Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Nov 1.
Published in final edited form as: Dev Sci. 2011 Aug 2;14(6):1283–1291. doi: 10.1111/j.1467-7687.2011.01072.x

Race preferences in children: insights from South Africa

Kristin Shutts 1, Katherine D Kinzler 2, Rachel C Katz 3, Colin Tredoux 4, Elizabeth S Spelke 3
PMCID: PMC3392197  NIHMSID: NIHMS384729  PMID: 22010888

Abstract

Minority-race children in North America and Europe often show less own-race favoritism than children of the majority (White) race, but the reasons for this asymmetry are unresolved. The present research tested South African children in order to probe the influences of group size, familiarity, and social status on children’s race-based social preferences. We assessed South African children’s preferences for members of their country’s majority race (Blacks) compared to members of other groups, including Whites, who ruled South Africa until 1994 and who remain high in status. Black children (3–13 years) tested in a Black township preferred people of their own gender but not race. Moreover, Black, White, and multiracial children (4–9 years) tested in a racially diverse primary school showed in-group bias by gender but not by race: all favored people who were White. Relative familiarity and numerical majority/minority status therefore do not fully account for children’s racial attitudes, which vary with the relative social status of different racial groups.

Introduction

Young children tend to like people who are similar to them, favoring playmates of their own age (French, 1987) and gender (Maccoby & Jacklin, 1987). Children also prefer others of their own race, but with an asymmetry: White children living in countries where White people are the numerical majority typically favor own-race members more than minority-race children do (for reviews, see Aboud, 1988; Aboud & Skerry, 1984; Brand, Ruiz & Padilla, 1974). Here we probe the nature of children’s asymmetric racial attitudes through studies conducted in a racially and ethnically diverse society where White people comprise a numerical minority yet have high social status: South Africa.

Previous research conducted in the US has shown that White children prefer unfamiliar White over Black and Asian children (e.g. Dunham, Baron & Banaji, 2006; Fox & Jordan, 1973; Kinzler, Shutts, DeJesus & Spelke, 2009; Kircher & Furby, 1971), attribute more positive traits to White children (Bigler & Liben, 1993; Kowalski, 2003; Williams, Best & Boswell, 1975), and show own-race preferences on implicit attitude measures (Baron & Banaji, 2006; Dunham et al., 2006). Similar own-race preferences have been observed in White children in Canada (Doyle & Aboud, 1995), the United Kingdom (Rutland, Cameron, Bennett & Ferrell, 2005), Australia (Griffiths & Nesdale, 2006), and New Zealand (Vaughan, 1978).

In the same countries, children from minority racial and ethnic groups tend to show weaker in-group favoritism (Aboud, 1988; Aboud & Skerry, 1984; Newman, Liss & Sherman, 1983; Ramsey & Myers, 1990; Spencer, 1984). For example, Aboud and Skerry (1984) reviewed 37 studies of attitudes of Black children living in countries where the majority population is White. Black children showed an own-race preference in only 27% of studies, and otherwise showed no preference (57%) or a White-race preference (16%). Similar patterns have been observed in children from other racial and ethnic minority groups in North America (Corenblum & Annis, 1987; Teplin, 1976), Europe (Alexandre, Monteiro & Waldzus, 2007; Davey & Mullin, 1980; Milner, 1973), and Oceania (e.g. Griffiths & Nesdale, 2006; Vaughan, 1964, 1978).

Three general tendencies could contribute to this asymmetry: Children may prefer members of groups that are larger (Brown & Bigler, 2002), more familiar (Cameron, Alvarez, Ruble & Fuligni, 2001; Cantor, 1972), or higher in status (Bigler, Brown & Markell, 2001; Brown & Bigler, 2002; Nesdale & Flesser, 2001). All three tendencies have been reported in adults (Cunningham, Nezlek & Banaji, 2004; Keltner, Young, Heerey, Oemig & Monarch, 1998; Lucken & Simon, 2002; Zajonc, 1968; Zebrowitz, White & Weineke, 2008) and are confounded in countries where the racial asymmetry has been reported. Moreover, children may become sensitive to all three of these factors through their personal encounters, their exposure to the media, or their observations of the social behavior of their families. The primary factors underlying children’s asymmetric race preferences are therefore unresolved.

Because of South Africa’s unique demographic profile and history, studying the preferences of its young citizens can help shed light on asymmetric racial attitudes. Like the US, South Africa is a multi-racial country where Whites historically held power over Blacks and other groups, and where Whites still possess highly disproportionate levels of wealth and education (Lehohla, 2005; Statistics South Africa, 2008). If social status is the primary force behind asymmetries in children’s racial preferences, then White South African children should prefer own-race people more than Black children do. Nevertheless, Blacks are the largest racial group in South Africa (approximately 79%; Lehohla, 2005). Due to a history of colonization and apartheid, South African Blacks have retained many aspects of their ethnic and cultural identities (Thompson, 2001). Most Black South Africans live in racially separated communities, speak more than one African language, and report feeling that they belong to a distinctive community with a unique culture (Bornman, 2006). If group size or familiarity are the primary forces behind asymmetric racial preferences, then South African children should show the reverse asymmetry, with Black children preferring own-race people more than White children do.

