125

British Journal of Developmental Psychology (2014), 32, 125–140 © 2013 The British Psychological Society www.wileyonlinelibrary.com

Subjective social status and intergroup attitudes among ethnic majority and minority children in Portugal Allard R. Feddes1*, Maria-Benedicta Monteiro2 and Mariline G. Justo2 1 2

Department of Social Psychology, University of Amsterdam, the Netherlands Centre for Psychological Research and Social Intervention, Lisbon University Institute (ISCTE-IUL), Lisbon, Portugal A measure of subjective social status (SSS) was examined among high (White), and low (Black and Roma) ethnic status children in Portugal within a developmental design including 6–8-year-old and 9–12-year-old children. White children favoured their in-group over the Black and Roma out-groups on the SSS measure, social preferences and positive as well as negative trait attributions. Generally, the Black and Roma showed equal SSS, preferences and trait attribution for their in-group and the high status White out-group, but not the other low-status out-group. With age White children generally demonstrated higher SSS for Black and Roma, preferred them more and attributed more positive traits. For low-status groups, an age effect was found only for Black children who preferred the Roma more with age and attributed more positive traits. Changes on preferences and trait attribution depending on age-group were mediated by SSS. It is concluded that minority group’s SSS does not parallel the objective status hierarchy but, rather, is a dynamic reorganisation of group’s relative positions serving strategies to cope with their minority condition.

Social status is an important determining factor of intergroup relations in childhood influencing a variety of social behaviours including intergroup attitudes (e.g., Bigler, Spears Brown, & Markell, 2001), in-group favouritism (e.g., Aboud, 1988), experiences and judgements about groups’ exclusion and inclusion (e.g., Killen, Sinno, & Margie, 2007), perception of groups’ variability (e.g., Guinote, Mouro, Pereira, & Monteiro, 2007), friendship preferences (e.g., Jugert, Noack, & Rutland, 2011), contact effects on the development of interethnic attitudes (e.g., Feddes, Noack, & Rutland, 2009), attitudes towards multiculturalism (e.g., Verkuyten & Martinovic, 2006), and identity development (Rutland et al., 2012). The objective of the present research was to investigate the association between high- and low-status children’s perceptions of their in-groups’ and other out-groups’ relative ethnic status positions and their intergroup attitudes. This research provides a novel contribution to the existing body of research in three ways: First, we examine children’s group-based perceptions of social status and not social status per se as a predictor of intergroup attitudes. We adopted a method previously used among adolescents: The subjective social status (SSS) measure (Goodman et al., 2001). *Correspondence should be addressed to Allard R. Feddes, Department of Social Psychology, University of Amsterdam, Weesperplein 4, 1018 XA, Amsterdam, the Netherlands (email: [email protected]). DOI:10.1111/bjdp.12025

126

Allard R. Feddes et al.

Second, we take a developmental perspective by measuring that association in two age groups, namely in 6- to 8- and 9- to 11-year-old children. This allows for a better comparison with other research on the development of intergroup attitudes in childhood (see Doosje et al., 2013; Raabe & Beelmann, 2011). We test the prediction that SSS mediates the association between age group and intergroup attitudes. Third, these associations are investigated in a multiple intergroup setting with one high-status (White) and two low-status groups (Roma and Black).1 Verkuyten and Zaremba (2005) note that research on the development of intergroup attitudes tends to focus on dyadic in-group versus out-group relations ignoring the dynamics of current multi-ethnic contexts in Western cities. Indeed, self-categorization theorists have argued that because social categories are relational, the context in which they are framed will be crucial to how any given category is used (Oakes, Haslam, & Turner, 1994). Previous research has provided evidence that even 5-year-olds’ stereotypes of the in-group differed depending upon the gender context (Sani, Bennett, Mullally, & MacPherson, 2003) or upon the multicultural context (Durkin, Nesdale, Dempsey, & McLean, 2012) in which the groups had been placed. We first provide some background information about the ethnic minority groups included in the study before reviewing research findings on the relationship between status and intergroup attitudes. Black Africans living in Portugal arrived massively from the Portuguese African ex-colonies after the independence period (1975) and represent now 23% of the 450 000 immigrant population (SEF, 2011). European reports regarding immigrant children in Portugal indicate that there is a high rate of school drop-out and a proportionally lower success rate among pupils of immigrant origin as compared with the majority White population (ECRI, 2013). Regarding the relationships between the native population and immigrant minorities, the same report states that there has been an increase in racist websites in recent years, targeting in particular Roma and Black immigrants. The Roma community, now estimated at 40 000–60 000 people, continues to experience serious problems in relation to equal rights and integration in Portugal (ECRI, 2013). Roma face many difficulties in the fields of employment, housing, and education. Discrimination is experienced in everyday life and a feeling of mutual mistrust between members of the Roma community and the majority population persists.

