.889Chinese peer 0.780 1.160 1.210 2.199Dutch peer 0.427 0.429 1.226 0.810 0.621 2.421Deviance 6,999.99 9,148.32

Note. The (co)variances indicate the unexplained variation and covariation of the three dependent variables in eachmodel. Effects in the samerow that have different subscripts are significantly different at p < .01. Assimilation and integration are dummy variables coded 1 or 0.**p < .01.

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Turkish Peer Chinese Peer Dutch Peer

Peer

Lik

ing

Assimilation

Integration

Separation

Figure 1. Effects of acculturation orientations on peer liking ofthe three target groups.

Acculturation Preferences and Peer Evaluation 7

maintaining her heritage culture (assimilation orintegration).

Results for the Dutch emigrants differed signifi-cantly from the findings for the immigrant peers,p < .01 (see Table 2). The Dutch peer was likedmost when she favored separation, and least likedwhen she favored assimilation (see Figure 1): Alldifferences were significant, p < .05, and betweensmall and medium, g2

partial ranging from 0.01 to0.04. Thus, the evaluations of the acculturationstrategies of the Dutch peers formed the mirrorimage of those of the immigrant peers. This sup-ports the importance of the intergroup context forchildren’s judgments rather than migration per se.

National identification, age, and gender. Next, weexamined whether the effects of the acculturationstrategies on peer liking were qualified by interac-tions with national identification, age, and genderof the participants. To enhance the interpretabilityof our results, interaction terms were computedafter national identification, age, and the dummyvariables were centered around their means. Forgender, we included a contrast (coded +0.5 for girlsand �0.5 for boys). Results are shown in Table 3.Adding the three main effects and the six interac-tions significantly improved the fit of the model,v2(27) = 73.65, p < .01.

The main and interaction effects for age were notsignificant (and therefore not included in Table 3).Gender only had a main effect with girls being over-all more positive than boys. National identification

had no main effect on the liking of the peers but itsinteraction with assimilation (vs. separation) wassignificant. To examine this significant interactionwe conducted simple slope analyses. We examinedthe difference between assimilation and separationfor children who identified strongly (1 SD > M) ver-sus weakly (1 SD < M) with the Netherlands. Asexpected, results showed that higher compared tolower identifiers liked the assimilating versus sepa-rating immigrant peer more, respectively, b = 2.06versus b = 1.40 for the Turkish peers, and b = 1.72versus b = 0.96 for the Chinese peers, all ps < .01.These effects are shown in Figure 2.

Table 3Moderating Effects of National Identification, Age, and Gender

Perceived national belonging Liking

Turkish Chinese Dutch Turkish Chinese Dutch

Assimilation (vs. separation) 1.53** 1.42** 1.23** 1.73** 1.34** �0.78**Integration (vs. separation) 1.07** 0.98** 0.92** 0.81** 1.16** �0.42**National identification 0.05 0.04 0.03 0.05 0.11 0.23***Assimilation 0.14 0.00 0.16 0.38** 0.43** 0.01*Integration 0.03 �0.01 �0.02 0.10 0.22 �0.34*

Gender 0.28** 0.21** 0.06 0.49** 0.27** 0.51***Assimilation 0.70** 0.82** 0.60** �0.12 0.26 0.14*Integration 0.60** 0.51** 0.18 0.18 0.21 0.42

(Co)varianceTurkish peer 1.157 2.799Chinese peer 0.742 1.119 1.158 2.152Dutch peer 0.406 0.409 1.207 0.731 0.571 2.289Deviance 6,944.61 9,074.67

Note. The (co)variances indicate the unexplained variation and covariation of the three dependent variables in each model. Nationalidentification was measured on 5-point scale and gender was included as contrast coded �0.5 for males and +0.5 for females. Theresults for age were not significant and therefore not shown.*p < .05. **p < .01.

0

0.5

1

1.5

2

2.5

LowIdentifiers:

Turkish Peer

LowIdentifiers:

Chinese Peer

High Identifiers:

Turkish Peer

High Identifiers:

Chinese Peer

Peer

Lik

ing

Assimilation

Separation

Figure 2. Interaction effects of national identification and accul-turation strategies on the liking of the Turkish and the Chineseimmigrant peers.

8 Verkuyten, Thijs, and Sierksma

For the Dutch emigrants there was a positivemain effect of national identification, which wasqualified by a significant interaction with integra-tion (vs. separation). Furthermore, the additionalanalysis showed that the interaction with assimila-tion versus integration was significant as well,b = 0.35, p < .05. Simple slope analyses showed thathigher identifiers but not lower identifiers liked theDutch emigrant peer more when she preferredseparation rather than integration, respectively,b = 0.71, p < .01, and b = 0.13, ns. By contrast,lower identifiers but not higher identifiers liked theDutch peer less when she preferred assimilationrather than integration, respectively, b = �0.66,p < .01, and b = �0.07, ns. This pattern of findingsis shown in Figure 3. It indicates that lower identifi-ers were less positive about the Dutch emigrantpeer when she rejected the Dutch culture (assimila-tion), whereas higher identifiers were more positivewhen she exclusively maintained Dutch culture andrejected the Turkish host culture (separation).