Studies of children in South Africa thus provide a unique opportunity to investigate how the factors of group size, familiarity, and status affect children’s developing race preferences. Curiously, however, the race preferences of children in contemporary South Africa have not received recent research attention. Two notable studies of South African children’s racial attitudes were conducted during the apartheid years: Using the classic Clark and Clark Doll Test (1947), Gregor and McPherson (1966) found that both Black and White South African children evaluated a doll with light pigmentation, blond hair, and blue eyes more favorably than a doll with dark pigmentation, black hair, and brown eyes. Critical of the use of dolls to assess children’s racial preferences, Fincham (1978) investigated Black and White South African preschoolers’ preferences using drawings that depicted children with identical physiognomy but different skin color and hair type (Fincham, 1978). The investigators found a preference for white stimuli in White children and no consistent preferences in Black children. It is not clear, however, how children perceived the drawn figures, or whether children in contemporary South Africa would show the same preferences.

Contemporary South Africa stands in stark contrast to South Africa during the apartheid years. The recent history of South Africa has brought sweeping changes to the country’s laws, values, and aspirations (see Finchilescu & Tredoux, 2010, and accompanying articles in the Journal of Social Issues; Volume 66, Issue 2). Contemporary South Africa is a vibrant mix of racial and ethnic groups, with 11 official languages, including Xhosa (the native language of most Black children in the present studies) and English (the language of commerce, and the native language of some White South Africans and some people of mixed racial heritage). Among the individuals held in greatest esteem are the Nobel laureates Nelson Mandela (the Xhosa founder of modern South Africa) and Archbishop Desmond Tutu (the country’s Xhosa religious leader). These leaders, and the national constitution and government, promote South African unity over racial or ethnic divisions. The context of contemporary South Africa may lead South African children of all races to use nationality more than race to organize their social world.

In five studies, we assessed the social preferences of children in a multiracial urban school (Studies 1, 4, and 5) and in a racially homogeneous Xhosa township (Studies 2 and 3). We presented participants with photographs of adults and children living in contemporary South Africa, and assessed both children’s categorization of, and preferences between, members of different social groups. We used photographs of real South Africans who were unfamiliar to children and who were given no racial or ethnic labels, in order to increase the naturalness of the task. Moreover, each task showed photographs of many different people in order to decrease the likelihood that children’s reactions to particular photographs would have undue influence on the general pattern of findings.

Study 1

Study 1 probed the race preferences of Black, Coloured (children of mixed heritage), and White children attending a racially diverse primary school in Cape Town. Children viewed Xhosa (the region’s predominant Black South African ethnic group) faces paired with White, Coloured, or foreign Black African faces. Children were asked to point to the person they preferred.

Method

Participants

The participants were 87 children (43 boys; M age = 6.5 years; range = 4–8 years). Children were identified by parents as Black (23), Coloured (36), White (15), or other (Indian, > 1 racial category, or no report: 13; see SOM for more information on children’s social identities).

Materials

Children viewed 48 faces of men and women photographed in Cape Town and categorized by race, ethnicity, and nationality by local adults (see SOM). Equal numbers of male and female faces were classified as Xhosa (N = 24), Coloured (N = 8), White (N = 8), or foreign Black African (N = 8). Photographs (approximately 8 × 10 cm) appeared in pairs on a laptop computer.

Design

On every trial, children viewed one photograph of a Xhosa adult face paired with a photograph of a same-gender adult of different race or nationality (White, Coloured, or foreign Black African). Across trials, eight pairs of each type were intermixed, blocked by gender (with male pairs first for half the children). The lateral position of the Xhosa face was counterbalanced both within each gender and pairing condition and across children.

Procedure

Children were tested individually in a quiet room at their school. The task was administered in English (the language of instruction) by a White female tester who asked them to ‘point to the person you like’. She then presented the 24 pairs of photographs in series, providing neutral feedback on children’s choices.

Results

As a group (N = 87), participants preferred faces that were not Xhosa, choosing the Xhosa face on only 36% of trials (Chance = 50%; t(86) = −9.59, p < .001). Children’s preferences depended, however, on the type of comparison face. Children preferred White over Xhosa faces (M = 78%; t(86) = 12.30, p < .001) and Coloured over Xhosa faces (M = 79%, t(86) = 13.05, p < .001), but chose Xhosa over foreign Black African faces (M = 66%, t(86) = 8.53, p < .001). All these effects were observed for faces of each gender, for male and female children analyzed separately, and for the children identified as Black, Coloured, and White analyzed separately (Figure 1 and SOM). There was no correlation between participant age and preferences.

Figure 1.

Figure 1

Open in a new tab

Studies 1 and 2: Race preferences of Coloured, Black, and White children tested in a multiracial school, and of Black children tested in a Black township. The graphs show participants’ mean percentage choices of Xhosa (Black South African) faces when compared to White South African, Coloured, and foreign Black African faces. Error bars represent standard error and asterisks indicate bars that are significantly above or below chance according to one-sample t tests (*** p < .001; ** p < .01; * p < .05). One-sample t-test values (left to right) are as follows: t(35) = −8.57, −10.09, 4.17; t(22) = −8.93, −7.89, 4.72; t(14) = −2.79, −2.92, 6.61; t(31) < 1, < 1, = 2.90).

