Child Development, March/April 2005, Volume 76, Number 2, Pages 451 466

Social Norms and Self-Presentation: Childrens Implicit and Explicit Intergroup Attitudes
Adam Rutland and Lindsey Cameron
University of Kent

Alan Milne and Peter McGeorge

University of Aberdeen

Two studies examined whether social norms and childrens concern for self-presentation affect their intergroup attitudes. Study 1 examined racial intergroup attitudes and normative beliefs among children aged 6 to 16 years (n 5 155). Accountability (i.e., public self-focus) was experimentally manipulated, and intergroup attitudes were assessed using explicit and implicit measures. Study 2 (n 5 134) replicated Study 1, focusing on national intergroup attitudes. Both studies showed that children below 10 years old were externally motivated to inhibit their in-group bias under high public self-focus. Older children were internally motivated to suppress their bias as they showed implicit but not explicit bias. Study 1, in contrast to Study 2, showed that children with low norm internalization suppressed their out-group prejudice under high public self-focus.

An American journalist (Hockstader, 2001) asked three Arab eighth graders to name their heroes and, after some nervous glances and fluttering smiles, they named a suicide bomber who killed himself and 21 others at a Tel Aviv disco. The nervous glances and fluttering smiles suggest that these Arab children were unsure about expressing their genuine attitudes to an American reporter. Nevertheless, maybe because they were being interviewed among their peers and within their local school community, these children eventually stated what we presume to be their actual beliefs. This example indicates that although children often harbor negative attitudes to others, they are also sensitive to their audience and social norms. This article examines how social norms and childrens concern for self-presentation may affect their expression of prejudice. There is a long tradition in social psychology of studying how social norms influence prejudice (Asch, 1956; Pettigrew, 1958; Sherif, 1936). Social norms prescribe appropriate attitudes, values, and behavior in a given situation. Several studies on adults have shown that changing audience norms of prejudice expression can have a strong effect on peoples measured intergroup attitudes (e.g., Blanchard, Crandall, Brigham, & Vaughan,

This research was supported by a grant (R00022318) from the Economic and Social Research Council. The authors extend their gratitude for feedback to Dominic Abrams, Rupert Brown, Sheri Levy, Roger Giner-Sorolla, and the anonymous reviewers. Much appreciation is extended to the teachers and children for their participation. Correspondence concerning this article should be addressed to Adam Rutland, Department of Psychology, University of Kent, Canterbury CT2 7NP, UK. Electronic mail may be sent to [email protected].

1994; Jetten, Spears, & Manstead, 1997; Monteith, Deneen, & Tooman, 1996). A contemporary social norm in Western societies is to avoid expressing discriminatory attitudes or behavior toward individuals based on their ethnic or racial group membership (Dovidio & Gaertner, 1991). This social norm, like many others, is conveyed through the mass media (see Graves, 1999), national laws (e.g., equal opportunity legislation), and special multicultural education programs (e.g., Bigler, 1999; McLeod, 1993), which promote tolerance and appreciation of cultural diversity in elementary schools and in wider society. Research provides evidence that children attend to the social norm that blatant or straightforward racial discrimination is inappropriate (Killen, Lee-Kim, McGlothlin, & Stangor, 2002; Killen, Pisacane, Lee-Kin, & Ardila-Rey, 2001; Killen & Stangor, 2001; Rutland, 2004; Theimer, Killen, & Stangor, 2001). Studies with college students have also shown that concern about expressing unacceptable prejudicial beliefs (i.e., self-presentation) is related to reported levels of racial intergroup bias (e.g., Plant & Devine, 1998). However, related research with children has been mixed. For example, Doyle, Beaudet, and Aboud (1988) found that first- through fifth-graders scores on the Childrens Social Desirability measure (CSD; Crandall, Crandall, & Katkovsky, 1965) did not relate significantly to their racial intergroup attitudes. However, a problem has been noted with the CSD (see Levy & Troise, 2001) insofar as it is a global measure of the tendency to present oneself in an unrealistically positive way (e.g., in terms of manners
r 2005 by the Society for Research in Child Development, Inc. All rights reserved. 0009-3920/2005/7602-0010

452

Rutland, Cameron, Milne, and McGeorge

and the expression of negative feeling states) rather than a specific measure of concerns about expressing prejudicial social attitudes. Levy and Troise (2001) developed and validated a measure to examine childrens social concern about expressing negative intergroup attitudes. In two studies with fifth- and sixth-grade children, higher social concern was related to more positive attitudes toward overweight and elderly people. This finding is compatible with studies that have shown that children provide more positive attitudes when evaluating the group of which the experimenter is a member (Jahoda, Thomson, & Bhatt, 1972; Katz, Sohn, & Zalk, 1975; Lawrence, 1991). Together these studies suggest that children may engage in selfpresentation behavior when high in social concern or in the presence of an in-group member. The studies presented here examined whether 6to 16-year-olds control their explicit expressions of prejudice toward some racial groups (Study 1) and nationalities (Study 2). Previous research suggests that prejudice suppression is related to social norms concerning discrimination (e.g., Blanchard et al., 1994; Crandall, Eshleman, & OBrien, 2002). Thus, a preliminary study was conducted to examine childrens social norm about out-group prejudice. The childrens internalization of this social norm was also examined, and external pressure (i.e., public self-focus) to comply with the norm was manipulated. Nonetheless, explicit bias suppression may occur even without external pressure; thus, implicit intergroup attitudes were also measured. The studies considered any distinctive developmental trends in childrens explicit in-group and out-group attitudes (Aboud, 2003, Cameron, Alvarez, Ruble, & Fuligni, 2001; Nesdale, 2004). Previous adult research suggests that the suppression of prejudice is motivated by either internal or external reasons (e.g., Fazio, Jackson, Dunton, & Williams, 1995; Plant & Devine, 1998). Externally motivated suppression represents compliance to normative pressure from others (Kelman, 1958) whereas internal suppression represents egalitarian or humanitarian motivations to be fair minded (Monteith, 1993; Plant & Devine, 1998). The developmental literature suggests two alternative hypotheses regarding the age at which children will show evidence of external and internal motivation to suppress their prejudice (Abrams, Rutland, & Cameron, 2003; Killen, Lee-Kim, et al., 2002; Killen & Stangor, 2001; Piaget, 1965; Quintana, 1994; Ruble, Alvarez, Bachman, & Cameron, 2004; Selman, 1971, 1980). First, research on social perspective taking suggests that children below 8 years old are relatively

poor at coordinating and integrating various psychological perspectives such as first-, second-, and third-person perspectives (Quintana, 1994, 1999; Selman, 1971, 1980). Therefore, only older children may be aware of the social norm against explicit prejudice expression and will be externally motivated to suppress their prejudice under high public self-focus. Alternatively, developmental research suggests that young children are aware of the social norm that racial discrimination is inappropriate (Killen, LeeKim, et al., 2002; Killen & Stangor, 2001). For example, Killen and Stangor (2001) found that 6- to 12-year-olds rated racial exclusion as a negative behavior and unacceptable. Young children should, therefore, be externally motivated to suppress their racial prejudice under high public self-focus (Crandall et al., 2002). In contrast, the development of social identification and awareness of group deviancy in middle childhood should mean that children begin to internalize the social norm and become internally motivated to control their prejudice (Abrams, Rutland, & Cameron, 2003; Abrams, Rutland, Cameron, & Marques, 2003; Quintana, 1998; Ruble et al., 2004). Thus, older childrens explicit attitudes should be unaffected by a manipulation of public self-focus. This conception of the transition from externally motivated conformity to prejudice suppression based on internalization is similar to Piagets (1965) ideas about the development of moral reasoning. First, Piaget argued that the childs moral reasoning is heteronomous; namely, moral norms are understood as one-sided and originating from external authority. Second, from approximately 10 years of age childrens reasoning becomes autonomous as they internalize fundamental moral norms and begin to regulate internally their own moral behavior. This internalization process requires the development of empathy, perspective taking, and logical and counterfactual reasoning (Crandall et al., 2002). In line with cognitive-developmental theory (Aboud, 1988; Bigler & Liben, 1993; Katz et al., 1975), this account suggests that childrens levels of prejudice are related to the acquisition of specific social cognitive skills. The effect of manipulating public self-focus on childrens out-group attitudes should be dependent on whether children have internalized the social norm that explicit prejudice is unacceptable (Crandall et al., 2002). Children who show low levels of norm internalization should suppress their out-group prejudice under high public self-focus. In contrast, children with high levels of norm internalization (i.e., those high in internal motivation) should be unaffected by external pressure. This finding should be most evident with out-group prejudice because