Subjective social status and intergroup attitudes in childhood Societal status of the ethnic immigrant group has been shown to explain variance in the expression of intergroup attitudes among children. For example, research using measures of ethnic preference and trait attribution has shown that children as young as 4 years old are aware of ethnicity and distinguish between in-group and out-group members (see for reviews Aboud, 1988; Bigler & Liben, 2006, 2007; Lam & Corson, 2013; Nesdale, 2001). Previous work using controlled experiments manipulating social status suggests that perceptions of social status are related to the expression of intergroup attitudes (e.g., Bigler et al., 2001). However, this notion has, to our knowledge, not been directly tested

1 We use the term ‘Black’ to refer to Portuguese with an African ethnic background in accordance with APA guidelines that note it is currently the acceptable term to use as long as it is used as a proper name and capitalized (APA, 2006). This term is commonly used in the context of research on intergroup relations and status differentials in childhood and adolescence; examples of recent publications in British Journal of Developmental Psychology are development of children’s intergroup attitudes (Nesdale, Lawson, Durkin, & Duffy, 2010), young children’s responses to media representations of intergroup threat and ethnicity (Durkin et al., 2012), national and ethnic identification (Lam & Corson, 2013), group identity and peer relations (Rutland et al., 2012).

Subjective social status and intergroup attitudes

127

before. This may be, in part, because social status has been defined as socioeconomic status (SES), which is an external, objective measure that does not account for subjective, internalized perceptions of social status. Wilkinson (1999) argued that these subjective perceptions of relative ranking can be more important determinants of health than objective indicators, such as income, which assess material resources. Indeed, in several studies, SSS showed a stronger relationship with health indicators than objective measures did (e.g., Adler, Epel, Castellazzo, & Ickovics, 2000; Goldman, Cornman, & Chang, 2006). Previous work has typically addressed the relative importance of SES and SSS on individual characteristics in adolescent populations at an intra-individual (e.g., Adler et al., 2000; Sweeting, West, Young, & Kelly, 2011) or an intragroup level (Hogg & Hains, 1996). However, little is known about when and how children begin to develop subjective perceptions of social hierarchies and use them as a cognitive resource in intergroup relations. According to developmental intergroup theory (DIT; Bigler & Liben, 2006, 2007), which focuses on the cognitive processes involved in stereotyping and prejudice development, children are sensitive to relative importance of social groups. Bigler and Liben (2007) argued that this is partly due to the fact that children are exposed to explicit statements linking social groups to attributes such as ‘African Americans are hostile’, ‘girls are shy’, or ‘majority children perform better in school’. According to these authors, children may thereby learn which group is more or less important. These processes are expected to co-occur with the advent of the concrete operational stage of cognitive development between 5 and 7 and formal operational thinking from approximately 7 years onwards (see also Aboud, 1988). Besides this cognitive account, it can be predicted based on social identity developmental theory (SIDT; Nesdale, 1999, 2004) that children’s motivation to identify with particular groups regulates perceptions of importance of in- and out-groups and thereby their expression of intergroup attitudes (Rutland et al., 2012). Based on SIDT, it can be predicted that children from minority and majority groups will actively engage in social comparisons to enhance or protect their social identity, which, in turn, will be related to intergroup evaluations. Research findings on trait attribution and social preferences among children of minority and majority status groups support this notion and show that different patterns of expression of intergroup bias are contingent to the societal groups’ status. For example, Alexandre, Monteiro, and Waldzus (2007) found that comparison processes seem to be already at work in 9- to 12-year-old children. In line with Tajfel (1981; for a review, see Brown, 2010), it was found that White children showed a higher preference for their in-group than for the Black and Roma out-groups. Both Black and Roma children used a different comparison strategy: They equally preferred the White higher-status out-group and their in-group over the other low-status out-group. Similar patterns were found in the USA by Pfeifer, Ruble, Bachman, Cameron, and Fuligni (2007) among immigrant (low-status) and non-immigrant (high-status) 9- to 12-year-old children. By including two age groups in this study, we expect to better understand the development of SSS perceptions regarding ethnic groups in a given society, as well as their influence on intergroup attitudes in younger and older ethnic majority and minority children. Two recent meta-analyses on the development of ethnic prejudice in childhood and adolescence have shown that heterogeneous developmental pathways seem to happen according to the groups’ relative social status: For older majority children, the expression of in-group favouritism remains unchanged or decreases with age compared to younger children (Doosje et al., 2013; Raabe & Beelmann, 2011). This is in line with