Responses of Black, Coloured, and White children were subjected to a 3 (comparison face) × 2 (face gender) × 3 (child race) × 2 (child gender) ANOVA, with % choice of Xhosa faces as the dependent variable. This analysis revealed a main effect of comparison face (F(1.73, 117.54) = 130.82, p < .001, ηp2 = .66; Greenhouse-Geisser correction). Children chose Xhosa faces over foreign Black African faces more often than they chose Xhosa over Coloured or White faces (each p < .001, simple effects tests). There was a main effect of participant race (F(2, 68) = 3.17, p < .05, ηp2 = .09): Tukey HSD post-hoc tests revealed that the tendency to select non-Xhosa over Xhosa faces was smaller for White than for Black or Coloured children (ps < .05; Figure 1), whose choices did not differ. Effects of face gender and participant gender were small (see SOM).

Discussion

Black, Coloured, and White children in Study 1 showed highly similar race preferences: all preferred White and Coloured to Black South African people, even though Blacks comprise the largest proportion of South Africa’s population. Moreover, children did not exhibit systematic own-race favoritism: indeed, Black children showed a stronger preference for non-Xhosa faces than did White children. These findings cast doubt on the thesis that asymmetric race preferences are caused primarily by differences in the sizes or familiarity of different racial groups.

The children also preferred Black South Africans to Blacks from foreign African countries. Because foreign Africans in South Africa have little political power, and are generally disliked by South Africans (Crush & Dodson, 2007), the most parsimonious account of all the data is that children prefer members of higher-status groups. Nevertheless, features of the testing environment, procedure, or participant pool may have masked or attenuated other effects on children’s race preferences. Accordingly, we repeated the study with a different population of Black children, tested in a different setting: namely, a predominantly Black township.

Study 2

We administered the preference test of Study 1, as well as a categorization task, to Xhosa children living in Langa (population 50,000; Statistics South Africa, 2001), a predominantly Black township in the Cape Town area. Children were tested in a Xhosa home by a Xhosa tester speaking in Xhosa. In a single testing session, we assessed both children’s social preferences and their group categorizations of the people in the photographs used in Study 1.

Method

The method was the same as Study 1, except as follows. Participants (20 males, 12 females; M age = 10 years; range = 5–13 years) were recruited from the community by word of mouth. For one 24-trial block, children pointed to the person they liked; for another 24-trial block, children viewed the same pairs and pointed to the person who was Xhosa. Block order and participant gender were orthogonally counterbalanced.

Preliminary analyses revealed no effects of block order.

Results

When asked to point to the person who was Xhosa, children performed above chance whether the comparison face was White, Coloured, or foreign Black African (Chance = 50%; M = 97%, t(31) = 31.57, p < .001; M = 95%, t(31) = 23.99, p < .001; and M = 67%, t(31) = 4.86, p < .001; respectively). An ANOVA with comparison face, face gender, and participant gender as factors revealed only a main effect of comparison face (F(1.20, 36.05) = 56.37, p < .001, ηp2 = .65; Greenhouse-Geisser correction): children were more accurate at distinguishing Xhosa faces from Coloured and White faces than from foreign Black African faces (ps < .001, simple effects tests). Older children were more adept at distinguishing Xhosa from non-Xhosa faces (r = .59, p < .001).

Children showed no preference between Xhosa and non-Xhosa faces (Chance = 50%; M Xhosa Choices = 54%; t(31) < 1), between Xhosa and White faces, or between Xhosa and Coloured faces (Figure 1). On trials featuring one Xhosa and one foreign Black African face, children showed a Xhosa preference (Figure 1). Nevertheless, a 3 (comparison face) × 2 (face gender) × 2 (participant gender) ANOVA revealed no significant effects. There was no correlation between participant age and preferences.

Discussion

Xhosa children in Study 2 did not show a preference for members of their own race, even though sessions were administered in a Xhosa home situated in a predominantly Black township, by a Xhosa tester speaking to children in Xhosa. Children’s lack of own-race favoritism did not stem from a failure to categorize or from a lack of motivation to respond consistently: when asked to categorize Xhosa faces, children performed reliably for all face comparisons; moreover, they showed a reliable preference for Xhosa over foreign Black African faces.

Although Xhosa children in Langa did not show an own-race preference, they also did not show the marked preferences for White or Coloured South Africans observed among the children tested in Cape Town. In Study 5, we explored the differences between the findings of the two studies, asking whether they could stem from differences in the testing procedures (especially the race and language of the tester), test settings, or the social and academic experiences of the children in the two studies.

