Self-Identity and Prejudice in Children

Howard J. Ehrlich, Ph.D.*

The Prejudice Institute
2743 Maryland Avenue
Baltimore, MD 21218
www.prejudiceinstitute.org

This study is addressed to three issues critical to the development of programs of educational intervention for elementary school children. First, we are concerned with the development of self-attitudes in children over time. In this study we use a cross-sectional approach to uncover age-related differences looking specifically at the development of self-attitudes, the degree to which the developing attitudes were positive, and the presence of race/ethnic indicators as a part of this developing self-identity. Second, we examined the relation of self-attitude development to the development of prejudice. Here the indicator of prejudice was the child’s assignment of stereotypes to self and other racial/ethnic children. Third, we examined the relationship between self and other attitudes. Although the association between these two dimensions of self has been long-established (see Ehrlich, 1973), their association developmentally has not been well-explored. Finally, we look to see if there are discrete stages in the developmental process. That is, are there age-related changes in self and other attitudes that would point to periods where an educational intervention would be more effective. Nesdale (2004), summarizing the current research, says:

    There is currently little agreement on issues such as the age at which prejudice emerges in children, whether or not there are age-related phases or stages through which prejudice develops, what the psychological processes or mechanisms are which govern the acquisition of ethnic prejudice, and what impact is exerted by children’s emerging linguistic and cognitive abilities upon their acquisition and retention of ethnic prejudice. (p. 220)

 

The Study Sample

The Institute was granted access to The Enrichment Centers, Inc., a chain of private summer school programs (June-August, 2003) in the greater Washington, D.C. metropolitan area. Parental consent was given for 120 students to participate in a 20-minute interview. The children ranged in age from five to 16, were from middle class backgrounds, almost equally divided by gender, and ethnically quite diverse. Forty-four percent were White; 17%, Black; 26% Asian-Pacific; with the others including Middle-Easterners, Hispanic, and multiracial children. This diversity, which was touted in the school’s advertising, may have resulted in attracting parents and their children who were more open to an ethnically diverse school than might have been found by a strict random sampling technique.

The interviewers were more or less equally diverse, and two interviewers were assigned to each child. The primary interviewer sat very close to the child, maintaining eye contact and conducting the interview as a game. The second interviewer served as recorder, sitting slightly behind the child and out of easy visibility. The interviews were conducted in a private space, usually an empty classroom.

Indicators of Self Attitudes

Three components of self identity were operationalized in this project. First, we indexed the development level of the child’s attitude. Operationally, “development” refers to the number of discrete references to self the young person could present. The children were asked: “I have a really interesting question to ask you. OK? I would like you to give me five answers to the question ‘Who am I?’ Start each answer with ‘I am____’. So if I were doing it, I would start by saying, ‘I am [insert name]’. OK, now you go” If the child provided five descriptors, the interviewer asked: “That was great! Can you give me three more ‘I ams’ ?”

The second component of self we examined was that of favorable self-attitudes, what some have referred to as self-esteem. Operationally, we simply counted the number of positive self descriptors the child used in answering “Who am I?” Examples of positive self descriptors are-------Male 10, good at helping my friends; F9, good to my dogs; F7, a nice person; F15, hard-working; F11, happy; F7, good with technology; M8, very interesting;

For the younger children, who worked hard to reply, only 1 percent of their self-representations could be coded as positive. There appeared to be a linear age progression in self evaluative statements with 40 percent of those 11 years and older citing positive self evaluations.

The third component of self was racial/ethnic identity. Did the child mention anything that could be construed as signifying a conception of self as a member of a racial/ethnic category in response to the “who am I” question?

Racial/ethnic identity was also indexed by two independent questions. We followed the “who am I” question with a nondirective prompt: “If you were talking on the telephone to someone who had never seen you, how would you describe yourself to them?” The responses were coded by the presence or absence of any indicator of a race/ethnic identity. Finally, we asked very directively, “What do you call yourself? For example, some people call themselves African Americans or Japanese, or Vietnamese or Latino or Jewish or Korean or Iranian–what do you call yourself?”

Who am I?

Development. Would there be a discernible developmental difference in the expression of the children’s identity by age, gender, or race?