Childrens Intergroup Attitudes

453

external pressures to suppress out-group prejudice are typically higher than for in-group bias (Brewer, 1979; Cameron et al., 2001; Nesdale, 2004; Rutland, Brown, Ahmavaara, Arnold, & Samson, 2003) and our measure of norm internalization was a measure of belief about prejudice toward the out-group. Implicit measures of intergroup attitudes were used in our studies to demonstrate that older children are internally motivated to control their prejudice. Namely, at the controlled (i.e., explicit) level older children should not show bias, whereas they should show bias at the uncontrolled (i.e., implicit) level. Research into childhood intergroup attitudes has typically used explicit measures (e.g., Doyle et al., 1988; Williams, Best, Boswell, Mattson, & Graves, 1975). Arguably, the sole use of such measures increases the likelihood that children will engage in self-presentation and internally control their prejudice (Brown, 1995; Nesdale, 2001; Powlishta, Serbin, Doyle, & White, 1994). Therefore, the present studies used a modified version of the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998) to measure childrens uncontrolled implicit attitudes. Research with adults has shown that the IAT is resistant to masking by self-presentation strategies (e.g., Kim, 2003; Monteith, Voils, & Ashburn-Nardo, 2001). Dissociation between explicit and implicit measures is typically observed, especially for socially sensitive issues and particularly for racial attitudes (Fazio et al., 1995; Greenwald et al., 1998). Thus, dissociation between explicit and implicit measures, and implicit bias in all age groups, was predicted. It was expected that older children, who should be internally motivated, would show significant implicit bias but no explicit prejudice. The first study examined whether children engaged in self-presentation by controlling their racial in-group bias and out-group prejudice. This involved measuring their explicit and implicit intergroup attitudes while manipulating public self-focus using a video camera. The first study also examined the childrens internalization of the racial prejudice social norm (i.e., a personal normative belief measure) using social exclusion vignettes. Three age groups were included: 6 to 8 years, 10 to 12 years, and 14 to 16 years. The two youngest age groups were included because intergroup bias typically declines between these ages among White children (see Aboud, 1988; Brown, 1995). A sample was also taken from adolescents, given the commonly found resurgence of bias during this period (Rutland, 1999; Teichman, 2001). A preliminary study (see the Appendix for full details) was conducted on a cohort of children sim-

ilar to those used in the present studies to determine whether discriminatory judgments based on a persons racial category (e.g., Black and Asian) and national category (e.g., American, German, British) were perceived as unacceptable. This study found that prejudice based on both racial and national group membership was perceived as illegitimate. However, the social-norm-prohibiting bias premised on racial group membership was stronger than that for national group membership, especially in relation to Germans. These findings suggest that children are more likely to control their explicit racial bias than national bias. Study 1 Method Participants and Design Participants were 155 White British children (57 males, 98 females) from a predominately White middle-class suburban or rural area outside a large metropolitan city in the southeast region of England. Approximately 3% to 4% of the school population was from a visible racial minority, and most of the teachers were White British. Included were forty-five 6- to 8-year-old children (M 5 7 years 10 months, SD 5 9 months), sixty-two 10- to 12-year-old children (M 5 10 years 8 months, SD 5 7 months), and fortyfour 14- to 16-year-old children (M 5 15 years 5 months, SD 5 5 months). Procedure Children were given a personal normative belief measure, an explicit intergroup bias measure, and an implicit intergroup bias measure, administered in counterbalanced orders. The second author tested children individually in their schools. Assignment to the high or low public self-focus condition was random within each age group. Gender was evenly distributed between conditions. Public self-focus was manipulated using a video camera (Macrae, Bodenhausen, & Milne, 1998). In the high public self-focus condition, children were videotaped completing the explicit and implicit attitude measures. The children were explicitly told they were being videotaped and the recording taken may be watched later by the researcher and her colleagues (i.e., other adults, teachers). Children were shown themselves on screen using the video cameras LCD monitor. This procedure was repeated between each task to ensure the salience of the video camera. In the low public self-focus condition, the

454

Rutland, Cameron, Milne, and McGeorge

children were told the battery for the video camera was not functioning. Then, the lens cap was placed on the video camera and it was switched off and pointed down toward the floor. The public-self focus manipulation was successful according to a study (see the Appendix for further details) using a subsample of children tested using the public self-consciousness scale (Fenigstein, Scheier, & Buss, 1975). Stimulus Materials and Measures Personal normative belief measure. The preliminary study described earlier established that the social norm was the nonexpression of explicit racial bias. Children were presented with social exclusion vignettes to determine whether they had internalized this social norm. There were two versions of the vignettes, with the content of the story and the genders of the characters varied to ensure gender relevance. In the male vignette, because of racial group membership, White boys excluded two Black boys from a soccer game. Similarly, in the female vignette, White girls excluded two Black girls from eating their school lunch (see the Appendix for both vignettes).The selection of the vignettes was based on a small-scale pilot study with 6- to 16-year-old participants. This showed that British children of all ages related to the activities described and that the vignettes maintained their attention. Each vignette included three black-and-white line drawings presented with cartoon captions read by the second author, except among the 14- to 16-year-olds because research suggests that the use of cartoons does not maximize the attention of older children (e.g., Nucci, 1981). Instead, adolescents were presented with typed event descriptions and asked to read these out loud to themselves. Next, the children were asked two questions. The first question was: How bad do you think the boys/ girls in the story were who wouldnt let the other children join in? The 6- to 12-year-olds answered on a 4-point scale of smiley faces. The scale was: OK (smiley face 5 1), bad (little frown 5 2), very bad (bigger frown 5 3), and very, very bad (very big frown 5 4). The valence of the faces from left to right was counterbalanced. Smetana and colleagues (Smetana, 1981; Smetana & Braeges, 1990; Smetana, Kelly, & Twentyman, 1984) have used this scale previously when studying childrens moral development. The choice of this scale also involved consideration of the developmental literature that suggests children are likely to perceive instances of racial exclusion as unacceptable (Killen, Lee-Kim, et al., 2002; Killen & Stangor, 2001). Therefore, a scale

was selected that was sensitive enough to identify possible variability in childrens beliefs about the acceptability of racial discrimination. The 14- to 16-year-olds answered using a simple 4-point numbered scale with the words (OK, bad, very bad, and very very bad) written below. The children were next asked a second question: What do you think the teacher should do? and were required to select one of three responses: not punish (scored 1), punish a little (scored 2), and punish a lot (scored 3). The childrens responses to the two questions were significantly correlated (r 5 .50, po.001). However, 23 children of our total sample failed to answer the second question. It was decided to maximize statistical power by using the first question only as an indication of personal normative belief. There was a marginal main effect for age group on personal normative beliefs, F(2, 154) 5 2.76, po.1. This effect revealed an increase in norm internalization with age (Ms 5 3.27, 3.56, and 3.61 for 6- to 8-year-olds, 10- to 12-year-olds, and 14- to 16-year-olds, respectively). The marginal significant difference was between the 6- to 8-year-olds and the 14- to 16-year-olds (Tukeys honestly significant different [HSD] test, po.1). The childrens responses to the first question were strongly skewed, with 3% responding OK, 8% responding bad, 26% responding very bad, and 63% responding very very bad. Moreover, the childrens responses were equally skewed within each age group. Thus, it was inappropriate to use the childrens responses as a continuous variable within a regression model. Instead, a median split was performed on the childrens responses, classifying children who responded OK, bad, and very bad into the low norm internalization group (n 5 57, 37%) and the children who responded very very bad into the high norm internalization group (n 5 98, 63%). There were equivalent distributions of children from each age group in the low and high groups. In the context of this study, children in the high group should show more concern than those in the low group about publicly expressing racial bias because they had internalized to a greater extent the norm that racial discrimination is unacceptable. Explicit intergroup bias measure. The Multiple-Response Racial Attitude (MRA) measure was used to derive separate indexes of in-group bias and out-group prejudice (Aboud, 2003). The children were presented with 20 adjectives, 10 positive and 10 negative. These adjectives were taken from the Preschool Racial Attitude Measure II (Pram II) Series A (Williams et al., 1975). The positive adjectives were: clean, wonderful, healthy, good, nice, happy, friendly, kind, helpful, and smart. The negative

Childrens Intergroup Attitudes

455

adjectives were: unfriendly, mean, dirty, cruel, stupid, selfish, sick, naughty, sad, and bad. To ensure the children understood the meaning of each word, they were given a corresponding definition. Initially, children were presented with two copies of each word and two cups labeled White and Black. The cups were also identified using simple gender-neutral cartoon faces. The children were explicitly told they could put the adjective in the Black cup, the White cup, or both cups. For instance, one item read, Some children are friendly. They often share their toys with other children. Who is friendly? Is it the Black child, the White child or more than one child who is friendly? To ensure understanding of the task, initially each child was asked to practice assigning adjectives by attributing four nonevaluative items to each of the categories (Black-Gutman & Hickson, 1996). These items were likes to run, likes to sing, likes TV, and likes music. The number of adjectives assigned to each cup was computed. Four scores were calculated: a positive and a negative trait score for each racial group, each with a possible range of 0 to 10. The higher the scores on positive traits and the higher the scores on negative traits, the more positive and negative, respectively, were the childs ratings. A separate in-group bias score ranging from 10 (very unfavorable) to 110 (very favorable) was created from the number of in-group positive evaluations minus the number of in-group negative evaluations. In addition, a separate out-group bias score was determined from the number of out-group negative evaluations minus the number of out-group positive evaluations. This score ranged from 110 (very unfavorable) to 10 (very favorable). Therefore, the higher the childrens ingroup bias and out-group bias scores, the more was their racial in-group favoritism and racial out-group prejudice, respectively. Implicit intergroup bias measure. The IAT (Greenwald et al., 1998) was used as a computer-based implicit measure. All children were tested on this task in their schools using the same Mac notebook computer supplied by the researchers. The IAT is a task devised to measure uncontrolled or automatic concept attribute associations. The underlying assumption of the test is that strongly associated (compatible) attribute concept pairs should be easier to classify together than weakly associated or opposed (incompatible) attribute concept pairs. Typically, in the IAT participants are presented with a series of words on a computer screen, which are exemplars of a concept (e.g., names associated with the Black racial category; i.e., Latishia, Ebony) and an attribute, (e.g., pleasant and unpleasant words).