128

Allard R. Feddes et al.

socio-normative theory (Francßa & Monteiro, 2013; Monteiro, de Francßa, & Rodrigues, 2009; Rutland, Cameron, Milne, & McGeorge, 2005), which states that due to a greater awareness of antiracism social norms, with age majority children will display more positive and less negative attitudes towards out-groups. For minority groups, instead, it can be expected that, with age, they will display a more positive attitude towards their in-group which will be revealed by an attribution of more positive traits, less negative traits to, and higher preferences for the in-group than for out-groups (i.e., Aboud, 1988; Alexandre et al., 2007; Corenblum, Annis, & Young, 1996; Rutland et al., 2012). For out-group evaluations among minority groups, no age-related changes are expected based on previous meta-analytic results by Raabe and Beelmann (2011) and Doosje et al. (2013).

The present research Based on the review above, we expect that White children will favour their in-group over the Black and Roma out-groups on measures of SSS, preferences, and trait attribution (Hypothesis 1a). Compared to younger White children, older children are expected to perceive low-status out-groups to be of higher SSS, demonstrate a higher preference for, and assign more positive and less negative traits to low-status out-groups (Hypothesis 1b). The Black and Roma children are generally expected to perceive their in-group as having a higher SSS, prefer their in-group more, and assign more positive and less negative traits to their in-group compared to the other low-status, but not to the White high-status out-group (Hypothesis 2a). With age, we expect no differences for low-status children on measures of out-group SSS, out-group preferences, and trait attributions (Hypothesis 2b). Finally, we expect that changes in out-group preferences and trait attributions towards out-groups across age groups will be at least partially mediated by SSS (Hypothesis 3).

Method Participants Participants were 139 children from first and second (younger participants) and the fourth (older participants) grades of six public primary schools serving the same lower-middle-class community in Lisbon metropolitan area. There were, respectively, 25 younger (Mage = 6.16, SD = 0.37) and 25 older (Mage = 10.17, SD = 0.70) White participants (60% male), 26 younger (Mage = 6.77, SD = 0.51) and 25 older (Mage = 10.04, SD = 0.98) Black participants (39% male), and 17 younger (Mage = 6.94, SD = 0.90) and 21 older (Mage = 9.76, SD = 0.89) Roma participants (53% male). Percentages of ethnic minorities in each school were around 30%.

Procedure Participants with parental permission were individually interviewed by a female White experimenter. Children were told that a story would be written about a school that was very similar to theirs and their help was needed. The experimenter read the questions aloud and recorded children’s answers on an answering form. Three versions of the interview were created varying the order of presentation of target groups. The order of the measures was kept constant, and all materials were controlled for gender. First, children were asked about SSS followed by their social preferences and trait attributions. Before presenting the measures, children’s ethnic self-categorization (Milner, 1983) was

Subjective social status and intergroup attitudes

129

assessed. The children were shown three pictures (about 4 9 4 inches). The picture of the Black group contained three faces of Black children; the picture of the Roma group, three Roma children; and the picture of the White group, three White children. Children were then asked to choose the group to which they belonged. All children categorized themselves correctly.