Perceived Host National Belonging

In Step 1, the intercept-only model yielded adeviance statistic (loglike) of 7,322.208 for perceivednational belonging. The mean ratings were signifi-cantly lower than the zero scale midpoint, p < .05for the Chinese peer and p < .01 for the other peers.Thus, overall, the peers were not strongly seen asbelonging to their host countries. In addition, therewere significant differences between the three targetpeers. Reflecting the social status positions and asshown in Table 1, the children reported highernational belonging for the Chinese peer (high-statusimmigrant), followed by the Turkish peer (low-status immigrant), and the Dutch peer (in-groupemigrant), all differences p < .01.

In Step 2, we entered the two dummy variablesfor the acculturation strategies of the peers. This ledto a significant improvement in model fit for bothnational belonging, v2(6) = 322.22, p < .01. As shownin Table 2, the effects of these dummies were similar.All target peers were regarded more strongly asbelonging to the host nation when they favoredassimilation rather than separation, g2

partial rangingfrom 0.18 to 0.25 (indicating large effects), and whenthey favored integration rather than separation,g2

partial ranging from 0.11 to 0.14 (indicating mediumto large effects). Additional analyses (with a differentreference category) showed that the assimilatingpeers were perceived to belong more strongly to thehost nation than the integrating peers, p < .01, g2

partial

ranging from 0.02 to 0.03. The effect of assimilationversus separation was stronger for the Turkishcompared to the Dutch peers, v2(1) = 7.91, p < .01,but all other effects were similar.

National identification, age, and gender. Addingnational identification, age, and gender and theirinteractions significantly improved the fit of themodel, v2(27) = 55.38, p < .01. There were no main orinteraction effects for national identification and forage on perceived national belonging of the target peer,but there were significant effects for gender(see Table 3). Compared to boys, girls were overallmore likely to consider the minority peers as conation-als but equally likely to regard the Dutch emigrantpeer as a Turkish national. However, there were sig-nificant interactions with the acculturation strategies.The effects of assimilation (vs. separation) and integra-tion (vs. separation) on perceived national belongingof the immigrant peers were stronger for girls versusboys. However, these effects were also significant forboys: for assimilation, b = 1.18, p < .001, and integra-tion, b = .77, p < .001, of the Turkish peer, and forassimilation, b = 1.01, p < .001, and integration,b = 0.72, p < .001, of the Chinese peer.

For the emigrant peer, the effect of assimilation(vs. separation) on perceived national belongingwas significantly stronger for girls compared toboys, and this was also found for the effect ofassimilation versus integration, b = 0.42, p < .05(not shown in Table 3). However, boys alsoregarded the emigrant peer more a Turkish nationalwhen she showed assimilation rather than separa-tion, b = 0.93, p < .01, but the difference betweenassimilation and separation was not significant.

Mediating Role of Host National Belonging

We tested our prediction that perceived nationalbelonging mediates the effect of acculturation strategy

0

0.20.4

0.60.8

11.2

1.41.6

1.82

Low Identifiers High Indentifiers

Peer

Lik

ing Assimilation

Integration

Separation

Figure 3. Interaction effects of national identification and accul-turation strategies on the liking of Dutch emigrant peers.

Acculturation Preferences and Peer Evaluation 9

on children’s peer liking. Note that we could onlytest this for the main effects of acculturation strategy,and not for its interactions with national identifica-tion because these interactions were not significantlyrelated to perceived national belonging.

In addition to a significant relation between theproposed mediator (perceived national belonging)and the independent variable (acculturation strat-egy), mediation requires that the effect of the inde-pendent variable is substantially reduced when themediator is included as an additional predictor inthe regression analysis (Baron & Kenny, 1986).When perceived national belonging of each tar-get peer was included in the prediction ofchildren’s peer liking, the model fit was signifi-cantly improved, v2(3) = 200.927, p < .01. Perceivednational belonging had a positive effect on thelikings of all target peers, but the effect was stron-ger for the immigrants than for the emigrant peer,b = 0.49 and b = 0.50 for, respectively, the Turkishand the Chinese peer, and b = 0.21 for the Dutchpeer, all p < .01. When perceived national belong-ing was added to the regression models, the effectsof the acculturation strategies were reduced for theTurkish peer, respectively, b = 0.98, p < .01, andb = 0.28, p < .05, for assimilation and integration(vs. separation), and, b = 0.70, p < .01, for assimila-tion versus integration. For the Chinese peer theeffects were smaller as well, respectively, b = 0.63and b = 0.67, both ps < .01, and b = �0.04, ns.Because all but one of those effects were stillsignificant, this indicates the possibility of partialmediation.

To examine whether these reductions were sub-stantial, we conducted Sobel tests for the indirecteffects of the acculturation strategies on the likingsof the immigrant peers through their perceivednational belonging (MacKinnon, Warsi, & Dwyer,1995). As expected, results indicated that perceivednational belonging carried a significant portion ofthe effects of assimilation (vs. separation; z = 8.98,p < .01) and integration (vs. separation; z = 7.97,p < .01) on the evaluation of the Turkish peer aswell as the Chinese peer, respectively, z = 9.57,p < .01, and z = 8.17, p < .01. Likewise, perceivednational belonging explained a significant part ofthe differential liking of the Turkish and Chinesepeers who preferred assimilation versus integration,respectively, z = 4.53, p < .01, and z = 4.54, p < .01.Together these analyses show that the effect ofimmigrants’ acculturation strategy on children’sevaluation is in part due to perceived host nationalbelonging. Thus, the acculturation strategy adoptedaffects the extent to which the immigrant peer is

considered a Dutch national and this in turn isrelated to the liking of the peer.