Studies 1 and 2 suggest that social status may play a role in guiding children’s race preferences. Children in both samples preferred Black South Africans when they were paired with lower status foreign Africans, but not when they were paired with higher status White and Coloured South Africans. Nevertheless, the children in Studies 1 and 2 made judgments of adults, who may especially evoke status concerns. Accordingly, in Study 3 we assessed the race and gender preferences of Xhosa children living in Langa using photographs of children as well as adults. We included a test for gender preferences because gender is a pervasive social category (Ruble, Martin & Berenbaum, 2006), providing a yardstick against which race preferences could be evaluated.

Study 3

We presented Xhosa children in Langa with pairs of photographs of South African adults and children who differed in race (Black vs. White) or gender. Social preferences were measured as in Study 2.

Method

The method was the same as Study 2 except as follows. Xhosa children (N = 64; 32 boys; M age = 7 years; range = 3–11 years) viewed photographs of child and adult faces (eight Black and eight White at each age and gender), categorized as Xhosa or White by local adults (see SOM). For the race condition, the photographs were arranged in 16 Black-White pairs of the same gender and age. For the gender condition, the photographs were arranged into 16 male-female pairs of the same race and age.

Design

Half the children were tested in the race condition, and half in the gender condition. Children viewed one eight-trial block featuring pairs of child faces followed by a block featuring pairs of adult faces. In the race condition, male and female pairs were blocked in counterbalanced order across participants; the lateral positions of Black and White faces were counterbalanced within and across participants and blocks. In the gender condition, the race and lateral positions of the faces were similarly counterbalanced.

Results

In the race condition, children showed no preference between Black and White faces (Figure 2, left). A 2 (face gender: same vs. different from participant) by 2 (face age: child vs. adult) by 2 (participant gender) ANOVA revealed only an effect of face gender (F(1, 30) = 4.20, p < .05, ηp2 = .12): children showed no preference between Black and White faces on trials featuring faces of their own gender (M Black Choices = 48%, t(31) < 1), but showed a preference for White faces on trials featuring faces of the opposite gender (M White Choices = 60%; t(31) = 2.92, p < .01). We chose to analyze effects of face gender according to whether faces matched participants’ own gender because children typically show robust preferences for children their own gender (Ruble et al., 2006).

Figure 2.

Figure 2

Open in a new tab

Studies 3–5: Race preferences of Black children tested in a Black township, and of Coloured, Black, and White children tested in a multiracial school. The graphs show participants’ mean percentage choices of Black faces (as opposed to White faces). Error bars represent standard error and asterisks indicate bars that are significantly below chance according to one-sample t-tests (*** p < .001; ** p < .01; * p < .05). One-sample t-test values (left to right) are as follows: t(31) = −1.81, p = .08; t(35) = −10.29; t(22) = −4.45; t(14) = −3.67; t(11) = −2.24.

In the gender condition, children showed an own-gender preference (Figure 3, left). A 2 (face race: Black vs. White) by 2 (face age) by 2 (participant gender) ANOVA revealed an interaction of face pair race and participant gender (F(1, 30) = 4.41, p < .05, ηp2 = .13): boys showed a greater own-gender preference on Black trials than on White trials (M = 70% vs. 56%, respectively; p < .01), whereas girls did not (M = 67% and 68%, respectively). There was no correlation between participant age and performance in the race or the gender condition.

Figure 3.

Figure 3

Open in a new tab

Studies 3–5: Gender preferences of Black children tested in a Black township, and of Coloured, Black, and White children tested in a multiracial school. The graphs show participants’ mean percentage choices of own-gender faces. Error bars represent standard error and asterisks indicate bars that are significantly above chance according to one-sample t-tests (*** p < .001). One-sample t-test values (left to right) are as follows: t(31) = 5.42; t(35) = 7.37; t(22) = 5.03; t(14) = 4.97; t(11) = 4.91.

Race and gender preferences were compared through a 2 (participant gender) × 2 (condition: gender vs. race) × 2 (face age) ANOVA, which revealed only an effect of condition (F(1, 60) = 24.16, p < .001, ηp2 = .29). Xhosa children’s own-gender preference reliably exceeded their own-race preference.

Discussion

Whether they were evaluating photographs of children or adults, Black children showed no preference for Blacks over Whites. In contrast, these children reliably preferred others of their own gender when tested with the same set of photographs. Children’s clear own-gender preferences indicate that they understood the task, were motivated to perform it, and extracted social information from the photographs. Children’s contrasting performance on the gender and race trials therefore suggests that these children prefer other children and adults of their own gender but not of their own race.

Study 4

We next used the method of Study 3 to test the race and gender preferences of the children from the multiracial Cape Town school.

Method

The children from Study 1 (N = 87) were given the child face trials from Study 3 after completion of Study 1. All children saw eight Black–White race preference trials (four with faces of girls and four with faces of boys), and eight male–female gender preference trials (four with faces of White children, and four with faces of Black children), in a within-subjects design. Approximately half of the children were tested on race preferences before gender preferences. As in Study 1, the tester was a White female who administered the task in English.