Measures Subjective social status To assess SSS, we modified a measure previously used by Goodman et al. (2001) who used two ‘ladders’ to assess adolescents’ individual SSS. Each participant was presented with three pictures of, respectively, a White, Black, and Roma group (used in the self-categorization task) and was shown a 10-rung ‘social ladder’. Children were asked to assign their in-group and the two out-groups to one of the ladder rungs according to ‘the importance of the group in Portugal’. The use of the term ‘important’ is in line with the notion of Bigler and Liben (2006, 2007) that children learn about the relative importance of minority and majority groups in their societies. The experimenter explained that the first rung was where the least important group should be; the second rung was for less important groups, and so on until the tenth rung, where the most important group should be. No further additional information was given in regard to the context for such importance. Children should take each of the three pictures that were laid on the table in front of them and put it on a ladder rung. It was possible for the children to place two or all three groups on the same rung of the ladder. A score was obtained for each target group ranging from 1 (the least important) to 10 (the most important) within each participant ethnic group.

Social preferences This measure was adapted from Francßa and Monteiro (2002), who used four intergroup contact situations to ask participant children to indicate their social preferences. In the present research, children were asked to indicate how much they would like to engage in the following four contact situations with a child of each target group: ‘Sharing a chewing gum’, ‘cooperating on a school assignment’, ‘sharing a secret’, and ‘having a [target group] child living at his/her home’. Answers were given on a scale varying from 1 (not at all) to 5 (very much). The scale was reliable (alphas for, respectively, the White, Black, and Roma target groups were .65, .82, and .82). The mean of the four questions was taken as an overall preference score for each target group.

Trait attribution measure Twelve traits were used derived from the Multi-response Racial Attitude questionnaire (MRA; Doyle & Aboud, 1995) including six positive items (e.g., nice) and six negative items (e.g., mean). Participants were asked ‘How many (target group) children would you describe as (each trait)?’, on a scale ranging from 1 (none) to 5 (all). A pictorial format was used to make it easier to answer that question. The option 1 (none) was depicted with a cross (X), the option 2 (almost none) was depicted with two persons, with increasing number of persons until option 5 (all) which was depicted with 25 persons. Cronbach’s alphas for, respectively, the positive and negative evaluations of target groups were .63

130

Allard R. Feddes et al.

and .72 for the White, .82 and .76 for the Black, and .83 and .77 for the Roma. As the reliabilities were satisfactory, two separate positive and negative trait attribution scores were computed by averaging, respectively, the six positive and the six negative items. Higher scores on the positive attribution scale indicate more positivity, and higher scores on the negativity attribution scale more negativity.

Results Patterns of results of SSS, social preferences, and trait attributions To examine the effects of age and ethnic status on the outcome variables regarding the three target groups, 3 (participant’s ethnic background: White, Black, and Roma) by 3 (target’s ethnic background: White, Black, and Roma) 9 2 (age group: 6–8 vs. 9–12 years old) GLM repeated-measures analyses were performed on each measure. Participants’ ethnic status and age groups were between-subject factors, and SSS, preferences, and trait attributions were within-subject factors. When sphericity assumptions were not met, the Greenhouse–Geisser correction was applied (see Field, 2009). Post-hoc multiple comparison procedures with Bonferroni’s correction were used to assess the significance of differences between pairs of cell means. The means and standard deviations are given in Tables 1 and 2, as well as results of t-tests for differences between age groups and GLM repeated measures for differences across target groups.

Table 1. Means and standard deviations for subjective social status (SSS; 1 = low status; 10 = high status) and social preferences (1 = not at all; 5 = very much) for, respectively, the younger (6–8 years) and older (9–12 years) White, Black, and Roma participants (N = 139): t-tests for differences between age groups and GLM repeated measures for differences between target groups SSS White target M (SD) White participants Total 9.06 (2.00)a Younger 8.60 (2.57)a Older 9.52 (1.08)a t 1.65 Black participants Total 8.04 (2.53)a Younger 7.69 (3.10)a Older 8.40 (1.76)a t 1.01 Roma participants Total 6.39 (3.14)a Younger 6.88 (3.12)a Older 6.00 (3.18)a t 0.86

Black target M (SD)

Social preferences Roma target M (SD)

White target M (SD)

Black target Roma target M (SD) M (SD)