For the Dutch emigrant peer there was no evi-dence for perceived national belonging mediatingbetween the effect of acculturation strategy onliking. In fact, when perceived national belongingwas added to the model all effects of acculturationstrategy increased in strength: the main effects ofassimilation (vs. separation; b = �1.04, p < .001)and integration (vs. separation; b = �0.61, p < .001),and the main effect of assimilation versus integra-tion (b = �0.43, p < .01). Sobel tests showed thatthere were significant and positive indirect effectsof assimilation (vs. separation; z = 4.29, p < .001),and integration (vs. separation; z = 4.14, p < .001),and assimilation versus integration (z = 2.68.p < .01) through perceived national belonging. Thismeans that there was suppression rather than medi-ation. Apparently there are two opposing pathwaysby which the acculturation strategies influence chil-dren’s liking of the Dutch peer. On one hand, andconsistent with the intergroup context, childrenliked this peer more when she maintained herDutch culture (separation and integration vs. assim-ilation) and when she did not adopt the Turkishhost culture (separation vs. integration and assimi-lation). However, under those conditions she wasalso regarded as being less Turkish, and similar tothe immigrant peers, perceived host nationalbelonging was in itself positively associated withchildren’s liking.

Discussion

For the first time, this study showed a clear andstrong causal effect of acculturation strategies ofimmigrant peers on native children’s evaluation ofthese peers. Medium to large effects (Cohen, 1988)of the acculturation strategies on peer evaluationswere found, indicating that the native majority chil-dren responded quite strongly to the different waysin which immigrants adapt to the host society. Inaddition, the findings show that the effect of theacculturation strategies on peer liking (a) in partdepends on perceived host national belonging, (b)differs between types of immigrant groups, (c) isrelated to the intergroup context rather than tomigration per se, and (d) differs for higher andlower national identifiers.

Overall, the children valued adoption of theDutch culture by immigrant peers whereas heritageculture maintenance was valued less. Assimilatingimmigrant peers were liked most followed by

10 Verkuyten, Thijs, and Sierksma

integrating peers and then separating peers. Thispattern of liking was clear and significant for thelow-status group of Turkish immigrants. For therelatively higher status Chinese, separation was alsodisliked most but there was no significant differ-ence between the liking of the assimilating and inte-grating peer.

These findings indicate that majority children aremore positively inclined toward immigrant peerswhen they feel that these peers value the host cul-ture to the extent that they want to adopt it. Inaddition, peers who adopt the host culture becomemore similar to the native majority and similaritycontributes to a sense of shared identity (Dovidioet al., 2007). According to the common in-groupidentity model, a shared identity improves attitudestoward former out-group members (Gaertner &Dovidio, 2000). Our findings show that the accul-turation strategies affected perceptions of hostnational belonging and thereby the evaluation ofthe acculturating peers. Thus, perceived hostnational belonging was a mediating processbetween acculturation preference and the evalua-tion of immigrant peers. Importantly, this mediat-ing role of perceived national belonging was foundfor both the lower (Turks) and the higher (Chinese)status immigrant group. This shows that a sharedidentity improves the attitude toward quite differ-ent immigrant groups. However, the effect was notfully mediated by host national belonging whichindicates that there are also other processesinvolved in this relation. One possibility that couldbe examined in future studies is that cultural simi-larity as such increases liking (Byrne, 1971) withoutnecessarily enhancing perceived host nationalbelonging.

By focusing on immigrant peers of a lower status(Turkish) and a higher status (Chinese) group, wewere able to show that the effect of acculturationpreferences on peer liking was stronger for theformer compared to the latter group. In general, theTurkish peers were evaluated more negatively thanChinese peers. In addition, the assimilating Turkishpeer was liked more than the integrating and sepa-rating ones. This shows that the lower status groupwas liked most when the heritage culture was notmaintained. One reason for this might be that ado-lescents have been found to consider cultural main-tenance by Muslim Turks as threatening to Dutchidentity and culture (Velasco Gonz�alez et al., 2008).For the Chinese peer, the evaluation of assimilationand integration was similar and more positive thanof separation. Thus, for this group the focus was onthe adoption of the Dutch culture, and whether this

was combined with heritage cultural maintenancedid not seem to matter for the peer liking. It shouldbe noted that the differences in liking of the Turkishand Chinese peer might also be related to the factthat the children were first asked about the leastliked Turkish group and then the Chinese. Thismight have led to a contrast effect in which theChinese peer was judged more positively. How-ever, it might also have created a negative responsepattern across the vignettes. More important, theorder of the groups cannot explain why particularlythe integrating Chinese peer was evaluated morepositively than the integrating Turkish peer,whereas the differences between the assimilatingand separating Turkish and Chinese peers werelow. Furthermore, the finding that overall theChinese peers were liked more and were morestrongly considered Dutch nationals than the Turk-ish peers is in agreement with their different socialstatus position and with previous findings (Hagen-doorn, 1995; Verkuyten & Kinket, 2000). So, it isnot very likely that the fixed order in which the tar-get immigrant groups were presented did affect thefindings for the acculturation evaluations.