Furthermore, participants have to categorize these words as quickly as possible by pressing a left or right key on a keyboard. Pilot work with 6- to 7-year-old children indicated that they had problems understanding a word-based IAT. Therefore, a completely pictorial-based version of the IAT was developed. Instead of using stereotypical Black and White names, unfamiliar Black and White faces with neutral facial expressions were used as concepts. The faces were approximately 3 4 cm. The sex of faces was matched with each childs sex. In addition, as an alternative to pleasant and unpleasant words, simple line drawings of happy or sad cartoon faces were used as attributes. The cartoon faces varied in shape (e.g., squares, triangles, circles) and were 4 4 cm. Pilot work indicated that some children had difficulty using keypad responses; therefore, children were required to respond by making movements with the mouse (toward or away). In addition, arrows were attached to the computer screen to indicate the appropriate direction of response for each stimulus category. This pictorial version of the IAT was piloted successfully with 23 children between the ages of 6 and 11 years. This pictorial version of the IAT, in line with previous research using the IAT (e.g., Greenwald et al., 1998; Monteith et al., 2001), involved a sequence of five blocks, which together allowed for an assessment of childrens uncontrolled association between concepts and an attribute. Block 1 trials introduced the initial concept discrimination and required children to distinguish between unfamiliar Black and White faces by assigning one concept to a response by using an away mouse movement and the other to a response by using a toward mouse movement. The attribute dimension was introduced in Block 2. The children were presented with simple line drawings of cartoon faces and asked to categorize these faces as either happy or sad by using the same response modes as with Block 1. The first two blocks of the IAT were important because they allowed the children to learn the assignments of particular stimuli to certain response modes (i.e., toward or away) to be used in Blocks 3 and 5. Next, the concepts and attributes were superimposed in Block 3. The stimuli for the concepts and attributes appeared in alternate trials within this block. This was termed the stereotype-consistent block because the White faces (ingroup) were paired with happy cartoon faces and the Black faces (out-group) were paired with sad cartoon faces. Block 4 was similar to Block 1 except that the children were presented with the same stimuli as in Block 1 and they responded using the opposite mouse movements. This was important because it

456

Rutland, Cameron, Milne, and McGeorge

allowed the children to learn the response assignments for concepts used in Block 5. Block 5 was called the stereotype-inconsistent block because the target concepts were reversed and combined with the same attribute dimensions as in Block 3. This meant that the target White faces were paired with sad cartoon faces and the target Black faces were paired with happy cartoon faces. Implicit intergroup bias (i.e., an IAT effect) was shown if the White children recorded quicker response times (RTs) in the stereotype-consistent block compared with the stereotype-inconsistent block. An IAT score was calculated by subtracting the RTs in the stereotype-consistent block from the RTs in the stereotype-inconsistent block. This meant that a higher IAT score indicated more implicit bias. The order in which stereotype-consistent and stereotypeinconsistent blocks were presented was counterbalanced, as were response assignments (i.e., toward and away mouse movements). There were 16 trials involving the presentation of stimuli (i.e., racial faces or cartoon faces) within Blocks 1, 2, and 4 (i.e., 8 happy and 8 sad, or 8 Black and 8 White) and 32 trials in the critical stereotype-consistent and stereotype-inconsistent blocks (Blocks 3 and 5). A shorter version of the IAT was used with the 6- to 8-year-olds because there was evidence that the length of the task was creating boredom and fatigue within this age group. In this shorter version of the IAT only 12 trials involved presentation of stimuli (i.e., racial faces or cartoon faces) within Blocks 1, 2, and 4 (i.e., 6 happy and 6 sad, or 6 Black and 6 White). However, there were still 32 trials in the critical stereotypeconsistent and stereotype-inconsistent blocks.

this measure is relatively uninfluenced by spuriously fast or slows RTs (cf. Stuart & Ord, 1987). Nonetheless, analysis was also conducted using the filtering or identification procedures routinely used for full versions of the IAT (see Greenwald et al., 1998). These include (a) recoding latencies under 300 ms to 300 ms and those over 3,000 ms to 3,000 ms; (b) using latencies from all trials (i.e., including latencies on which errors occurred) except for the first two trials of each block, which generally have atypically high latencies; and (c) using logarithm transformations of latencies for statistical significance tests because of their reduced statistical noise. The results using this filtering procedure were not different from those obtained using the childrens median RTs; therefore, only the analyses on the median RTs are reported. Cunningham, Preacher, and Banaji (2001) addressed internal reliability regarding the IAT and found it had satisfactory interitem reliability (a 5 .78) and test retest reliability (a 5 .68). Internal reliability was less of an issue within the study given the use of the median RT.

Results To ascertain whether the pattern of findings fits the hypotheses, a series of analyses of variance (ANOVAs) were conducted with age group (6 8 vs. 10 12 vs. 14 16), public self-focus (high vs. low), and internalization of norm (high vs. low) as between-participants variables. Separate between-participant ANOVAs were conducted on childrens explicit in-group and out-group bias measures. Next, consideration was given to the implicit bias measure; a mixed ANOVA was completed, with IAT block (stereotypical vs. counterstereotypical) as a withinparticipants variable and age group (6 8 vs. 10 12 vs. 14 16), public self-focus (high vs. low), and norm (high vs. low) as between-participants variables. The correlations among variables are presented

Preparation of IAT Data for Analysis Inspection of the response latencies in each trial showed the usual impurities for speeded tasks (Greenwald et al., 1998), with a very small number of extremely fast and extremely slow responses. These outlying scores typically indicate either responses undertaken before presentation of the stimulus (i.e., anticipations) or momentary inattention. The responses in the tails of the latency distribution, although lacking theoretical interest, are troublesome as they distort means and inflate variances. To deal with these few problematic responses (cf. Barnett & Lewis, 1984), a standard accommodation procedure was adopted (e.g., Sroufe, Sonies, West, & Wright, 1973), namely, the use of childrens median RTs within each block. The use of the median RT as a robust estimator is common in psychology because

Table 1 Correlations Among Variables in Study 1 1 1. 2. 3. 4. 5. Age In-group bias Out-group prejudice IAT score Personal normative belief .27 .12 .04 .15 2 3 4 5

.76 .06 .05

.01 .05

.09

Note. IAT 5 Implicit Association Test. po.01.

Childrens Intergroup Attitudes

Table 2 Mean Explicit Racial In-Group Bias (Standard Deviations) for Each Age Group as a Function of Public Self-Focus Public self-focus Age group 6 8 years 10 12 years 14 16 years Total High 0.81 (3.68) n 5 23 1.34 (2.51) n 5 32 0.05 (0.50) n 5 21 0.80 (2.61) Low 3.22 (4.17) n 5 21 0.53 (2.06) n 5 34 0.04 (0.83) n 5 23 1.14 (2.94) Total 2.07 (4.08) n 5 44 0.92 (2.31) n 5 66 0.05 (0.68) n 5 44 n 5 154

457

Note. In-group bias scores could range from 10 (very unfavorable to in-group) to 110 (very favorable to in-group).

in Table 1. Finally, partial correlations were conducted to examine the relationship between implicit and explicit measures.

olds, t(42) 5 2.02, po.05 (Levenes test for equality of variance, F 5 .26, p4.05; see Table 2). Moreover, only the 6- to 8-year-olds in the low public self-focus condition, t(22) 5 3.70, po.001, and the 10- to 12-yearolds in the high public self-focus condition, t(31) 5 3.03, po.01, had in-group bias scores significantly higher than 0. The 6- to 8-year-olds in the low public self-focus condition showed in-group bias, but those in the high public self-focus condition did not. Simple main effects within the public self-focus condition revealed an age effect in the low condition, F(2, 79) 5 10.33, po.001, but not in the high condition. Post hoc analysis in the low public self-focus condition showed that 6- to 8-year-olds demonstrated higher in-group bias than did both 10- to 12year-olds and 14- to 16-year-olds (Tukeys HSD tests, po.05). Table 2 shows that in-group bias decreased with age in the low public self-focus condition, whereas no such age trend was evident in the high public self-focus condition. Explicit Out-Group Prejudice