Results and discussion

On race trials, children showed a significant preference for White children (M = 78%; t(86) = 11.65, p < .001); see also Figure 2, middle. A 3 (participant race: Black, Coloured, White) × 2 (participant gender) × 2 (face pair gender: same as vs. different from participant) ANOVA revealed only a weak and hard to interpret three-way interaction of the three factors (F(2, 68) = 3.56, p < .05, ηp2 = .10): the lowest preference for White children was shown by White males on own-gender trials (M = 56%) (see SOM).

On gender trials, children showed a significant preference for faces of their own gender (M = 76 %; t(86) = 11.60, p < .001); see also Figure 3, middle. A 3 (participant race: Black, Coloured, White) × 2 (participant gender) × 2 (face pair race: Black vs. White) ANOVA indicated a main effect of participant gender (F(1, 68) = 4.82, p < .05, ηp2 = .07), qualified by an interaction of participant gender and participant race (F(2, 68) = 6.86, p < .005, ηp2 = .17). Black and White girls tended to show stronger own-gender preferences than their male peers; the opposite was true for Coloured children (see SOM). To compare own-gender to own-race preferences, we conducted a further analysis of the responses of the Black and White children only. For this analysis, children’s race choices were re-scored as choices of their own race vs. the opposite race, parallel to the scoring of gender choices. The scores were analyzed by an ANOVA with trial type (race vs. gender) as a within-subject factor and participant race (Black, White) and gender as between-subjects factors. There was a main effect of participant gender (F(1, 34) = 12.22, p < .005, ηp2 = .26): Girls generally showed stronger own-group preferences (M = 68%) than boys (M = 53%). There were also main effects of trial type (F(1, 34) = 29.88, p < .001, ηp2 = .47) and participant race (F(1, 34) = 30.66, p < .001, ηp2 = .47), which were qualified by an interaction of the two factors (F(1, 34) = 10.77, p < .005, ηp2 = .24). Black and White participants showed equally strong own-gender preferences, but differed in the strength of their own-race preferences (p < .001): White children preferred own-race photographs, whereas Black children preferred other-race (i.e. White) photographs.

Study 5

These studies reveal very similar gender preferences among children of different races tested in different communities: namely, Black children in Langa and children of all races in the multiracial Cape Town school showed robust preferences for people of their own gender. Nevertheless, the studies reveal somewhat different patterns of race preferences. Children in a multiracial Cape Town school showed a preference for White children (Study 4), whereas Black children in the predominantly Black township of Langa showed no overall preference between Blacks and Whites (Study 3). What accounts for this difference?

Two procedural differences between Studies 3 and 4 concern the order of testing and the race and language of the tester (Xhosa in Study 3 but not 4; see Aboud, 1988, and Annis & Corenblum, 1986). To equate for these possible effects, we conducted a fifth study with a new group of 12 Black children (six boys; M age = 7.5 years; range = 5–9 years) from the same multiracial school in Cape Town. A Xhosa tester speaking in Xhosa (the home language of all 12 children) tested children’s preferences using only child face pairs. Much like participants in Study 4, children in Study 5 showed a significant preference for White over Black faces on race trials (Figure 2, right) and a significant preference for own-gender face on gender trials (Figure 3, right).

Although Black children in the multiracial Cape Town school were tested in Xhosa by a Xhosa tester, they still showed a significant preference for White over Black children. Black South African children did not show an own-race preference in any of the five studies.

General discussion

The present research begins to shed light on the sources of children’s developing race-based social preferences. Taken together, the findings from these studies suggest four primary conclusions, and raise a host of interesting questions for future research. First, group size and familiarity do not appear to be the primary determinants of children’s race preferences. Blacks are by far the largest racial group in South Africa, yet Black children in two communities showed no preference for Black adults or children. In racially homogeneous Langa, moreover, Blacks are far more frequently encountered than Whites, but are not preferred. The pervasive preference for Whites observed in children living in majority-White countries, therefore, is not likely explained by the greater size and familiarity of that racial group.

Second, children’s racial preferences appear to reflect the relative status of different racial groups in South Africa. In the multiracial Cape Town school, children of all races showed a preference for White and Coloured over Black South Africans. This pattern mirrors the relative wealth, education, and opportunity of these racial groups in South Africa (Lehohla, 2005; Statistics South Africa, 2008). In Langa, Xhosa children showed no preference between Black and White faces; this finding resembles that found in most studies of African-American children in the US (Aboud & Skerry, 1984). In a related line of research, we recently found that Xhosa children attending school in English preferred Black speakers of English to Black speakers of Xhosa (Kinzler, Shutts & Spelke, in press). Since English is viewed (by many adults) as a language of opportunity and status in South Africa (De Klerk, 2000; Deumert, 2010), both children’s language-based social preferences and their racial preferences may be sensitive to social status. Experiments that directly measure or manipulate children’s perceptions of the social status of different races and languages are needed to test this suggestion.

Third, children develop race-based social preferences even when they are raised in a country that emphasizes national unity over racial and ethnic divides (see also Singh, Choo & Poh, 1998). Indeed, the results of the present studies are similar to findings from studies of children’s racial attitudes conducted during the apartheid years (Fincham, 1978; Gregor & McPherson, 1966). Despite their country’s recent efforts to bridge racial distinctions and promote unity, South African children still show sensitivity to race. In future work, it will be important to understand the information and experiences that contribute to children’s perceptions and preferences based on race.