The fact that the intergroup context plays animportant role in the evaluation of acculturatingpeers is further demonstrated by the findings forthe liking of the Dutch emigrants. The existingresearch on majority member’s evaluation of immi-grants has not considered emigrants (but seeGieling et al., 2011). This is unfortunate because thismakes it impossible to test whether migration perse or rather the intergroup context is important.The quote heading this article suggests that accul-turation strategies are evaluated in the light of themigration process and that cultural adaptation isexpected of both immigrants and emigrants. How-ever, our findings show that this is not the case atall. In fact, the separating Dutch emigrant peer wasliked most. Thus, the one who maintained Dutchculture without adopting the host Turkish culturewas liked most, followed by the peer who main-tained Dutch culture while also adopting Turkishculture (integration), and the assimilating peer wasliked least. The latter peer was considered moreTurkish than the former ones, which suggests thatthe latter is viewed as relatively less similar to one-self and therefore evaluated more negatively. Thus,if children feel that emigrants want to maintaintheir heritage culture they will be positivelyinclined toward them, whereas assimilation towardthe new culture implies lower similarity and thatthe Dutch culture is not strongly valued and sup-ported (Abrams & Rutland, 2008; Marques & P�aez,

Acculturation Preferences and Peer Evaluation 11

1994). This pattern of findings clearly shows that in-tergroup considerations are important for the evalu-ation of acculturating preferences of migrants. Thisgoes against the common rhetoric about immi-grants’ having to assimilate because as a principlethe rule “when in Rome do as the Romans do”should apply to all migrants, including in-groupemigrants (“we would do the same”). It should benoted that the children were not insensitive to thisrule when judging the Dutch peer. There was asuppression effect for the liking of the Dutchemigrant, which suggests that children did appreci-ate it when this peer could be regarded as a Turkishnational. However, this effect could not counter thedirect effects of the acculturation strategies, andoverall our findings show that there were substan-tial differences in the way in which children evalu-ated immigrants and emigrants with similaracculturation strategies (see also Gieling et al.,2011).

A last set of findings indicating the importantrole of the intergroup context relates to nationalidentification. National identification did not mod-erate the relation between acculturation preferencesand perceived host national belonging, but highercompared to lower identifiers were more positivetoward Turkish and Chinese peers who assimi-lated to Dutch culture than toward peers who pre-ferred separation. Immigrant peers who desireheritage culture maintenance without host cultureadoption can be assumed to pose an element ofidentity threat for majority members, but notequally to all children. Particularly, higher identify-ing children are more concerned about identitythreats (Nesdale et al., 2005; Pfeifer et al., 2007). Inaddition, higher identifiers probably will feel morevalued by immigrants who fully want to assimi-late to the host culture. Furthermore, higher identi-fiers will feel more valued by Dutch emigrantswho want to maintain their Dutch culture. Thefindings show that higher identifiers were muchmore positive about emigrants who preferred sep-aration from Turkish culture compared to integra-tion and assimilation. Thus, the exclusivemaintenance of Dutch culture was favored and theadoption of Turkish culture, either in combinationwith Dutch culture or not, was liked less. Loweridentifiers were particularly negative about theemigrant peer who preferred assimilation andequally positive about integration and separation.This indicates that Dutch cultural maintenance isrelatively important for lower identifiers, but forthem it can go together with the adoption of theTurkish host culture.

Limitations and Directions for Future Studies

Some limitations of the present research shouldbe discussed. For example, we used single-itemmeasures that have been used successfully in previ-ous research (e.g., Verkuyten & Thijs, 2001; Yee &Brown, 1992) but that could be improved. In addi-tion, the vignettes were developed in discussionswith children and tried to give a realistic picture ofan acculturating peer but this meant that theycontained various types of information. The empha-sis was on heritage cultural maintenance and hostculture adoption but there was also information onself-feelings and social contacts. Although researchsuggests that it matters little for resulting inter-group attitudes whether the focus is, for example,on perceived social contacts or cultural adoption(Tip et al., 2012; Van Acker & Vanbeselaere, 2011),future studies could examine whether the type ofinformation matters for native children’s judg-ments.

Furthermore, it would have been preferable toextend the experimental research by systematicallyvarying the gender of the peers presented in thevignettes. We could only use female names and thismight imply a conservative test case because thereare more stereotypes about exclusion regardingmales than females (Schneider, 2004) and Dutch ado-lescents tend to evaluate female immigrants morepositively than male immigrants (Poppe & Verkuy-ten, 2012). Hence, it is possible that the effects wouldhave been even stronger for male targets.