Explicit In-Group Bias The childrens in-group bias scores were submitted to a 3 (age) 2 (public self-focus) 2 (internalization of norm) between-participants ANOVA. A significant Age Public Self-Focus interaction was found, F(2, 154) 5 4.19, po.05. A main effect for age, F(2, 154) 5 5.62, po.01, was also found that revealed that overall in-group bias became gradually less evident with age (see Table 2). Post hoc analysis indicated that 6- 8-year-olds showed more in-group bias than did 14- 16-year-olds (Tukeys HSD tests, po.05). Simple main effects within each age group revealed that lower in-group bias in the high public self-focus condition compared with the low public self-focus condition was only evident among the 6- to 8-year-

Table 3 Mean Explicit Racial Out-Group Prejudice (Standard Deviations) for Each Norm Group as a Function of Public Self-Focus Public self-focus Norm group High Low Total High 0.23 (1.77) n 5 52 1.05 (4.01) n 5 32 0.48 (3.48) Low 0.13 (3.69) n 5 45 0.91 (3.19) n 5 34 1.54 (2.68) Total 1.86 (2.82) n 5 97 0.16 (3.62) n 5 57 n 5 154

Note. Out-group prejudice scores could range from 110 (very unfavorable to in-group) to 10 (very favorable to in-group).

The childrens out-group prejudice scores were submitted to a 3 (age) 2 (public self-focus) 2 (internalization of norm) between-participants ANOVA. A significant Public Self-Focus Norm interaction was found, F(1, 154) 5 5.09, po.05. Simple effects within each norm group revealed that higher out-group prejudice in the low public selffocus condition compared with the high public selffocus condition was found only among the low norm group, t(55) 5 2.04, po.05 (see Table 3). Children who showed less evidence of norm internalizing were most influenced by the public self-focus manipulation, significantly decreasing their out-group prejudice in the high versus low public self-focus condition. As expected, the ANOVA found a main effect for age, F(2, 154) 5 5.13, po.01, and a main effect for public self-focus, F(1, 154) 5 3.94, po.05. The main effect for age revealed a decrease in out-group prejudice with age: 6- to 8-year-olds (M 5 1.25, SD 5 4.33), 10- to 12-year-olds (M 5 39, SD 5 3.01), and 14- to 16-year-olds (M 5 .05, SD 5 .83). Note, however, that the only significant difference was between the 6- to 8-year-olds and 10- to 12-year-olds (Tukeys HSD tests, po.05). There was evidence of out-group prejudice among the youngest age group as their scores were marginally above 0 (one-sample t test significant at p 5 .06). As expected, the main effect for public self-focus showed that out-group prejudice was higher in the low versus high condition (see Table 3).

458

Rutland, Cameron, Milne, and McGeorge

Implicit Intergroup Bias The childrens median RTs on the stereotype-consistent and stereotype-inconsistent blocks of the IAT were submitted to a 3 (age) 2 (public self-focus) 2 (internalization of norm) 2 (IAT block) mixed design ANOVA, with IAT block as the within-participants variable. A main effect for age was found, F(1, 142) 5 83.74, po.001, as was a main effect for IAT block, F(1, 142) 5 23.29, po.001. There were no other main effects or interactions. The age main effect showed that the childrens responses gradually became quicker with age: 6- to 8-year-olds (M 5 2079.52 ms, SD 5 442.68), 10- to 12-year-olds (M 5 1540.58 ms, SD 5 424.48), and 14- to 16-year-olds (M 5 943.15 ms, SD 5 180.73). Post hoc analysis indicated that all comparisons between age groups were significant (Tukeys HSD tests, po.05). The main effect for IAT block revealed that the childrens responses were significantly quicker on the stereotypeconsistent block (M 5 1433.41 ms, SD 5 590.32) than on the stereotype-inconsistent block (M 5 1605.43 ms, SD 5 607.86). Thus, children in all age groups showed evidence of implicit intergroup bias. Relationship Between Implicit and Explicit Measures The correlations partialing out age between the IAT score and explicit bias measures were all nonsignificant: explicit in-group bias score (r 5 .07) and explicit out-group bias score (r 5 .02). Nonsignificant correlations were evident between the IAT score and the explicit in-group bias measure for 6- to 8-yearolds (r 5 .05), 10- to 12-year-olds (r 5 .12), and 14- to 16-year-olds (r 5 .12). The correlations between the IAT score and the explicit out-group prejudice score were also nonsignificant for 6- to 8-year-olds (r 5 .01), 10- to 12-year-olds (r 5 .08), and 14- to 16-year-olds (r 5 .02). Discussion The findings of this study suggest that social norms and childrens concern for self-presentation affect their explicit racial intergroup attitudes. The children below 10 years old suppressed their explicit in-group bias under high public self-focus. The childrens explicit racial in-group bias was highest among the 6- to 8-year-olds and then declined with age. Though significant, this developmental trend was only evident under low public self-focus. As expected, children with low norm internalization significantly decreased their out-group prejudice under high public self-focus unlike those with high norm

internalization. Finally, children over 10 years showed significant evidence of implicit intergroup bias but not of explicit in-group bias or out-group prejudice. These findings provide partial support for the hypothesis that with age there is a transition from externally motivated conformity to internally motivated bias suppression. The youngest age group was aware of the social norm that racial discrimination is negative behavior and inappropriate. Thus, under high public self-focus they were externally motivated to suppress their in-group bias. In contrast, the older age groups showed no evidence of externally motivated in-group bias suppression. Instead, they showed evidence of internally motivated bias suppress as they showed implicit but not explicit in-group bias. However, a developmental transition from externally motivated to internally motivated suppression was not evident on the out-group prejudice measure. Noticeably, children with low norm internalization irrespective of age were externally motivated to suppress their explicit out-group prejudice under high public self-focus. However, the older children, unlike the youngest age group, showed evidence of internal motivation to suppress their out-group prejudice as they evidenced implicit but not explicit prejudice. These findings suggest that whether children are externally motivated to suppress in-group bias is dependent on their age, whereas the suppression of out-group prejudice is more dependent on the degree to which children have internalized the norm against prejudice. As predicted, the level of norm internalization moderated the affect of public self-focus on outgroup prejudice. This finding is compatible with research on adults that suggests that the expression and suppression of prejudice closely follows normative beliefs held within a particular society (Crandall et al., 2002; Sherif, 1936; Turner, 1991). Childrens suppression of their prejudice may also depend on the prevailing normative context regarding the acceptability of any given form of prejudice. Therefore, children are unlikely to show prejudice suppression when the social norm within their environment is to tolerate a particular form of prejudice (e.g., prejudice based on sexual orientation or nationality). As noted earlier, a preliminary study of norms prevalent in British childrens social environment indicated that the norm against bias based on racial group membership was stronger than that for national group membership, especially regarding Germans. Thus, the findings of our preliminary study and previous research (Bennett, Lyons, Sani, & Barrett, 1998; Rutland, 1999; Verkuyten, 2001) suggest that British children are less likely to engage in self-presentation and control their explicit national prejudice.

Childrens Intergroup Attitudes

459

Study 2 To test whether British children suppress their explicit national intergroup bias, a second study was conducted replicating the basic design of our first experiment except with an emphasis on childrens national rather than racial attitudes. This second study examined childrens attitudes toward the British in-group and German out-group. Germans were chosen because previous research has shown that British school children perceive them as a salient out-group and are willing to express intergroup bias toward this group (Abrams, Rutland, & Cameron, 2003; Barrett, Wilson, & Lyons, 2003; Rutland, 1999). As in Study 1, childrens personal normative beliefs regarding out-group prejudice were measured. It was expected that children would judge national discrimination as more tolerable, and therefore less punishable, than racial discrimination. Thus, in Study 2 there should be less evidence of explicit national bias suppression. Specifically, the childrens explicit national in-group and out-group attitudes should be unaffected by the public self-focus manipulation. In accordance with previous research, it was anticipated that children would show less national intergroup bias with age (e.g., Barrett et al., 2003; Lambert & Klineberg, 1967). However, given the salient intergroup rivalries between Germany and Britain (see Abrams, Rutland, & Cameron, 2003; Rutland, 1999, 2004), significant intergroup bias was anticipated into early adolescence (Teichman, 2001). In early adolescence, when self-identity is a central issue (Erikson, 1968) and the collective self is also gaining importance (Quintana, 1999; Ruble et al., 2004), individuals should favor negative characterizations of out-groups as a means of self-enhancement. In middle adolescence this motivation may be less evident given the emergence of a more established and secure sense of self (Phinney, 1990). Finally, as in Study 1, dissociation between implicit and explicit intergroup attitudes and implicit bias at all ages was predicted. Method Participants and Design Participants were 134 White British children (33 males, 90 females) from the same demographic background as the children in Study 1. Included were forty-six 6- to 8-year-olds (M 5 7 years 8 months, SD 5 11 months), forty-five 10- to 12-yearolds (M 5 11 years 0 months, SD 5 8 months), and forty-three 14- to 16-year-olds (M 5 14 years 11 months, SD 5 14 months).