6.74 (3.06)b 5.40 (3.15)b 8.08 (2.34)b 3.41**

4.60 (3.57)c 3.64 (3.21)b 5.56 (3.72)c 1.95†

4.11 (0.66)a 4.36 (0.54)a 3.86 (0.69)a 2.87**

3.20 (1.06)b 2.73 (1.15)b 3.67 (0.73)a 3.46**

2.54 (1.05)c 1.98 (0.97)c 3.10 (0.82)b 4.41***

8.29 (2.48)a 7.77 (2.83)a 8.84 (1.97)a 1.56

4.69 (3.49)b 3.50 (2.92)b 5.92 (3.66)b 2.61*

3.73 (0.76)a 3.67 (0.77)a 3.79 (0.75)a 0.55

3.84 (1.07)a 3.52 (1.26)a 4.18 (0.68)a 2.33*

2.55 (1.16)b 2.04 (0.96)b 3.09 (1.11)b 3.62**

5.53 (3.41)a 5.76 (3.33)a 5.33 (3.55)a 0.38

5.74 (3.78)a 5.18 (3.96)a 6.19 (3.66)a 0.82

3.17 (0.98)a 3.10 (0.96)ab 3.23 (1.02)a 0.38

2.70 (1.06)b 2.59 (0.88)a 2.79 (1.20)a 0.59

3.84 (0.84)c 3.76 (0.91)b 3.89 (0.79)b 0.46

Note. T-values are given for differences between younger and older children: *p < .05; **p < .01; ***p < .001; †p < .10; different superscripts (a, b, and c) indicate significant differences between targets for each measure across participants (Bonferroni test).

Subjective social status and intergroup attitudes

131

Table 2. Means and standard deviations for positive and negative evaluations (1 = none; 5 = all) for, respectively, the younger (6–8 years) and older (9–12 years) White, Black, and Roma participants (N = 139): t-tests for differences between age groups and GLM repeated measures for differences between target groups Positive evaluations White target M (SD) White participants Total 4.28 (0.53)a Younger 4.28 (0.58)a Older 4.29 (0.48)a t 0.04 Black participants Total 4.15 (0.50)a Younger 4.17 (0.57)a Older 4.13 (0.43)a t 0.24 Roma participants Total 3.87 (0.65)a Younger 3.84 (0.69)a Older 3.89 (0.64)a t 0.21

Black target M (SD)

Negative evaluations

Roma target M (SD)

White target M (SD)

Black target M (SD)

Roma target M (SD)

3.57 (0.90)b 3.09 (0.92)b 4.05 (0.57)a 4.42***

2.94 (1.08)c 2.51 (1.14)c 3.37 (0.84)b 3.01**

2.51 (0.96)a 2.36 (1.05)a 2.65 (0.55)a 1.16

2.89 (0.92)b 2.99 (1.08.)b 2.63 (0.68)a 0.71

3.40 (1.04)c 3.65 (1.11)c 3.20 (0.70)b 1.72†

4.11 (0.65)a 4.12 (0.68)a 4.11 (0.62)a 0.05

3.05 (0.98)b 2.81 (0.99)b 3.30 (0.93)b 1.80†

2.53 (0.71)a 2.40 (0.82)a 2.65 (0.56)a 1.27

2.54 (0.80)a 2.45 (0.91)a 2.69 (0.44)a 0.79

3.33 (0.90)b 3.46 (1.06)b 3.20 (0.70)b 1.04

3.27 (0.87)b 3.06 (0.99)b 3.44 (0.74)b 1.38

3.82 (0.68)a 3.97 (0.79)a 3.71 (0.56)ab 1.20

2.83 (0.66)a 2.75 (0.58)a 2.90 (0.72)a 0.66

3.20 (0.81)b 3.11 (0.86)b 3.27 (0.78)b 0.61

3.04 (0.84)ab 2.68 (0.77)a 3.33 (0.80)b 2.57*

Note. T-values are given for differences between younger and older children: *p < .05; **p < .01; ***p < .001; †p < .10; different superscripts (a, b, and c) indicate significant differences between targets for each measure across participants (Bonferroni test).