However, the use of only female names impliesa gender match between female participants andpeers and this might explain why girls were morepositive about the acculturating peers, why theyregarded the immigrant peers as more Dutch, andwhy (some of) the effects of acculturation strategyon perceived national belonging were stronger forgirls (but also significant for boys). More specifi-cally, it might be argued that some of our findingswere due to cross-categorization effects becausewhereas girls shared one or two characteristics withthe peers (either gender, or gender and ethnicity),boys could share only one characteristic (ethnicity;see Crisp & Hewstone, 2007). However, thereare two reasons to assume that the role of cross-categorization was limited. First, research amongchildren (using ethnicity and gender as twocharacteristics) has not found clear evidence forcrossed-categorization effects in which sharing onecharacteristic leads to more positive evaluations thansharing none of the two characteristics (Verkuyten,Weesie, & Eijberts, 2011). Second, the effects of the

12 Verkuyten, Thijs, and Sierksma

acculturation strategies on participants’ liking ofout-group and in-group peers were unrelated togender. The gender differences found might havemore to do with girls tending to be less concernedwith ethnic groups and ethnic group differences(Verkuyten & Thijs, 2001). This interpretation issupported by our findings that girls indicated lowernational identification than boys and that they morestrongly considered the Turkish and Chinese peersas being Dutch. Despite these differences the pat-tern of findings for the effect of acculturation strate-gies on host national belonging and peer liking wassimilar for girls and boys.

We focused on older children and young adoles-cents (8–13 years) and no age differences werefound. Thus, early adolescents did not evaluate theacculturating peers more positively than the olderchildren, and the differences in liking of the threeacculturation strategies were independent of age.The restricted age range might be one reason forthis and future studies could examine a wider agerange. However, a recent meta-analysis of agedifferences in ethnic and racial prejudice found onlya very slight decrease in prejudice in late childhood(8–10 years) and no general trend after 10 years ofage (Raabe & Beelman, 2011). In addition, experi-mental research with older adolescents and withadults have yielded similar results as the currentstudy, for example, by showing that majority mem-bers prefer assimilation of immigrants, and thenintegration, rather than separation (e.g., Maison-neuve & Teste, 2007; Van Acker & Vanbeselaere,2012; Van Oudenhoven et al., 1998; but see Nigburet al., 2008). However, future studies should exam-ine systematically whether there is an age trend inthe evaluation of acculturation strategies of immi-grants. Developmental intergroup research hasshown that with age a more complex understand-ing of groups develop (Horn, 2003; Killen &Rutland, 2011). Adolescents compared to childrenmight be more concerned about the maintenanceand functioning of the Dutch in-group and there-fore endorse assimilation more strongly. However,their better understanding of the importance ofgroup identity and autonomy might also imply thatthey endorse integration more strongly becausethey recognize and accept that an immigrant peeralso wants to hold on to his or her own culture.Both processes might be at work at the same timewith the result that no age effects are found. Thus,more research on age differences is needed and thisresearch could also investigate children’s reasoningabout the implications and acceptance of differentacculturation strategies.

Our research was conducted in schools with arelatively low percentage of ethnic minority chil-dren and there were very small differences in thejudgments of children from different schools. In theNetherlands, as in other European countries, thereis a strong national debate on immigration andintegration but the great majority of schools arerather homogeneous because only around 15% ofthe population has an immigrant background.Furthermore, the multiethnic school is also anexception in the United States (Pettigrew, 2004).However, future studies could examine ethnicallymixed schools in which there are more opportuni-ties for intergroup contact. Contact with students ofdifferent ethnic or racial groups is associated withless prejudice (Tropp & Prenovost, 2008) and itmight also positively affect native children’s judg-ments about acculturating peers. However, thisdoes not have to mean that children in morediverse contexts make no evaluative distinctionbetween assimilating, integrating, and separatingimmigrant peers, or between immigrant andemigrant peers. Future studies should examine theprecise role of intergroup contact on children’s per-ceptions and evaluations of acculturating peers.

Another interesting extension of the currentresearch would be to focus not only on native major-ity group children but also on children of immigrantfamilies. A dynamic intergroup perspective on accul-turation (Brown & Zagefka, 2011) implies that theacculturation preferences of natives and of immi-grants are important to consider. There can be adynamic interplay between the perceptions of bothgroups and the concordance or fit between the accul-turation preferences of the native majority childrenand of immigrant children can be an importantdeterminant of the intergroup relations (Bourhis,Moise, Perrault, & S�enecal, 1997; Nigbur et al., 2008).In addition, relations between immigrant and ethnicminority group children are increasingly importantin many social settings but little is known about theways in which children of minority families reacttoward new immigrants (Pfeifer et al., 2007).

Conclusion

The present research has obvious strengths andthe findings have some practical implications. As forstrengths, the research goes beyond the substantialliterature on children’s ethnic and racial prejudiceby focusing on the evaluation of immigrants’ accul-turation strategies (Brown, 2010; Levy & Killen,2008). Migration and issues surrounding immigra-tion are important in many parts of the world and

Acculturation Preferences and Peer Evaluation 13

also in the lives of children, but little is knownabout children’s understanding of migrants. Theresearch also goes beyond the bulk of the accultura-tion literature by focusing on the evaluations ofnative majority children, rather than on the accultur-ation preferences and strategies of immigrant youth.In addition, it is one of the first investigations thatexamined the evaluation of the acculturation strate-gies of different immigrant groups and that made acomparison with native emigrants. This made itpossible to show that the evaluations depend on thestatus position of the minority group and on the in-tergroup context rather than on being a migrant perse. Furthermore, it is one of the few contributions tohave adopted an experimental approach and toexamine a specific theoretical mechanism (perceivedhost national belonging) that mediates between per-ceived acculturation strategies and the evaluation ofimmigrants.