Each testing session involved the same procedure and measures as Study 1. Stimulus Materials and Measures Personal normative belief measure. This measure was described in Study 1. However, within the vignettes, British (instead of White) boys or girls excluded two German (instead of Black) boys or girls because of their group membership. As in Study 1, the childrens personal normative belief scores were strongly skewed, with 4% responding OK, 7% responding bad, 34% responding very bad, and 55% responding very very bad. The childrens responses were also equally skewed in each age group. Therefore, a median split was performed on the childrens responses classifying children who responded OK, bad, and very bad into the low norm internalization group (n 5 61, 45%) and the children who responded very very bad into the high norm internalization group (n 5 73, 55%). There were equivalent distributions of children from each age group in the low and high norm groups. Explicit intergroup bias measure. This measure was the same as used in Study 1, though with a focus on national attitudes. The children were presented with two cups labeled British and German and marked with the British and German flags. Implicit intergroup bias measure. This measure was described in Study 1. To measure implicit national bias, neutral faces different from Study 1 were used. Unfamiliar White faces were used with either the British or German flags superimposed on the face. A pilot study with 6- to 10-year-olds indicated that most children were familiar with both the British and German flags. Nevertheless, before testing, all children were shown large (29 20 cm) pictures of the British and German flags and were explicitly told which country each flag represented. Results Initially, a series of ANOVAs were conducted to examine whether the childrens ratings in the social inclusion vignettes differed between Study 1 and Study 2. Next, separate ANOVAs were conducted on the childrens explicit in-group and out-group bias measures, with age group (6 8 vs. 10 12 vs. 14 16), public self-focus (high vs. low), and internalization of norm (high vs. low) as between-participants variables. Regarding the implicit bias measure, a mixed ANOVA was completed, with IAT block (stereotypical vs. counterstereotypical) as a within-participants variable and age group (6 8 vs. 10 12 vs. 14 16), public

460

Rutland, Cameron, Milne, and McGeorge

Table 4 Correlations Among Variables in Study 2 1 1. 2. 3. 4. 5. Age In-group bias Out-group prejudice IAT score Personal normative belief .26 .11 .20 .04 2 3 4 5

Explicit In-Group Bias The childrens in-group bias scores were submitted to a 3 (age) 2 (public self-focus) 2 (internalization of norm) between-participants ANOVA. This analysis found a significant Age Public Self-Focus interaction, F(2, 132) 5 5.21, po.01, and a main effect for age, F(2, 132) 5 4.92, po.01. The main effect of age showed that in-group bias became less apparent with age (see Table 5). Post hoc analysis indicated that 14- to 16-year-olds showed less in-group bias than did both 6- to 8-year-olds and 10- to 12-yearolds (Tukeys HSD tests, po.05). Simple main effects analysis showed lower in-group bias in the high public self-focus condition than in the low public self-focus condition among 6- to 8-year-olds, t(44) 5 2.21, po.05 (Levenes test for equality of variance, F 5 .08, p4.05). Lower in-group bias under high public self-focus versus low public self-focus was also evident among the 14- to 16-year-olds, t(41) 5 2.35, po.05 (Levenes test for equality of variance, F 5 4.67, po.05). In contrast, among the 10to 12-year-olds, higher in-group bias was shown in the high public self-focus condition versus the low public self-focus condition, t(42) 5 2.08, po.05 (Levenes test for equality of variance, F 5 2.50, p4.05; see Table 5). Only the 6- to 8-year-olds in the low public self-focus condition, t(24) 5 4.70, po.001, and the 10- to 12-year-olds in the high, t(21) 5 4.22, po.001, and low, t(21) 5 2.45, po.05, public selffocus conditions had in-group bias scores significantly higher than zero. This demonstrates that although 6- to 8-year-olds showed in-group bias in the low public self-focus condition, this bias disappeared under high public self-focus. The 10- to 12year-olds, on the contrary, showed in-group bias under low public self-focus, and heightening public

.62 .02 .09

.03 .08

.12

Note. IAT 5 Implicit Association Test. po.05. po.001.

self-focus (high vs. low), and norm (high vs. low) as between-participants variables. The correlations among variables are presented in Table 4. Finally, partial correlations were conducted to examine the relationship between implicit and explicit measures.

Personal Normative Beliefs Two 3 (age) 2 (study: racial, national) betweenparticipants ANOVAs were conducted, one using badness and one using punishment as the dependent variable. In the former analysis no main effects or interactions were found. In contrast, the latter analysis, a significant Age Study interaction was found, F(2, 266) 5 3.59, po.05. Simple main effects within each age group showed that only the two oldest age groups rated social exclusion based on nationality as significantly less punishable than that based on racial group membership: 10- to 12-yearolds, t(99) 5 2.18, po.05 (M 5 2.11, SD 5 .68; M 5 2.42, SD 5 .68 for nationality-and racial-based exclusion, respectively), and 14- to 16-year-olds, t(84) 5 2.03, po.05 (M 5 1.71, SD 5 .64; M 5 2.02, SD 5 .76, for nationality-and racial-based exclusion, respectively). In contrast, the 6- to 8-year-olds perceived no significant difference between the level of punishment appropriate for social exclusion based on national and racial group membership (M 5 2.47, SD 5 .66; M 5 2.27, SD 5 .78, for nationality-and racial-based exclusion, respectively). A main effect for age, F(2, 266) 5 11.90, po.01, also revealed that with age children thought the characters should experience less punishment: 6- to 8-year-olds (M 5 2.35, SD 5 .73), 10- to 12-year-olds (M 5 2.32, SD 5 .69), and 14- to 16-year-olds (M 5 1.87, SD 5 .72). Post hoc analysis showed the 14- to 16-year-olds thought the story characters should be punished significantly less than did the 6- to 8-year-olds and 10- to 12-yearolds (Tukeys HSD tests, po.05).

Table 5 Mean Explicit National In-Group Bias (Standard Deviations) for Each Age Group as a Function of Public Self-Focus Public self-focus Age group 6 8 years 10 12 years 14 16 years Total High 0.67 (4.29) n 5 21 3.00 (3.29) n 5 22 0.32 (0.84) n 5 22 1.11 (3.38) Low 3.16 (3.36) n 5 25 1.18 (2.26) n 5 22 0.70 (1.81) n 5 20 1.75 (2.79) Total 2.02 (3.98) n 5 46 2.07 (2.93) n 5 44 0.05 (0.68) n 5 42 n 5 132

Note. In-group bias scores could range from 10 (very unfavorable to in-group) to 110 (very favorable to in-group).

Childrens Intergroup Attitudes

461

self-focus only increased in-group bias. The 14- to 16year-olds showed no evidence of in-group bias in either public self-focus condition. Explicit Out-Group Prejudice The childrens out-group prejudice scores were submitted to a 3 (age) 2 (public self-focus) 2 (internalization of norm) between-participants ANOVA. As expected, there were no significant main effects or interactions. Moreover, the out-group prejudice scores of the 6- to 8-year-olds (M 5 1.13), 10- to 12-year-olds (M 5 .80), and 14- to 16-yearolds (M 5 .35) were not significantly higher than zero. These findings indicate that at no age were the children showing out-group prejudice. Furthermore, as predicted, the childrens prejudice ratings were unaffected by the public-self manipulation and were unrelated to their personal normative beliefs. Implicit Intergroup Bias The childrens median RTs on the stereotypeconsistent (British1happy/German1sad) and stereotype-inconsistent (British1sad/German1happy) blocks of the IAT were submitted to a 3 (age) 2 (public self-focus) 2 (internalization of norm) 2 (IAT block) ANOVA, with IAT block as the withinparticipants variable. This analysis found a main effect for age, F(2, 120) 5 51.67, po.001, and a main effect for IAT block, F(1, 120) 5 35.18, po.001. No other main effects or interactions were found. The age main effect showed that the childrens RTs gradually became quicker with age: 6- to 8-year-olds (M 5 1778.25 ms, SD 5 524.65), 10- to 12-year-olds (M 5 1340.54 ms, SD 5 456.48), and 14- to 16-yearolds (M 5 861.81 ms, SD 5 157.47). Post hoc analysis indicated that all comparisons between age groups were significant (Tukeys HSD tests, po.05). The main effect for IAT block indicated that the childrens RTs were significantly quicker on the stereotype-consistent block (M 5 1233.24 ms, SD 5 543.11) than on the stereotype-inconsistent block (M 5 1444.63 ms, SD 5 628.95). Thus, children in all age groups showed evidence of implicit intergroup bias. Relationship Between Implicit and Explicit Measures The relationships between explicit and implicit measures were examined through correlations partialing out age. Overall, the IAT score was not associated with the explicit intergroup bias (r 5 .05),