Subjective social status The three-way interaction between target group, participant’s ethnicity, and age group was not significant, F(3.797, 252.519) = 0.54. No significant interaction was found between target group and age group, F(1.899, 252.519) = 2.28, ns. This means that perceptions of SSS of the White, Black, and Roma groups were generally stable across age groups. As expected, a significant interaction was found between target group and children’s ethnic background, F(3.797, 252.519) = 7.26, p < .001, g2p = .10. In line with Hypothesis 1a, independent of age, White children favoured their in-group over Black and Roma out-groups; the results showed a clear status hierarchy with White children perceiving their in-group as having the highest status and the Black having a significantly higher status than the Roma out-group (see Table 1). T-tests showed that older White children perceived both Black and Roma children to be of significantly higher status compared to judgements made by younger White children, which is in line with Hypothesis 1b. A different pattern was found for minority groups. Pairwise comparisons showed that both younger and older Black children saw themselves of equal status as the White, but of a higher status than the Roma out-group (see Table 1). Unexpectedly, Roma children did not show significant SSS differences between the target groups. Hypothesis 2a was therefore supported for Black children only. Hypothesis 2b was supported for Roma children: t-tests showed that no age differences in SSS of out-groups were present. Among Black children, Hypothesis 2b was partly supported as no age effects were found

132

Allard R. Feddes et al.

in regard to the White out-group, but with age, the Roma were evaluated as having higher status.

Patterns of intergroup preferences A three-way interaction between target group, children’s ethnicity, and age was found, F (4, 266) = 3.58, p < .01, g2p = .05. This interaction qualified a significant two-way interaction between target group and participant’s ethnicity F(4, 266) = 38.23, p < .001, g2p = .37. As can be seen in Table 1, post-hoc pairwise comparisons showed three different patterns across the three groups. White children generally showed a preference for their in-group over both out-groups, in line with Hypothesis 1a. T-tests showed that with age, White children showed a higher preference for both out-groups, in line with Hypothesis 1b. Black children expressed an equal preference for the Black and the White target groups, but preferred the Black over the Roma target group, in line with Hypothesis 2a. The Roma children preferred their in-group over both the Black and (against the expectation) the White out-group, thereby partly supporting Hypothesis 2a. T-tests across age groups showed that with age, Black children showed no increase in preference for the White out-group, but preferred the Roma out-group more, which goes against Hypothesis 2b. No significant changes in out-group preferences were found among the Roma children, in line with Hypothesis 2b.

Positive trait attributions A significant three-way interaction between target group, children’s ethnicity, and age group was found, F(1.840, 243.100) = 4.43, p < .01, g2p = .06. This interaction qualified a two-way interaction between target group and age group, F(1.828, 243.100) = 4.85, p < .05, g2p = .04. Children generally rated out-groups more positively with age as can be seen in Table 2. Three different patterns across the three groups were found: White children were generally more positive towards their in-group than towards both out-groups, supporting Hypothesis 1a. T-tests showed that both out-groups were evaluated significantly more positive with age, supporting Hypothesis 1b. As can be seen in Table 2, pairwise comparisons showed that Black and Roma children evaluated their in-group more positively than the other low-status group, but not the high-status White out-group, supporting Hypothesis 2a. T-tests showed that older Black children evaluated the Roma out-group as more positive (marginally), which is against Hypothesis 2b, while no increase in positivity was found in regard to the White out-group. In line with Hypothesis 2b, t-tests showed that there were no changes in positive trait attribution towards out-groups among Roma children.

Negative trait attributions Also for negative traits, a significant three-way interaction between target group, children’s ethnicity, and age group was found, F(3.670, 244.033) = 3.55, p < .05, g2p = .05. This interaction qualified a significant two-way interaction between target group and participant’s ethnicity, F(3.670, 244.033) = 7.85, p < .001, g2p = .11. Five different patterns across the three groups were found (see Table 2): In line with Hypothesis 1a, pairwise comparisons showed that White children were generally more negative towards both out-groups than to their in-group. In line with Hypothesis 1b, t-tests showed that the Roma out-group was evaluated (marginally) less negative. Against the

Subjective social status and intergroup attitudes

133

predictions, however, no reduction in negative trait attribution towards the Black out-group was found. Pairwise comparisons showed that Black children generally evaluated their in-group and the White out-group as equally negative. In contrast, they evaluated the Roma out-group as more negative, which is in line with Hypothesis 2a. Contrary to the expectations, Roma children evaluated their in-group as equally negative as the White and Black out-groups. T-tests showed no age differences in regard to out-group evaluations among Black and Roma children, which is in line with Hypothesis 2b.