The applied implications of the research relate tothe question of how support for heritage culturemaintenance could be encouraged among majoritychildren. It is clear that immigrants’ cultural main-tenance without host culture adoption is rejected bymost children. Segregation implies low perceivedhost national belonging and is evaluated rathernegatively by native peers (compare Wilson & Rod-kin, 2012). This means that encouraging culturaladoption among immigrants would have beneficialeffects for the majority children’s attitudes. In addi-tion, cultural adoption has beneficial effects forimmigrant youth’s adjustment to and functioning inthe host society (Berry, Phinney, Sam, & Vedder,2006), and these beneficial effects might even bestronger when an assimilation strategy is adopted.For some immigrants such a strategy might befavorable psychologically and for developing cross-ethnic friendships (Wilson & Rodkin, 2012), butothers find it undesirable or impossible to relin-quish their minority culture and identity. Mostimmigrants prefer a strategy of integration in whichthe value and distinctiveness of their heritage cul-ture is affirmed but in the context of attachmentand connection with the host society (Berry et al.,2006; Verkuyten, 2005). Therefore, encouragingmajority member’s acceptance of this dual identitystrategy is probably the most promising step towardmore positive intergroup relations among childrenand adolescents. This encouragement is importantbecause with age, adolescents might more stronglyview immigrants as a threat to the norms and val-ues of the majority culture and therefore becomeless tolerant and accepting of them (Gieling, Thijs,& Verkuyten, 2010; Gieling et al., 2011).

References

Abrams, D., & Rutland, A. (2008). The development ofsubjective group dynamics. In S. Levy & M. Killen(Eds.), Intergroup attitudes and relations in childhoodthrough adulthood (pp. 47–65). Oxford, UK: Oxford Uni-versity Press.

Abrams, D., Rutland, A., & Cameron, L. (2003). Thedevelopment of subjective group dynamics: Children’sjudgments of normative and deviant in-group and out-group individuals. Child Development, 74, 1840–1856.doi:10.1046/j.1467-8624.2003.00641.x

Abrams, D., Rutland, A., Pelletier, J., & Ferrell, J. (2009).Group nous and social exclusion: The role of theory ofsocial mind, multiple classification skill and social experi-ence of peer relations within groups. Child Development,80, 224–243. doi:10.1111/j.1467-8624.2008.01256.x

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychologicalresearch: Conceptual, strategic and statistical consider-ations. Journal of Personality and Social Psychology, 51,1173–1182. doi:10.1037/0022-3514.51.6.1173

Barrett, M. (2007). Children’s knowledge, beliefs and feelingsabout nations and national groups. New York: PsychologyPress.

Berry, J. W. (1997). Immigration, acculturation, and adap-tation. Applied Psychology: An International Review, 46,5–68. doi:10.1111/j.1464-0597.1997.tb01087.x

Berry, J. W., Phinney, J. S., Sam, D. L., & Vedder, P.(Eds.) (2006). Immigrant youth in cultural transition.Mahwah, NJ: Erlbaum.

Bigler, R. S., Spears Brown, C., & Markell, M. (2001).When groups are not created equal: Effects of groupstatus on the formation of intergroup attitudes in chil-dren. Child Development, 72, 1151–1162. doi:10.1111/1467-8624.00339

Bourhis, R., Moise, L. C., Perrault, S., & S�enecal, S. (1997).Towards an interactive acculturation model: A socialpsychological approach. International Journal of Psychol-ogy, 32, 369–386. doi:10.1080/002075997400629

Brown, R. (2010). Prejudice: It’s social psychology (2nd ed.).Oxford, UK: Wiley-Blackwell.

Brown, R., & Zagefka, H. (2011). The dynamics ofacculturation: An intergroup perspective. Advances inExperimental Social Psychology, 44, 129–184. doi:10.1016/B978-0-12-385522-0.00003-2

Byrne, D. E. (1971). The attraction paradigm. New York:Academic Press.

Cohen, J. (1988). Statistical power analysis for the behavioralsciences. Hillsdale, NJ: Erlbaum.

Crisp, R. J., & Hewstone, M. (2007). Multiple social cate-gorization. Advances in Experimental Social Psychology,39, 163–254. doi:10.1016/S0065-2601(06)39004-1

Curtis, R. C., & Miller, K. (1986). Believing another likesor dislikes you: Behaviors making the beliefs come true.Journal of Personality and Social Psychology, 51, 284–290.doi:10.1037/0022-3514.51.2.284

Dovidio, J. F., Gaertner, S. L., & Saguy, T. (2007). Anotherview of “we”: Majority and minority group perspec-

14 Verkuyten, Thijs, and Sierksma

tives on a common ingroup identity. European Review ofSocial Psychology, 18, 296–330. doi:10.1080/10463280701726132

Fox, J. (1997). Applied regression analysis, linear models, andrelated methods. Thousand Oaks, CA: Sage.

Gaertner, S. L., & Dovidio, J. F. (2000). Reducing intergroupbias: The common in-group identity model. Philadelphia,PA: Psychology Press.

Gaertner, S. L., Dovidio, J. F., Guerra, R., Rebelo, M.,Monteiro, M. B., Riek, B. M., et al. (2008). The commonin-group identity model: Applications to children andadults. In S. R. Levy & M. Killen (Eds.), Intergroup atti-tudes and relations in childhood through adulthood (pp. 204–219). Oxford, UK: Oxford University Press.