in-group bias (r 5 .00), and out-group bias (r 5 .10) scores. Nonsignificant correlations were evident between the IAT score and the explicit in-group bias measure for 6- to 8-year-olds (r 5 .05), 10- to 12-yearolds (r 5 .17), and 14- to 16-year-olds (r 5 .14). The correlations between the IAT score and the explicit out-group prejudice score were also nonsignificant for 6- to 8-year-olds (r 5 .09), 10- to 12-yearolds (r 5 .07), and 14- to 16-year-olds (r 5 .10). Discussion The two oldest age groups perceived racial prejudice as warranting more punishment than national prejudice. In contrast, the youngest age group saw these two forms of prejudice as equally punishable. This finding seems compatible with research that suggests that children below 8 years old are relatively less attentive than older children to what is normative when engaging in social reasoning about inclusion and exclusion (Abrams, Rutland, & Cameron 2003; Abrams, Rutland, Cameron, et al., 2003; Killen, Crystal, & Watanabe, 2002; Killen & Stangor, 2001). As predicted, children irrespective of age did not show evidence of self-presentation by controlling their explicit national out-group prejudice under high public self-focus. This finding contrasts with the result from Study 1, where children with low norm internalization showed external motivation to control their prejudice by suppressing their racial out-group bias under high public self-focus. These findings combined suggest that childrens bias suppression is dependent on the prevailing norms within the childs social environment. Similar to Study 1 and counter to our prediction, the 6- to 8-year-olds showed significantly less ingroup bias under high public self-focus versus low public self-focus. The young children were aware of the social norm that national bias is a relatively inappropriate behavior. Thus, in line with Study 1, they were externally motivated to control their in-group bias. Notably, the youngest children, unlike the oldest children, showed significant in-group bias under low public self-focus and inhibited this bias under high public self-focus. The oldest age group failed to show significant in-group bias in either public self-focus condition. Therefore, it is reasonable to conclude that only the youngest age group showed significant suppression of explicit in-group bias. In contrast, among the 10- to 12-year-olds, heightening public self-focus only increased national in-group bias. This finding is particularly interesting because it suggests that increasing public accountability

462

Rutland, Cameron, Milne, and McGeorge

can actually increase in-group bias. Previous research has shown that when the norm is to tolerate discrimination and the context involves salient intergroup rivalry, early adolescents can show high levels of intergroup bias (Abrams, Rutland, & Cameron, 2003; Black-Gutman & Hickson, 1996; Rutland, 1999; Teichman, 2001). The 10- to 12-yearolds seemed internally motivated to express ingroup bias, as they showed significant national ingroup bias even in the low public self-focus condition. Heightening public accountability may only have introduced external pressure in line with their internal motivation, resulting in increased national in-group bias. Notably, higher public accountability did not increase national out-group prejudice among 10- to 12-year-olds. This finding is compatible with the literature that suggests the expression of ingroup bias is both more common and acceptable than out-group prejudice (Aboud, 2003; Brewer, 1979; Cameron et al., 2001; Mummendey & Otten, 1998; Nesdale, 2004). As expected, the children showed significantly less intergroup bias with age, though they were still showing significant intergroup bias in early adolescence. Finally, as in Study 1, children of all ages showed implicit intergroup bias and they showed no dissociation between the implicit and explicit measures. General Discussion Taken together, the two studies provide support for the claim that social norms and childrens concern about self-presentation influence their intergroup attitudes. Both studies showed that children below 10 years old were externally motivated to control their in-group bias under high public self-focus. Study 1 also demonstrated that children with low norm internalization (i.e., those externally motivated to control prejudice) suppressed their racial prejudice under high public self-focus. In contrast, as predicted based on our preliminary study, children in Study 2 did not suppress their explicit national prejudice under high public self-focus. In fact, the 10- to 12-year-olds increased their national in-group bias under high public-self focus. These findings indicate that suppression of out-group prejudice is closely related to social norms in the childrens social environment. The studies found little evidence supporting the developmental prediction that only older children would be externally motivated to suppress their intergroup bias under high public accountability. Instead, the older children showed evidence of internally motivated bias suppression because they

showed implicit but not explicit intergroup bias. The developmental prediction of a transition with age from externally to internally motivated bias suppression found some support. In both studies, the youngest children were aware that discrimination was inappropriate and suppressed their in-group bias under heightened public accountability. Furthermore, one might expect that changes in norm internalization would mediate the effect of public self-focus on childrens in-group bias. Unfortunately, the nature of our data set made it impossible to perform full mediation analysis because the childrens responses on the normative belief measure were strongly skewed and age was not a continuous factor (i.e., there were no 9-, 12-, or 13-year-olds in each study). Future studies need to investigate directly this possible mediation process. This research should avoid measuring explicitly childrens normative beliefs because of social desirability concerns that are likely to result in skewed responses. Instead, future studies could adopt a minimal group paradigm (e.g., Bigler, Jones, & Lobliner, 1997; Nesdale & Flesser, 2001) and attempt to manipulate directly childrens norms regarding in-group and out-group bias (e.g., Ojala & Nesdale, 2004). Evidence for a developmental transition from externally to internally motivated bias suppression was evident only for the in-group attitude measures. The young children were not externally motivated to suppress their out-group prejudice in either study. As expected, there was no evidence of externally motivated out-group prejudice suppression in Study 2, and the suppression of out-group prejudice in Study 1 was dependent on the degree to which the children had internalized the norm against prejudice. These contrasting findings on the in-group and out-group attitude measures need explanation. One explanation might be that there exist different social norms regarding the acceptability of in-group bias and out-group prejudice. Research suggests that ingroup bias is both more common and tolerated than is out-group prejudice (Aboud, 2003; Brewer, 1979, 1999; Hewstone, Rubin, & Willis, 2002; Mummendey & Otten, 1998; Nesdale, 2004). Young children may have internalized less than the older children the norm against in-group bias; therefore, under high accountability they were more likely to be externally motivated to suppress their in-group bias. In contrast, children irrespective of age may have internalized the strong norm against out-group prejudice. Thus, only children in the low norm internalization group were likely to be externally motivated to suppress their out-group prejudice. Further research

Childrens Intergroup Attitudes

463

is required to validate this explanation and should independently examine childrens social norms regarding in-group bias and out-group prejudice. Social identity theorists maintain that self-presentational concerns may drive intergroup attitudes when individuals wish to win a positive self-identity within their social group (Abrams, 1994; Barreto & Ellemers, 2000). However, the present studies were unable to examine this proposition without including identification measures. Future studies should examine how in-group identification might moderate the affect of social norms on childrens self-presentation of intergroup attitudes. Another potential extension of this research would be the inclusion of a private self-focus manipulation (Wicklund & Duval, 1971), such as seeing ones own mirror image. The present studies found evidence that older children are internally motivated to suppress their bias because they showed implicit but not explicit intergroup bias. A private self-focus manipulation should result in children who are internally motivated to suppress their intergroup bias. The developmental account of a transition from externally to internally motivated bias suppression, partially supported by our studies, acknowledges the importance of sociocognitive skills. These sociocognitive skills include empathy, perspective taking, and logical and counterfactual reasoning, as these are important in the internalization of norms (Crandall et al., 2002). Adult research on the suppression of prejudice also indicates the important role of self-conscious emotions such as guilt and shame (e.g., Monteith, Devine, & Zuwerink, 1993). Future research into childrens social norms and self-presentation of intergroup attitudes should include measures of these sociocognitive and emotional factors. The dissociation found between implicit and explicit measures suggests that children were engaging in self-presentation. The findings indicate that from an early age children showed implicit intergroup bias, though whether these biases were explicitly expressed depended on social norms and childrens motivation to control their bias. These two studies are, to our knowledge, the first successful attempts at adapting the IAT to measure implicit attitudes among young children. Future research should attempt to include both explicit and implicit measures, thus providing more subtle and socially sensitive measures of intergroup attitudes in children. There are positive implications for childhood prejudice reduction from the finding that social norms and childrens concern about self-presentation influence their intergroup attitudes. Instead of facing the challenge of changing individually childrens

attitudes, these studies suggest that changing the normative climate in childrens social environment can induce significant attitude change (Crandall et al., 2002; Monteith, 1993). Indeed, when children identify with a desirable group in their environment that condemns a prejudice, they are more likely to succeed in the battle for norm internalization. To conclude, our findings indicate that social norms and childrens concern for self-presentation affect their intergroup attitudes. There was evidence of racial prejudice suppression in Study 1, especially among children with lower levels of norm internalization. As expected, Study 2 provided no evidence of national prejudice suppression. Furthermore, there was some evidence for a developmental transition from externally to internally motivated bias suppression. Overall, the findings of the two studies suggest an original developmental account of intergroup attitudes that acknowledges that children are sensitive to social norms and concerned with self-presentation.