Test of mediation of age effects by SSS To examine the hypothesis that age differences in social preferences and trait attribution were mediated by SSS, multiple regression analyses were performed separately for each ethnic group following the method described in Baron and Kenny (1986). Mediation analyses were only conducted if there was evidence of a change on the outcome variables depending on age group (see Tables 1 and 2). First, a regression analysis was conducted with age group as independent variable and preferences and trait attributions as respective outcome variables. The age categories were dummy-coded (6–8 years = 0; 9–12 years = 1). Second, the respective outcome variable was regressed on SSS. Third, the respective outcome variable was regressed on both age category and SSS. The results are given in Table 3 including the bootstrap2 results. The bootstrap confidence intervals showed that the indirect effect of age on the outcome variables via SSS was significant in all cases. This provides evidence for SSS as a (partial) mediator of age differences in social preferences and trait attributions. As can be seen in Table 3, the association between age category and the respective outcome variable became weaker in all cases when SSS was included in the model. Specifically, for White participants, SSS was found to partially mediate the relationship between age and, respectively, greater preference for and more positive trait attribution to Black and Roma out-groups. Evidence for a full mediation by SSS was found for the reduction in negative trait attribution towards the Roma out-group. For Black participants, the relationship between age and greater preference for and more trait attribution to the Roma out-group was, respectively, partly and fully mediated by SSS.

Discussion The results on the measure of SSS indicate that already from the age of 6 years onwards, children are aware of group differences in their relative positions on the social hierarchy. This finding is in line with developmental intergroup theory (DIT) (Bigler & Liben, 2006, 2007), which poses that children are aware of relative importance of different groups and this is reflected in their intergroup attitudes. These findings also support the idea that a very powerful normative consensus exists on the hierarchy of ethnic groups and that children engage in intergroup comparisons (e.g., Chafel, 1986; Tajfel, 1981; Yee & Brown, 1992). The present study shows, however, that depending on ethnicity and age, children’s 2 Shrout and Bolger (2002) state that mediation can be demonstrated by showing that the indirect effect is significantly different from zero. They recommend a bootstrap technique (see also Hayes, 2009; Preacher & Hayes, 2008) that computes a confidence interval around the product term (a*b). If zero is not included in the interval, the indirect effect is significant. Bootstrapping is superior to the Sobel test because Sobel assumes that a*b is normally distributed. However, especially in small samples, a more narrow and asymmetrical distribution is found. The bootstrap interval converges to the actual distribution of the indirect effect.

134

Allard R. Feddes et al.

Table 3. Results of mediation analyses among White and Black participants with age as independent variable (0 = 6–8 years old; 1 = 9–12 years old), subjective social status (SSS) as mediator and, respectively, social preferences, positive and negative trait attributions as outcome variables in regard to Black and Roma outgroups Participants’ ethnicity

Target’s ethnicity

White

Black

Roma

Black

Roma

Regression SSS on age Step 1: Social preferences on age Step 2: Social preferences on SSS Step 3: Social preferences on age (& SSS) Step 1: Positive trait attribution on age Step 2: Positive trait attribution on SSS Step 3: Positive trait attribution on age (& SSS) SSS on age Step 1: Social preferences on age Step 2: Social preferences on SSS Step 3: Social preferences on age (& SSS) Step 1: Positive trait attribution on age Step 2: Positive trait attribution on SSS Step 3: Positive trait attribution on age (& SSS) Step 1: Negative trait attribution on age Step 2: Negative trait attribution on SSS Step 3: Negative trait attribution on age (& SSS) SSS on age Step 1: Social preferences on age Step 2: Social preferences on SSS Step 3: Social preferences on age (& SSS) Step 1: Positive trait attribution on age Step 2: Positive trait attribution on SSS Step 3: Positive trait attribution on age (& SSS)

b (sig.) .44 (p .54 (p .49 (p .29 (p

< < <

Subjective social status and intergroup attitudes among ethnic majority and minority children in Portugal.

A measure of subjective social status (SSS) was examined among high (White), and low (Black and Roma) ethnic status children in Portugal within a deve...
173KB Sizes 1 Downloads 3 Views