Gieling, M., Thijs, J., & Verkuyten, M. (2010). Tolerance ofpractices by Muslim actors: An integrative social-devel-opmental perspective. Child Development, 81, 1384–1399.doi:10.1111/j.1467-8624.2010.01480.x

Gieling, M., Thijs, J., & Verkuyten, M. (2011). Voluntary andinvoluntary immigrants and adolescents’ endorsement ofmulticulturalism. International Journal of Intercultural Rela-tions, 35, 259–267. doi:10.1016/j.ijintrel.2010.09.003

Gijsberts, M., Huijnk, W., & Vogels, R. (Eds.). (2011).Chinese Nederlanders: Van horeca naar hogeschool. Hague,the Netherlands: Social and Cultural Planning Office.

Goldstein, H. (1995). Multilevel statistical models. Oxford,UK: Oxford University Press.

Hagendoorn, L. (1995). Intergroup biases in multiplegroup systems: The perception of ethnic hierarchies.European Review of Social Psychology, 6, 199–228. doi:10.1080/14792779443000058

Hartmann, D., & Gerteis, J. (2005). Dealing with diversity:Mapping multiculturalism in sociological terms. Socio-logical Theory, 23, 218–240. doi:10.1111/j.0735-2751.2005.00251.x

Horn, S. S. (2003). Adolescents’ reasoning about exclusionfrom social groups. Developmental Psychology, 39, 71–84.doi:10.1037/0012-1649.39.1.71

Horn, S. (2006). Group status, group bias, and adoles-cents’ reasoning about treatment of others in schoolcontext. International Journal of Behavioral Development,30, 208–218. doi:10.1177/0165025406066721

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fitindexes in covariance structure analysis: Conventionalcriteria versus new alternatives. Structural EquationModeling, 6, 1–55. doi:10.1080/10705519909540118

Killen, M., & Rutland, A. (2011). Children and social exclu-sion: Morality, prejudice, and group identity. Oxford, UK:Wiley-Blackwell.

Killen, M., Rutland, A., Abrams, D., Mulvey, K. L., &Hitti, A. (2012). Development of intra- and intergroupjudgments in the context of moral and social-conven-tional norms. Child Development. Advance online publi-cation. doi:10.1111/cdev.12011

Levy, S. R., & Killen, M. (2008). Intergroup attitudes andrelations in childhood through adulthood. Oxford, UK:Oxford University Press.

MacKinnon, D. P., Warsi, G., & Dwyer, J. H. (1995). Asimulation study of mediated effect measures. Multivar-iate Behavioral Research, 30, 41–62. doi:10.1207/s15327906mbr3001_3

Maisonneuve, C., & Teste, B. (2007). Acculturation prefer-ences of a host community: The effects of immigrantacculturation strategies on evaluations and impressionformation. International Journal of Intercultural Relations,31, 669–688. doi:10.1016/j.ijintrel.2007.06.001

Malti, T., Killen, M., & Gasser, L. (2012). Social judgmentsand emotion attributions about exclusion in Switzer-land. Child Development, 83, 697–712. doi:10.1111/j.1467-8624.2011.01705.x

Marques, J. M., & P�aez, D. (1994). The “black sheepeffect”: Social categorization, rejection of ingroup devi-ates, and perception of group variability. EuropeanReview of Social Psychology, 5, 38–68. doi:10.1080/14792779543000011

Matera, C., Stefanile, C., & Brown, R. (2011). The role ofimmigrant acculturation preferences and generationalstatus in determining majority intergroup attitudes.Journal of Experimental Social Psychology, 47, 776–785.doi:10.1016/j.jesp.2011.03.007

Montreuil, A., & Bourhis, R. Y. (2001). Majority accultura-tion orientations toward values and devalued immi-grants. Journal of Cross-Cultural Psychology, 32, 698–719.doi:10.1177/0022022101032006004

Nesdale, D., Maass, A., Durkin, K., & Griffiths, J. (2005).Group norms, threat, and children’s racial prejudice.Child Development, 76, 652–663. doi:10.1111/j.1467-8624.2005.00869.x

Nigbur, C., Brown, R., Cameron, L., Hossain, R., Landau,A., Le Touze, D., et al. (2008). Acculturation, well-beingand classroom behaviour among white British and Brit-ish Asian primary-school children in the south-east ofEngland: Validating a child-friendly measure of accul-turation attitudes. International Journal of InterculturalRelations, 32, 493–504. doi:10.1016/j.ijintrel.2008.03.001

Oakes, P. J., Haslam, S. A., & Turner, J. C. (1994). Stereo-typing and social reality. Oxford, UK: Blackwell.

Pettigrew, T. F. (2004). Justice deferred: A half centuryafter Brown v. Board of Education. American Psycholo-gist, 59, 521–529. doi:10.1037/0003-066X.59.6.521

Pfeifer, J. H., Ruble, D. N., Bachman, M. A., Alvarez,J. M., Cameron, J. A., & Fuligni, A. J. (2007). Socialidentities and intergroup bias in immigrant and nonim-migrant children. Developmental Psychology, 43, 496–507.doi:10.1037/0012-1649.43.2.496

Poppe, E., & Verkuyten, M. (2012). Gender differences inprejudice: A study on evaluations of Muslims, Muslim menand Muslim women among Dutch adolescents. Utrecht, theNetherlands: Ercomer.