Finally, social status is not the only factor that contributes to children’s developing social group preferences. Langa children’s lack of a consistent White preference suggests that familiarity and group size also may modulate children’s racial preferences. In future work, it will be important to test children in a range of settings and collect detailed information about their exposure to members of different racial groups in order to clarify whether and how familiarity and group size are related to children’s racial attitudes. Experiments that manipulate group size and familiarity are necessary to test the causal effects of these factors on children’s racial attitudes (see Bigler & Liben, 2007, for a summary of such an approach using novel groups). Moreover, social status may not guide children’s gender preferences: children in the present studies did not favor males over females, even though, on average, men in South Africa earn more money and achieve higher levels of education than women (Budlender, 2002). Although social status may be the primary force behind children’s asymmetric race preferences, other factors evidently shape children’s social attitudes as well.

The present findings cast a clearer perspective on the developing race preferences of children worldwide. In many societies, including those of North America and Europe, children may show preferences for Whites over other races because they prefer adults and other children of higher status, and they view race as indicative of social standing. This suggestion, in turn, raises a host of questions for future research. For example, previous research suggests that children are attuned to various cues to social status (Bigler et al., 2001), including differences in prestige (Chudek, Birch, Heller & Henrich, under review), occupation (Bigler, Averhart & Liben, 2003), and power over activities and rewards (Bussey & Bandura, 1984). Which of the multifaceted aspects of social status matter most to children in the context of emerging racial group attitudes? Additionally, how do young children become attuned to status differences within their society (e.g. by watching interactions between adults; through their own personal social interactions and observations; from exposure to media)? Finally, why is social status such a potent variable for young children who have little or no direct immediate access to wealth, power, or advanced education themselves? The research findings presented here, together with future research, may shed light on these questions.

Since the end of the apartheid regime, South Africa has overcome many obstacles and made many strides in addressing social injustices. However, demographic data reveal that racial groups in South Africa still differ markedly in wealth and access to education (Lehohla, 2005; Statistics South Africa, 2008). The present findings suggest that South Africa’s youngest citizens may be influenced by their perceptions of social status differences such as these. Explicit messages of national unity and racial non-discrimination, such as those that have been promulgated in South Africa throughout the lives of its current generation of children, may not suffice to eradicate children’s social biases. As future research probes more deeply the sources of these biases, it should help to clarify both the societal and psychological changes that are needed to eliminate them.

Acknowledgments

This research was supported by Harvard Psychology Department Allport and Stimson Fund grants to KS and Kristina Olson; by a Harvard Mind, Brain, and Behavior Initiative Fellowship to KS; by a Harvard International Innovation Fund grant to ESS and Mahzarin Banaji; and by NIH Grant HD23103 to ESS. We thank Vuyokazi Amanda Sijila, Harold Sijila, Thando Mxutu, Linda Dlungana, and Thami Hlatshwayo for assistance with data collection; Bridget Spelke for taking the photographs of South African children; Kara Weisman for assistance with the manuscript; and Mahzarin Banaji and Kristina Olson for their ideas, support, and advice on this research.