References
Aboud, F. E. (1988). Children and prejudice. Oxford, England: Blackwell. Aboud, F. E. (2003). The formation of in-group favoritism and out-group prejudice in young children: Are they distinct attitudes? Developmental Psychology, 39, 48 60. Abrams, D. (1994). Social self-regulation. Personality and Social Psychology Bulletin, 20, 473 484. Abrams, D., & Brown, R. J. (1989). Self-consciousness and social identity: Self-regulation as a group member. Social Psychology Quarterly, 52, 311 318. Abrams, D., Rutland, A., & Cameron, L. (2003). The development of subjective group dynamics: Childrens judgments of normative and deviant in-group and outgroup individuals. Child Development, 74, 1840 1856. Abrams, D., Rutland, A., Cameron, L., & Marques, J. M. (2003). The development of subjective group dynamics: When in-group bias gets specific. British Journal of Developmental Psychology, 21, 155 176. Asch, S. E. (1956). Studies of interdependence and conformity: A minority of one against a unanimous majority. Psychological Monographs: General and Applied, 70, 1 70. Barnett, V., & Lewis, T. (1984). Outliers in statistical data (2nd ed.). New York: Wiley. Barreto, M., & Ellemers, N. (2000). You cant always do what you want: Social identity and self-presentational determinants of the choice to work for a low-status group. Personality and Social Psychology Bulletin, 26, 891 906. Barrett, M., Wilson, H., & Lyons, E. (2003). The development of national in-group bias: English childrens attributions of characteristics to English, American and German people. British Journal of Developmental Psychology, 21, 193 220.

464

Rutland, Cameron, Milne, and McGeorge pipeline? Journal of Personality and Social Psychology, 69, 1013 1027. Fenigstein, A., Schier, M. F., & Buss, A. H. (1975). Public and private self-consciousness: Assessment and theory. Journal of Consulting and Clinical Psychology, 43, 522 527. Graves, S. B. (1999). Television and prejudice reduction: When does television as a vicarious experience make a difference? Journal of Social Issues, 55, 707 728. Greenwald, A. D., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The Implicit Association Test. Journal of Personality and Social Psychology, 74, 1464 1480. Hewstone, M., Rubin, M., & Willis, H. (2002). Intergroup bias. Annual Review of Psychology, 53, 575 604. Hockstader, L. (2001, September 5). At Arab, Israeli schools, hatred is common bond: Conflict hardens 8thgraders stereotypes. Washington Post, p. A1. Jahoda, G., Thomson, S. S., & Bhatt, S. (1972). Ethnic identity and preferences among Asian immigrant children in Glasgow: A replicated study. European Journal of Social Psychology, 2, 19 32. Jetten, J., Spears, R., & Manstead, A. S. R. (1997). Strength of identification and intergroup differentiation: The influence of group norms. European Journal of Social Psychology, 27, 603 609. Katz, P. A., Sohn, M., & Zalk, S. R. (1975). Perceptual concomitants of racial attitudes in urban grade-school children. Developmental Psychology, 11, 135 144. Kelman, H. C. (1958). Compliance, identification and internalization: Three processes of attitude change. Journal of Conflict Resolution, 2, 51 60. Killen, M., Crystal, D. S., & Watanabe, H. (2002). Japanese and American childrens evaluations of peer exclusion, tolerance of differences, and prescriptions for conformity. Child Development, 73, 1788 1802. Killen, M., Lee-Kim, J., McGlothlin, H., & Stangor, C. (2002). How children and adolescents evaluate gender and racial exclusion. Monographs for the Society for Research in Child Development, 67(4, Serial No. 271). Killen, M., Pisacane, K., Lee-Kin, J., & Ardila-Rey, A. (2001). Fairness or stereotypes? Young childrens priorities when evaluating group exclusion and inclusion. Developmental Psychology, 37, 587 596. Killen, M., & Stangor, C. (2001). Childrens social reasoning about inclusion and exclusion in gender and race peer groups contexts. Child Development, 72, 174 186. Kim, D. Y. (2003). Voluntary controllability of the Implicit Association Test (IAT). Social Psychology Quarterly, 66, 83 96. Lambert, W. E., & Klineberg, O. (1967). Childrens views of foreign peoples: A cross-national study. New York: Appleton-Century-Crofts. Lawrence, V. W. (1991). Effect of socially ambiguous information on White and Black childrens behavioral and trait perceptions. Merrill-Palmer Quarterly, 37, 619 630. Levy, S. R., & Troise, D. M. (2001, April). Childrens social concerns with appearing prejudiced. In C. McKown

Bennett, M., Lyons, E., Sani, F., & Barrett, M. (1998). Childrens subjective identification with the group and ingroup favoritism. Developmental Psychology, 34, 902 909. Bigler, R. S. (1999). The use of multicultural curricula and materials to counter racism in children. Journal of Social Issues, 55, 687 705. Bigler, R. S., Jones, L. C., & Lobliner, D. B. (1997). Social categorization and the formation of intergroup attitudes in children. Child Development, 68, 530 543. Bigler, R. S., & Liben, L. S. (1993). A cognitive-developmental approach to racial stereotyping and reconstructive memory in Euro-American children. Child Development, 64, 1507 1518. Black-Gutman, D., & Hickson, F. (1996). The relationship between racial attitudes and social-cognitive development in children: An Australian study. Developmental Psychology, 32, 448 456. Blanchard, F. A., Crandall, C. S., Brigham, J. C., & Vaughan, L. A. (1994). Condemning and condoning racism: A social context approach to interracial settings. Journal of Applied Psychology, 79, 993 997. Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A cognitive-motivational analysis. Psychological Bulletin, 86, 307 324. Brewer, M. B. (1999). The psychology of prejudice: Ingroup love or out-group hate? Journal of Social Issues, 55, 429 444. Brown, R. J. (1995). Prejudice: Its social psychology. Oxford, England: Blackwell. Cameron, J. A., Alvarez, J. M., Ruble, D., & Fuligni, A. J. (2001). Childrens lay theories about in-groups and outgroups: Reconceptualizing research on prejudice. Personality and Social Psychology Review, 5, 118 128. Crandall, C. S., Eshleman, A., & OBrien, L. (2002). Social norms and the expression and suppression of prejudice: The struggle for internalization. Journal of Personality and Social Psychology, 82, 359 378. Crandall, V. C., Crandall, V. J., & Katkovsky, W. (1965). A childrens social desirability questionnaire. Journal of Consulting Psychology, 29, 27 36. Cunningham, W. A., Preacher, K. J., & Banaji, M. R. (2001). Implicit attitude measures: Consistency, stability and convergent validly. Psychological Science, 12, 163 170. Dovidio, J. F., & Gaertner, S. L. (1991). Changes in the expression and assessment of racial prejudice. In H. J. Knopke, R. J. Norrell, & R. W. Rogers (Eds.), Opening doors: Perspectives on race relations in contemporary America (pp. 119 148). Tuscaloosa: University of Alabama Press. Doyle, A., Beaudet, J., & Aboud, F. (1988). Developmental patterns in the flexibility of childrens ethnic attitudes. Journal of Cross Cultural Studies, 19, 3 18. Erikson, E. H. (1968). Identity: Youth and crisis. Oxford, England: Norton. Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic activation as an unobstrusive measure of racial attitudes: A bona fide

Childrens Intergroup Attitudes (Chair), The development and consequences of stereotype processes in childhood. Symposium conducted at the biennial meeting of the Society for Research in Child Development, Minneapolis, MN. Macrae, C. N., Bodenhausen, G. V., & Milne, A. B. (1998). Saying no to unwanted thoughts: Self-focus and the regulation of mental life. Journal of Personality and Social Psychology, 74, 578 589. McLeod, K. A. (Ed.). (1993). Multicultural education: The state of the art. Toronto, Canada: University of Toronto. Monteith, M. J. (1993). Self-regulation of prejudiced responses: Implications for progress in prejudice-reduction efforts. Journal of Personality and Social Psychology, 65, 469 485. Monteith, M. J., Deneen, N. E., & Tooman, G. D. (1996). The effect of social norm activation on the expression of opinions concerning gay men and Blacks. Basic and Applied Social Psychology, 18, 267 288. Monteith, M. J., Devine, P. G., & Zuwerink, J. R. (1993). Selfdirected versus other-directed affect as a consequence of prejudice-related discrepancies. Journal of Personality and Social Psychology, 64, 198 210. Monteith, M. J., Voils, C. I., & Ashburn-Nardo, L. (2001). Taking a look underground: Detecting, interpreting, and reacting to implicit racial biases. Social Cognition, 19, 395 417. Mummendey, A., & Otten, S. (1998). Positive-negative asymmetry in social discrimination. In W. Streobe & M. Hewstone (Eds.), European review of social psychology (Vol. 9, pp. 107 143). Chichester: Wiley. Nesdale, D. (2001). Development of prejudice in children. In M. Augoustinos & K. J. Reynolds (Eds.), Understanding prejudice, racism and social conflict (pp. 57 72). London: Sage. Nesdale, D. (2004). Social identity processes and childrens ethnic prejudice. In M. Bennett & F. Sani (Eds.), The development of the social self (pp. 219 245). East Sussex, England: Psychology Press. Nesdale, D., & Flesser, D. (2001). Social identity and the development of childrens group attitudes. Child Development, 72, 506 517. Nucci, L. (1981). Conceptions of personal issues: A domain distinct from moral or societal concepts. Child Development, 52, 114 121. Ojala, K., & Nesdale, D. (2004). Bullying and social identity: The effects of group norms and distinctiveness threat on attitudes towards bullying. British Journal of Developmental Psychology, 22, 19 35. Pettigrew, T. F. (1958). Personality and sociocultural factors in intergroup attitudes: A cross-cultural comparison. Journal of Conflict Resolution, 2, 29 42. Phinney, J. S. (1990). Ethnic identity in adolescents and adults: Review of research. Psychological Bulletin, 108, 499 514. Piaget, J. (1965). The moral judgment of the child. New York: Free Press. Plant, E. A., & Devine, P. G. (1998). Internal and external motivation to respond without prejudice. Journal of Personality and Social Psychology, 75, 811 832.

465

Powlishta, K., Serbin, L. A., Doyle, A., & White, D. R. (1994). Gender, ethnic, and body type biases: The generality of prejudice in childhood. Developmental Psychology, 30, 526 536. Quintana, S. M. (1994). A model of ethnic perspective taking ability applied to Mexican-American children and youth. International Journal of Intercultural Relations, 18, 419 448. Quintana, S. M. (1998). Childrens developmental understanding of ethnicity and race. Applied and Preventative Psychology, 7, 27 45. Quintana, S. M. (1999). Role of perspective-taking abilities and ethnic socialization in development of adolescent ethnic identity. Journal of Research on Adolescence, 19, 161 184. Ruble, D. N., Alvarez, J., Bachman, M., & Cameron, J. (2004). The development of a sense of we: The emergence and implications of childrens collective identity. In M. Bennett & F. Sani (Eds.), The development of social self (pp. 29 76). East Sussex, England: Psychology Press. Rutland, A. (1999). The development of national prejudice, in-group favoritism and self stereotypes in British children. British Journal of Social Psychology, 38, 55 70. Rutland, A. (2004). The development and self-regulation of intergroup attitudes in children. In M. Bennett & F. Sani (Eds.), The development of the social self (pp. 247 265). East Sussex, England: Psychology Press. Rutland, A., Brown, R. J., Ahmavaara, A., Arnold, K., & Samson, J. (2003, August). Developmental trends in childrens national in-group and out-group attitudes. Poster presented at the 11th European Conference on Developmental Psychology, Catholic University, Milano, Italy. Selman, R. L. (1971). Taking anothers perspective: Roletaking development in early childhood. Child Development, 42, 1721 1734. Selman, R. L. (1980). The growth of interpersonal understanding: Development and clinical analyses. San Diego, CA: Academic Press. Sherif, M. (1936). The psychology of social norms. New York: Harper. Smetana, J. G. (1981). Preschool childrens conceptions of moral and social rules. Child Development, 52, 1333 1336. Smetana, J. G., & Braeges, J. L. (1990). The development of toddlers moral and conventional judgments. MerrillPalmer Quarterly, 36, 329 346. Smetana, J. G., Kelly, M., & Twentyman, C. T. (1984). Abused, neglected, and nonmaltreated childrens conceptions of moral and social-conventional transgressions. Child Development, 55, 277 287. Sroufe, L. A., Sonies, B. C., West, W. D., & Wright, F. S. (1973). Anticipatory heart rate deceleration and reaction time in children with and without referral for learning disability. Child Development, 44, 267 273. Stuart, A., & Ord, J. K. (1987). Kendalls advanced theory of statistics: Vol. 1. Distributional theory. London: Griffin. Teichman, Y. (2001). The development of Israeli childrens images of Jews and Arabs and their expression in

466

Rutland, Cameron, Milne, and McGeorge Scottish (M 5 7.15, SD 5 2.27), and English (M 5 7.31, SD 5 2.37). Pairwise comparisons showed that all differences were significant at po.05.

human figure drawings. Developmental Psychology, 37, 749 761. Theimer, C. E., Killen, M., & Stangor, C. (2001). Young childrens evaluations of exclusion in gender-stereotypic peer contexts. Developmental Psychology, 37, 18 27. Turner, J. C. (1991). Social influence. Pacific Grove, CA: Brooks/Cole. Verkuyten, M. (2001). National identification and intergroup evaluations in Dutch children. British Journal of Developmental Psychology, 19, 559 571. Wicklund, R. A., & Duval, S. (1971). Opinion change and performance facilitation as a result of objective selfawareness. Journal of Experimental Social Psychology, 7, 319 342. Williams, J. E., Best, D. L., Boswell, D. A., Mattson, L. A., & Graves, D. J. (1975). Preschool racial attitude measure II. Educational and Psychological Measurement, 35, 3 18.

Public Self-Focus Manipulation Check

To check the success of the public self-focus manipulation, at the end of testing, a subsample of the 10- to 16-yearolds (n 5 33) completed a public self-consciousness 10-item scale. This scale was modeled on one originally used by Fenigstein et al. (1975) to measure public self-consciousness. Minor alterations were made to the wording of the items used by Fenigstein et al., making the items more accessible to children. In addition, the wording of the items was altered so that children responded according to how they were feeling at that time (i.e., immediately after testing had taken place). This scale has been used with children aged 10 years and above (e.g., Abrams & Brown, 1989). The children responded by putting a mark on a 3.5cm line to show how they felt. The line went from a lot (scored 7) to not at all (scored 0). Each childs item score was the distance between their response and the not at all end of the line. The childrens responses were subjected to reliability analysis resulting in a Cronbachs alpha of .72. Given this satisfactory degree of reliability, the childrens responses were collapsed to form a single scale by calculating a composite mean for public self-consciousness (the higher each childs mean score, the higher is the childs public self-consciousness). As expected, childrens public selfconsciousness scores were significantly higher in the high public self-focus condition (M 5 5.95, SD 5 .58) than in the low self-focus condition (M 5 5.13, SD 5 1.11), t(31) 5 2.67, po.05.

Appendix Preliminary Study of Normative Context

This study involved the adaptation of a technique used by Macrae et al. (1998) to measure the appropriateness of judgments based on different social categories. One hundred and eighteen 14- to 16-year-olds were asked to rate the appropriateness within their society of judgments formed solely on the basis of a persons social group membership. Participants responded on a 9-point scale ranging from 1 (appropriate) to 9 (inappropriate). Thirty-five social groups were used. These included two racial groups (Black and Asian), five national groups (French, German, Scottish, American, and Australian), and the in-group categories (White and English). The remaining 26 groups were filler items (e.g., criminals, teachers, athletes, rich people). Participants responses to the two racial groups and five national groups were subjected to reliability analyses, yielding Cronbachs alphas of .80 and .89, respectively. Given these satisfactory reliability coefficients, composite scores for both the racial and the national groups were computed. The composite score for appropriateness of judging someone based on their racial group (M 5 7.12, SD 5 2.10) was significantly higher than the composite score for the national groups (M 5 6.75, SD 5 2.10), t(116) 5 2.30, po.05. This finding shows that children thought it extremely inappropriate to judge someone on the basis of their racial group membership. Moreover, they deemed it significantly less inappropriate to judge someone on the basis of their nationality. The children perceived that judging the Germans (M 5 6.16, SD 5 2.90) solely on the basis of their nationality was significantly less inappropriate than judging Americans (M 5 6.60, SD 5 2.50), French (M 5 6.54, SD 5 2.68), Australians (M 5 7.21, SD 5 2.18),

Vignettes Used to Measure Personal Normative Belief

Female version: It was lunchtime and three girls were at a school dining table eating their lunch. There were two extra seats at the table. Next Latisha and Ebony, who were black girls, came up to the girls and asked if they could sit down beside them in the extra seats. The girls said no to Latisha and Ebony, and explained that they do not like mixing with Black people. This upset Latisha and Ebony. Male version: One day, in the school ground, three boys wanted to play a 5-a-side soccer game. However, they could not because they needed two extra boys. Next Malik and Lamar, who were Black boys, came up to the boys and asked if they could join in. The boys said no to Malik and Lamar, and explained that they do not like playing with Black people. This upset Malik and Lamar.