Raabe, T., & Beelman, A. (2011). Development of ethnic,racial, and national prejudice in childhood and adoles-cence: A multinational meta-analysis of age differences.Child Development, 82, 1715–1737. doi:10.1111/j.1467-8624.2011.01668.x

Acculturation Preferences and Peer Evaluation 15

Rasbash, J., Browne, W., Healy, M., Cameron, B., &Charlton, C. (2004). MLWin version 2.0. New York:Multilevel Models Project Institute of Education.

Schneider, D. J. (2004). The psychology of stereotyping. NewYork: Guilford.

Snijders, T. A. B., & Bosker, R. J. (1999). Multilevel analy-sis. An introduction to basic and advanced multilevel model-ing. London: Sage.

Tajfel, H., & Turner, J. C. (1979). An integrative theory ofintergroup conflict. In W. G. Austin & S. Worchel(Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks/Cole.

Tip, L., Zagefka, H., Gonz�alez, R., Brown, R., Cinirella,M., & Na, X. (2012). Is support for multiculturalismthreatened by …. threat itself? International Journal ofIntercultural Relations, 36, 22–30. doi:10.1016/j.ijintrel.2010.09.011

Tropp, L. R., & Prenovost, M. A. (2008). The role of inter-group contact in predicting children’s inter-ethnic atti-tudes: Evidence from meta-analytic and field studies. InS. Levy & M. Killen (Eds.), Intergroup attitudes and rela-tions in childhood through adulthood (pp. 236–248).Oxford, UK: Oxford University Press.

Van Acker, K., & Vanbeselaere, N. (2011). Bringingtogether acculturation theory and intergroup contacttheory: Predictors of Flemings’ expectations of Turks’acculturation behaviour. International Journal of Intercul-tural Relations, 35, 334–345. doi:10.1016/j.ijintrel.2010.06.004

Van Acker, K., & Vanbeselaere, N. (2012). Heritage cul-ture maintenance precludes host culture adoption andvice versa: Flemings’ perceptions of Turks’ accultura-tion behaviour. Group Processes and Intergroup Relations,15, 133–145. doi:10.1177/1368430211417263

Van Oudenhoven, J. P., Prins, K. S., & Buunk, B. P.(1998). Attitudes of minority and majority memberstowards adaptation of immigrants. European Journal ofSocial Psychology, 28, 995–1013. doi:10.1002/(SICI)1099-0992(1998110)28:6<995::AID-EJSP908>3.0.CO;2-8

Velasco Gonz�alez, K. V., Verkuyten, M., Weesie, J., &Poppe, E. (2008). Prejudice towards Muslims in theNetherlands: Testing integrated threat theory. BritishJournal of Social Psychology, 47, 667–685. doi:10.1348/014466608X284443

Verkuyten, M. (2005). The social psychology of ethnic iden-tity. Hove, UK: Psychology Press.

Verkuyten, M., & Kinket, B. (2000). Social distances in amulti-ethnic society: The ethnic hierarchy among Dutchpre-adolescents. Social Psychology Quarterly, 63, 75–85.doi:10.2307/2695882

Verkuyten, M., & Slooter, L. (2008). Muslim and non-Muslim adolescents’ reasoning about freedom of speechand minority rights. Child Development, 79, 514–528.doi:10.1111/j.1467-8624.2008.01140.x

Verkuyten, M., & Steenhuis, A. (2005). Preadolescents’understanding and reasoning about asylum seekerpeers and friendships. Applied Developmental Psychology,26, 660–679.

Verkuyten, M., & Thijs, J. (2001). Ethnic and gender biasamong Dutch and Turkish children in late childhood:The role of social context. Infant and Child Development,10, 203–217. doi:10.1002/icd.279

Verkuyten, M., & Thijs, J. (2002). Racist victimizationamong children in the Netherlands: The effect of ethnicgroup and school. Ethnic and Racial Studies, 25, 310–331.doi:10.1080/01419870120109502

Verkuyten, M., Weesie, J., & Eijberts, M. (2011). The eval-uation of perpetrators and victims of peer victimiza-tion: An extended crossed categorization approach.European Journal of Social Psychology, 41, 324–334. doi:10.1002/ejsp.777

Wilson, T. M., & Rodkin, P. C. (2012). Children’s cross-ethnic relationships in elementary schools: Concurrentand prospective associations between ethnic segrega-tion and social status. Child Development. Advanceonline publication. doi:10.1111/cdev.12020

Yee, M. D., & Brown, R. (1992). Self-evaluations andintergroup attitudes in children aged three to nine.Child Development, 63, 619–629. doi:10.1111/j.1467-8624.1992.tb01650.x

Zagefka, H., & Brown, R. (2002). The relationshipbetween acculturation strategies, relative fit and inter-group relations: Immigrant-majority relations inGermany. European Journal of Social Psychology, 32,171–188. doi:10.1002/ejsp.73

Zagefka, H., Tip, L. K., Gonzalez, R., Brown, R., & Cinni-rella, M. (2012). Predictors of majority members’ accul-turation preferences: Experimental evidence. Journal ofExperimental Social Psychology, 48, 654–659. doi:10.1016/j.jesp.2011.12.006

16 Verkuyten, Thijs, and Sierksma