References

  1. Aboud F. Children and prejudice. Cambridge, MA: Basil Blackwell; 1988. [Google Scholar]
  2. Aboud FE, Skerry SA. The development of ethnic attitudes: a critical review. Journal of Cross-Cultural Psychology. 1984;15:3–34. [Google Scholar]
  3. Alexandre JD, Monteiro MB, Waldzus S. More than comparing with majorities: the importance of alternative comparisons between children from different minority groups. International Journal of Psychology and Psychological Therapy. 2007;7:201–212. [Google Scholar]
  4. Annis RC, Corenblum B. Effects of test language and experimenter race on Canadian Indian children’s racial and self-identity. Journal of Social Psychology. 1986;126:761–773. [Google Scholar]
  5. Baron AS, Banaji MR. The development of implicit attitudes: evidence of race evaluations from ages 6, 10, and adulthood. Psychological Science. 2006;17:53–58. doi: 10.1111/j.1467-9280.2005.01664.x. [DOI] [PubMed] [Google Scholar]
  6. Bigler RS, Averhart CJ, Liben LS. Race and the workforce: occupational status, aspirations, and stereotypes among African American children. Developmental Psychology. 2003;39:572–580. doi: 10.1037/0012-1649.39.3.572. [DOI] [PubMed] [Google Scholar]
  7. Bigler RS, Brown CS, Markell M. When groups are not created equal: effects of group status on the formation of intergroup attitudes in children. Child Development. 2001;72:1151–1162. doi: 10.1111/1467-8624.00339. [DOI] [PubMed] [Google Scholar]
  8. Bigler RS, Liben LS. A cognitive-developmental approach to racial stereotyping and reconstructive memory in Euro-American children. Child Development. 1993;64:1507–1518. [Google Scholar]
  9. Bigler RS, Liben LS. Developmental intergroup theory: explaining and reducing children’s social stereotyping and prejudice. Current Directions in Psychological Science. 2007;16:162–166. [Google Scholar]
  10. Bornman E. National symbols and nation-building in the post-apartheid South Africa. International Journal of Intercultural Relations. 2006;30:383–399. [Google Scholar]
  11. Brand ES, Ruiz RA, Padilla AM. Ethnic identification and preference. Psychological Bulletin. 1974;81:860–890. [Google Scholar]
  12. Brown CS, Bigler RS. Effects of minority status in the classroom on children’s intergroup attitudes. Journal of Experimental Child Psychology. 2002;83:77–110. doi: 10.1016/s0022-0965(02)00123-6. [DOI] [PubMed] [Google Scholar]
  13. Budlender D for Statistics South Africa. Women and men in South Africa: 5 years on. 2002 Retrieved 31 March 2010 from: http://www.statssa.gov.za/publications/statsdownload.asp?PPN=womenandmen5yearson&SCH=2376.
  14. Bussey K, Bandura A. Influence of gender constancy and social power on sex-linked modeling. Journal of Personality and Social Psychology. 1984;47:1292–1302. doi: 10.1037//0022-3514.47.6.1292. [DOI] [PubMed] [Google Scholar]
  15. Cameron JA, Alvarez JM, Ruble DN, Fuligni AJ. Children’s lay theories about ingroups and outgroups: reconceptualizing research on prejudice. Personality and Social Psychology Review. 2001;5:118–128. [Google Scholar]
  16. Cantor GN. Effects of familiarization on children’s ratings of pictures of Whites and Blacks. Child Development. 1972;43:1219–1229. [Google Scholar]
  17. Chudek M, Heller S, Birch S, Henrich J. Prestige-biased cultural learning: Bystander’s differential attention to potential models influences children’s learning. (under review) [Google Scholar]
  18. Clark KB, Clark MP. Racial identification and preference in Negro children. In: Newcomb TM, Harley EL, editors. Readings in social psychology. New York: Holt; 1947. pp. 169–178. [Google Scholar]
  19. Corenblum B, Annis RC. Racial identity and preferences in Native and White Canadian children. Canadian Journal of Behavioral Sciences. 1987;19:254–265. [Google Scholar]
  20. Crush J, Dodson B. Another lost decade: the failures of South Africa’s post-apartheid migration policy. Tijdschrift voor Economische en Sociale Geografie. 2007;98:436–454. [Google Scholar]
  21. Cunningham WA, Nezlek JB, Banaji MR. Implicit and explicit ethnocentrism: revisiting the ideologies of prejudice. Personality and Social Psychology Bulletin. 2004;30:1332–1346. doi: 10.1177/0146167204264654. [DOI] [PubMed] [Google Scholar]
  22. Davey AG, Mullin PN. Ethnic identification and preferences of British primary school children. Journal of Child Psychology and Psychiatry. 1980;21:241–251. doi: 10.1111/j.1469-7610.1980.tb01799.x. [DOI] [PubMed] [Google Scholar]
  23. De Klerk V. To be Xhosa or not to be Xhosa … that is the question. Journal of Multilingual and Multicultural Development. 2000;21:198–215. [Google Scholar]
  24. Deumert A. Tracking the demographics of (urban) language shift – an analysis of South African census data. Journal of Multilingual and Multicultural Development. 2010;31:13–35. [Google Scholar]
  25. Doyle AB, Aboud FE. A longitudinal study of White children’s racial prejudice as a social-cognitive development. Merrill-Palmer Quarterly. 1995;41:209–228. [Google Scholar]
  26. Dunham Y, Baron AS, Banaji MR. From American city to Japanese village: a cross-cultural investigation of implicit race attitudes. Child Development. 2006;77:1268–1281. doi: 10.1111/j.1467-8624.2006.00933.x. [DOI] [PubMed] [Google Scholar]
  27. Fincham FD. Race concepts among Black and White preschool children in South Africa. Journal of Behavioural Science. 1978;2:291–295. [Google Scholar]
  28. Finchilescu G, Tredoux C. The changing landscape of intergroup relations in South Africa. Journal of Social Issues. 2010;66:223–430. [Google Scholar]