Approximately 14 percent of the children were not able to answer the question. The difficulty was greatest for the younger children (5-6 years). The age differences were substantial (chi square = 26.4, 3 df, p<.001). Among those responding, the mean and median response was approximately five self-descriptive terms. There were no gender differences in the number of responses. However, Asian children had significantly fewer responses than did Black or White children (chi square = 7.3, 3 df, p=.06).

Self-esteem. Once again age was the significant correlate. Self-esteem, the number of positive self descriptors, was a direct function of age. Of the 23 five-to-six year olds, only two cited a positive self attribute. By seven and eight, one out of four did so, and by age eleven approximately 40 per cent were describing themselves with favorable descriptors. Again, there were no gender differences. There was a very strong tendency for Asian children, as compared to black and white children, to avoid positive self descriptors. The black-white differences were minimal.

Self-description by telephone. The telephone task confronted the children with an existentially different task. While the former question evoked the child’s self-identity, this question shifted to potential responses made by others. Answering the question, who am I, permits answers which are subjective and non observable. For example, “I am good at math” or “I am a fast runner.” In contrast, telling someone how to recognize you means that you have to focus on external appearances. Many of the children tried to describe what they wearing then, but the interviewers were instructed to discourage descriptions of clothing unless they were unique. The pattern of telephone descriptions was quite similar to the “who am I” responses with regard to age, gender, and race/ethnicity. However, twice as many children in the telephone task cite their race/ethnicity. Among those responding to both questions (n=99), 53 percent do not mention their racial/ethnic background while 11 percent mention it in both questions. Race/ethnicity. Although the adult studies lead to the expectation that race/ethnicity would be a salient component of self-identity, this was not replicated in this sample of children. Eighty-two percent did not identify themselves on the “who am I” by their racial or ethnic background (with ethnicity construed in its most inclusive form). However, when asked specifically what they call themselves, almost all of the Asian and African American children (93% and 96% respectively) provide a ready racial/ethnic response. In contrast, 20% of the White American children were nonplused by the question and had no answer.

One American Indian child provided a temporary setback for the interviewing team when he declared that he called himself a “truck.” After some discussion, with him and his teacher, we were able to discern his logic: he was a Cherokee and that was the brand name of a Jeep truck; hence, he was a truck. Two of the younger children of color could not generalize from the illustrations in our question. One youngster, age five, let us know that she was “Tinkerbell,” while another 5-year-old told us he was an Okemon (a character in a Japanese video game).

Attitudes Toward Others

In adult investigations, the correlation between attitudes toward others and self-attitudes has been explored extensively. The research has generally found that persons who like themselves, like others. Persons who dislike themselves dislike others. What of children who are in the process of developing their own self-identity? How are attitudes to others expressed, and how is this expression manifest toward the race /ethnicity of others? To explore this, we employed a set of photographs mounted on a 9x12 cut of poster board. There were six boards, three featuring only boys and three for girls. The photographs, mainly full head shots, were sorted by age. Each chart featured a small age range and the children were shown only that chart displaying kids in their own age grouping. Each board contained seven to 11 pictures with typically seven ethnically distinguishable young people.

The interviewer introduced the task: “Now we are going to play a little game. I will read you a word and you point to the person who that word fits the best. There’s no right or wrong answer, of course. I’ll pick a word and you pick a face.” The words given were: hard-working, loud, honest, smelly, strange, athletic, and lazy. (We had also included “patriotic” but dropped it as it proved conceptually too difficult for many, particularly the younger children.) The children were also asked to select the “smartest” and the “dumbest,” and whom they would like as a “best friend.” We had two initial concerns. Would the children differentially select the photographs, that is, would they display a rudimentary assignment of stereotypes? Would their choices reflect an ingroup/outgroup bias?

To look for stereotype assignment, we tested the hypothesis that the selections significantly departed from a random distribution. To do so, we compared the childrens’ choices of African American, Asian American, and European American photos. The findings are displayed in Table 1.

                    Table 1.  Stereotype Assignments

Trait	χ²	p-value	Trend
Strange	14.3	<.01	Asians most frequently         selected
Lazy	9.3	<.05	Asians least frequently         selected
Loud	5.9	<.20	European Americans most often selected
Smelly	1.0	ns	No trend
Honest	5.9	<.20	African Americans most frequently selected
Athletic	22.5	<.001	Asians least frequently selected
Hardworking	8.3	<.05	European Americans least frequently selected
Dumbest	19.0	<.001	European Americans most frequently selected
Smartest	23.0	<.001	African Americans least frequently selected

Although there is some intrinsic interest in the content of this table to the student of stereotyping, the content is of secondary concern to us. The importance of the table is the frequency with which given traits were consensually assigned to specific groups. This is after all a sample of children who typically did not use their own race/ethnic identity in either self-identification or in self-description. However, they agreed beyond a chance level in eight of the nine instances in the assignment of these stereotypes.

To gain another perspective on the assignment of stereotypes, we examined the targets of the stereotyping. Did the children distinguish the stereotypes and targets by assigning more of the positive characteristics to their own group and more of the negative characteristics to other groups? Accordingly, we separated positive and negative stereotypes, sorting them by whether they were assigned to the kid’s ingroup or to an outgroup. Presumably, if the assignment were random there should be little difference in the depictions of outgroup and the membership group of the child. These data are presented in Table 2. Its central finding is clear. Regardless of the race/ethnic background of the children, the assignment of negative stereotypes to the outgroups exceeds their positive self attributions. Again we observe that while race/ethnicity is not yet well-developed in the children’s self identity, they have already “learned” to value their own groups while strongly disvaluing others.

As an aside, we again observe that there are clear group differences in the assignment of these stereotypes with the Black children in this sample engaged in less stereotyping of self and others and the White children most stereotyping.

We next examined each age grouping by stereotype assignments. Using a coefficient of variation (Coulter, 1989), we calculated how many children of a given age would need to be redistributed in order for the stereotype assignments to be equally distributed among them. Each age grouping (5-6, 7-8, 9-10, 11 up) had nine V scores, one for each stereotype. We then calculated the average V score by summing across the stereotypes within the age category since there was no basis for expecting an age-grading with regard to specific stereotypes. Logically and theoretically, however, we expected that as children got older, their perspectives on group stereotypes would come to be more homogeneous. This would be in keeping with our observation of age differences in the development of self-attitudes and self-esteem.

Contrary to expectations, the differences across the first three age categories were virtually zero, averaging two percent. In contrast, 16 percent of the children 11 years old and older would have to be “moved” for their responses to manifest a random distribution. While the age differences with regard to self and race/ethnic identity were smaller than expected, they were nevertheless apparent. In contrast, the assignment of stereotypes to others is virtually random among the younger children. By age 11, the number of stereotypes assigned to others increases and the variability of their assignment decreases.

Racial/Ethnic Identity

In this analysis we examine the self-identity and its relation to an index of racial/ethnic identity. The primary question directing us here is whether positive self-attitudes are related to a strong racial/ethnic identity and, conversely, negative self attitudes to a weak racial/ethnic identity. To begin, we constructed an index of race/ethnic identity based on the childrens’ responses to five questions.

Did the child cite race/ethnicity in response to “Who am I?

Did the child cite race/ethnicity in response to describing him/herself  over the telephone?

Did s/he identify father or mother by their identity?

Had the child been subject to teasing because of his/her racial/ethnic identity?

Had the child witnessed others being teased because of their racial/ethnic identity?

We scored the responses “0" if the response was absent and “2" if the response was present. In a few instances where coders were uncertain we accorded the response a “1". The range of scores was zero to 10. The index combines elements of both self and social identity along with the experience of racial/ethnic harassment. There were no children who received a score of 10, while 20% received a score of zero. The median value fell between scores of four and six.

With regard to the index scores, there were no differences by gender and age. As Table 3 shows, White children had substantially lower scores than Blacks and Asians both of whom displayed similar scores. The results by age indicate that the youngest children did have the lowest scores, but there was no consistent, statistically significant pattern with increasing age. The age bump for the nine-and-ten-year-olds, however, is not entirely unexpected and will be addressed in the next section.

Finally, we ran a correlation between the number of positive self-attitudes expressed in answer to “who am I” and the index scores. Our hypothesis, of course, was that the more favorable the self-attitudes, the stronger the child’s racial/ethnic identification. With a correlation of .85, (p<.04), the hypothesis was strongly confirmed.

What Have We Learned?

1. There were manifest differences in the expression of self-and social-identities by children at different age groups and by race/ethnicity. This is consistent with past research.

2. Looking at positive self-attitudes, we can see a progression in their expression by age. If there was any surprise to this observation, it was that 60 percent of the children 11 and older had still not developed positive self-descriptors —at least, well-enough to express one in eight possible answers to “who am I.”

3. In their expression of social identity (the telephone question), the children expressed greater awareness. Twice as many mentioned their race/ethnicity in answering this question than in answering the previous question. It appears that while many children had not yet incorporated their race/ethnicity in their self-identity, they recognized it as part of their social identity. They appeared to understand, in a preliminary fashion, that there was a difference in how others looked at them.

4. Prompting the children with racial/ethnic labels (recall that they were presented with seven different labels), yields quite different results. Almost all of the Asian (93%) and African American (96%) children had no difficulty in categorizing themselves. White children, in contrast, were somewhat less likely to invoke a racial/ethnic category. Only 80% did so. Again, we are led to the conclusion that identifying oneself in racial/ethnic terms lags behind the understanding of these concepts as applied to one’s self identity.

5. Self and social identity do seem to develop in a patterned manner. As the number of positive self-attitudes increases, so the strength of a racial/ethnic identity increases.

6. The increased awareness of racial/ethnic identity among those 9-10 years-old and its reversion by age 11 (Table 3) has its parallel in past observations of the peaking of prejudice at ages five to seven and a subsequent drop-off. Researchers have suggested that this may be a consequence of changes in the child’s presentation of self accompanying their developing awareness of social identities. (See Rutland, 2004.)

7. When presented with the task of assigning stereotypes to a diverse set of photographs, the children demonstrated a substantial consensus. Behaviorally, many giggled through the task and a few expressed reluctance. Nevertheless, their answers displayed clearly that they distinguished ingroups and outgroups. Further, as Table 2 indicated the children were far more prolific in their assignment of negative stereotypes to outgroups.

The Policy Options

This research was guided by two questions. Do self-attitudes and prejudice develop along the same trajectory? The answer is affirmative, and to the degree that they do, its implication for the educator is that programs geared to insulating children from prejudice might appropriately deal with self as well as social attitudes.

In the second question, we asked — Is there a chronological or developmental pattern to the child’s acquisition of prejudice? The answer to this would suggest that educational programs could be timed to coincide with the pattern. Our data indicate that grade seven and/or 10-11 years-of-age appears the most conducive for introducing a race/ethnic relations learning module.

Works Cited

Coulter, P.B. (1989) Measuring inequality. Boulder, CO:              Westview Press.
Ehrlich, H.J. (1973) The social psychology of prejudice. NY: Wiley.
Nesdale, D. (2004) Social identity processes and children’s ethnic prejudice. In M. Bennett and F.Sani (Eds.) The development of the social self ( pp. 219-245) NY: Psychology Press.
Rutland, A. The development and self-regulation of intergroup attitudes in children. In M. Bennett and F.Sani (Eds.) The development of the social self ( pp. 247-265) NY: Psychology Press.

* Howard J. Ehrlich is the executive director of The Prejudice Institute, a Baltimore-based think tank dedicated to policy research in all areas of prejudice, discrimination, ethnoviolence and social policy. Melissa Childs, Ari Fogelman, Jolanta Smolen, Laura Turner, and Summer Woo served as interviewers and assisted in the design of the interview schedule. Liz Wiemers assisted with the data analysis.


Table 2. Mean Number of Stereotypes Selected By Direction of Choice and Race/Ethnicity

	Positive Ingroup Choices N             Mean	Negative Outgroup Choices N                      Mean
Asian	38               9.50	70                     17.50
Black	15                3.75	37                      9.25
White	53               13.25	80                      20.0

     Note: Cell entries refer to the number of choices made by all        children in the specified group with the mean referring to the average number of choices per group. For example, the bottom left-hand cell indicates that all White students made 53 positive ingroup choices, an average for White students of 13.25.

Table 3. Mean Scores on the Race/Ethnicity Index
____________________________________________

R/E                                    Mean Score

Asian                                     3.6
Black                                     3.3
White                                     2.3
------------------------------------------------
AGE

5-6 years                                 2.2
7-8                                       2.9
9-10                                      3.7
11 years & older                          2.8

------------------------------------------------
GRADE IN SCHOOL

K                                         1.6
1-2                                       2.6
3                                         3.2
4-5                                       3.8
6                                         3.0
7-8-9                                     2.8

__________________________________________________

Note: The greater the index score, the greater the racial/ethnic identification.

Who We Are . Projects . Newsletter . Publications . Guest Book . Speakers . Internships . Factsheets . Links . Current Issues . HOME