  29. Fox DJ, Jordan VB. Racial preference and identification of black, American Chinese, and white children. Genetic Psychology Monographs. 1973;88:229–286. [PubMed] [Google Scholar]
  30. French DC. Children’s social interaction with older, younger, and same-age peers. Journal of Social and Personal Relationships. 1987;4:63–86. [Google Scholar]
  31. Gregor AJ, McPherson DA. Racial preference and ego-identity among white and Bantu children in the republic of South Africa. Genetic Psychology Monographs. 1966;73:217–253. [PubMed] [Google Scholar]
  32. Griffiths JA, Nesdale D. In-group and out-group attitudes of ethnic majority and minority children. International Journal of Intercultural Relations. 2006;30:735–749. [Google Scholar]
  33. Keltner D, Young RC, Heerey EA, Oemig C, Monarch ND. Teasing in hierarchical and intimate relations. Journal of Personality and Social Psychology. 1998;75:1231–1247. doi: 10.1037//0022-3514.75.5.1231. [DOI] [PubMed] [Google Scholar]
  34. Kinzler KD, Shutts K, DeJesus J, Spelke ES. Accent trumps race in guiding children’s social preferences. Social Cognition. 2009;27:623–634. doi: 10.1521/soco.2009.27.4.623. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Kinzler KD, Shutts K, Spelke ES. Language-based social preferences among children in South Africa. Language Learning and Development. doi: 10.1080/15475441.2011.583611. (in press) [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kircher M, Furby L. Racial preferences in young children. Child Development. 1971;42:2076–2078. [Google Scholar]
  37. Kowalski K. The emergence of ethnic and racial attitudes in preschool-aged children. Journal of Social Psychology. 2003;143:677–690. doi: 10.1080/00224540309600424. [DOI] [PubMed] [Google Scholar]
  38. Lehohla P of Statistics South Africa. Census 2001: Achieving a better life for all: Progress between Census ’96 and Census 2001 (report no. 03-02-16 (2001)) 2005 Retrieved 12 November 2009 from: http://www.statssa.gov.za/census01/html/C2001SumReport.asp.
  39. Lucken M, Simon B. Cognitive and affective experiences of minority and majority members: the role of group size, status, and power. Journal of Experimental Social Psychology. 2002;41:396–413. [Google Scholar]
  40. Maccoby EE, Jacklin CN. Gender segregation in childhood. In: Reese EH, editor. Advances in child development and behavior. Vol. 20. New York: Academic Press; 1987. pp. 239–287. [DOI] [PubMed] [Google Scholar]
  41. Milner D. Racial identification and preferences in ‘black’ British children. European Journal of Social Psychology. 1973;3:281–295. [Google Scholar]
  42. Nesdale D, Flesser D. Social identity and the development of children’s group attitudes. Child Development. 2001;72:506–517. doi: 10.1111/1467-8624.00293. [DOI] [PubMed] [Google Scholar]
  43. Newman MA, Liss MB, Sherman F. Ethnic awareness in children: not a unitary concept. Journal of Genetic Psychology. 1983;143:103–112. [Google Scholar]
  44. Ramsey PG, Myers LC. The salience of race in young children’s cognitive, affective and behavioral responses to social environments. Journal of Applied Developmental Psychology. 1990;11:49–67. [Google Scholar]
  45. Ruble DN, Martin CL, Berenbaum SA. Gender development. In: Damon W, Eisenberg N, editors. Handbook of child psychology. 6. Vol. 3. New York: Wiley; 2006. pp. 858–931. [Google Scholar]
  46. Rutland A, Cameron L, Bennett L, Ferrell J. Interracial contact and racial constancy: a multi-site study of racial intergroup bias in 3–5 year old Anglo-British children. Applied Developmental Psychology. 2005;26:699–713. [Google Scholar]
  47. Singh R, Choo WM, Poh LL. In-group bias and fair-mindedness as strategies of self-presentation in intergroup perception. Personality & Social Psychology Bulletin. 1998;24:147–162. [Google Scholar]
  48. Spencer MB. Black children’s race awareness, racial attitudes and self-concept: a reinterpretation. Journal of Child Psychology and Psychiatry. 1984;25:433–441. doi: 10.1111/j.1469-7610.1984.tb00162.x. [DOI] [PubMed] [Google Scholar]
  49. Statistics South Africa. City of Cape Town: Census 2001: Langa: Demographic profile by gender. 2001 Retrieved 8 March 2010 from: http://www.capetown.gov.za/en/stats/2001census/Documents/Langa.htm.
  50. Statistics South Africa. Income & expenditure of households 2005/2006: Statistical release P0100. 2008 Retrieved 16 November 2009 from: http://www.statssa.gov.za/publications/statsdownload.asp?PPN=P0100&SCH=4108.
  51. Teplin LA. A comparison of racial/ethnic preferences among Anglo, Black and Latino children. American Journal of Orthopsychiatry. 1976;46:702–709. doi: 10.1111/j.1939-0025.1976.tb00968.x. [DOI] [PubMed] [Google Scholar]
  52. Thompson L. The history of South Africa. 3. New Haven, CT: Yale University Press; 2001. [Google Scholar]
  53. Vaughan GM. The development of ethnic attitudes in New Zealand school children. Genetic Psychology Monographs. 1964;70:135–175. [PubMed] [Google Scholar]
  54. Vaughan GM. Social change and intergroup preferences in New Zealand. European Journal of Social Psychology. 1978;8:297–314. [Google Scholar]
  55. Williams JE, Best DL, Boswell DA. The measurement of children’s racialattitudes in the early school years. Child Development. 1975;46:494–500. [Google Scholar]
  56. Zajonc RB. Attitudinal effects of mere exposure. Journal of Personality and Social Psychology. 1968;9:1–27. [Google Scholar]
  57. Zebrowitz LA, White B, Wieneke K. Mere exposure and racial prejudice: exposure to other-race faces increases liking for strangers of that race. Social Cognition. 2008;26:259–275. doi: 10.1521/soco.2008.26.3.259. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES