Back to Journal Cover Page

CURRENT RESEARCH IN SOCIAL PSYCHOLOGY

Volume 7, Number 13
Submitted: January 7, 2002
Resubmitted: June 14, 2002
Accepted: June 14, 2002
Publication date: June 21, 2002

EFFECTS OF SELF-ESTEEM, STATUS, AND IDENTIFICATION ON TWO FORMS OF INGROUP BIAS

Christopher L. Aberson
Humboldt State University

Lauren M. Howanski
Franklin and Marshall College

ABSTRACT

A correlational study examined relationships between two forms of ingroup bias, status, ingroup identification, and personal and collective self-esteem subscales. Group identification correlated positively with ingroup bias for all status comparisons. Collective self-esteem (CSE) results suggested that different forms of CSE influence ingroup bias in several manners. Public CSE related positively to direct bias and indirect bias for high status ingroups. Private CSE related to greater direct ingroup bias. Lower membership CSE related to greater direct and indirect ingroup bias. Personal self-esteem did not predict ingroup bias. Results suggest a complex relationship between self-esteem, ingroup bias, identification, and status.

[225]
---------------
[226]

INTRODUCTION

Social identity theory proposes that individuals define themselves in terms of group memberships and use comparison strategies that enhance differences between groups in manners that favor ingroups (Tajfel 1982; Tajfel and Turner 1979). These strategies, termed ingroup bias, refers to favoring one's own group over other groups. The phenomenon is robust and well documented (see Brewer 1979; Mullen, Brown, and Smith 1992 for reviews). Several factors proposed to moderate ingroup bias include self-esteem (e.g., Hogg and Abrams 1990), identification with the ingroup (e.g., Spears, Djoose, and Ellemers 1999), status (e.g., Mullen, Brown, and Smith 1992), and ingroup bias dimension (e.g., Aberson, Healy, and Romero 2000).

The relationship between ingroup bias and self-esteem is an oft-debated topic in social identity research. Hogg and Abrams (1990) highlighted two corollaries regarding this relationship. The first states that intergroup discrimination enhances self-esteem. The second argues that depressed self-esteem promotes ingroup bias. Both corollaries suggest a central role for self-esteem in social identity theory. A qualitative literature review supports the proposition that successful discrimination enhances certain dimensions of self-esteem (Rubin and Hewstone 1998). We focus on self-esteem as a predictor of ingroup bias rather than as a product of ingroup bias (e.g., Hunter, Reid, Stokell, and Platow 2000).

A fundamental issue when conceptualizing self-esteem as a predictor of ingroup bias is whether self-esteem relates positively or negatively to ingroup bias. Researchers commonly ask, "Who shows more bias, low or high self-esteem individuals?" Hogg and Abrams' (1990) argue that depressed self-esteem leads to greater motivation for ingroup bias. Low self-esteem motivates ingroup bias; therefore, people lower self-esteem should be more likely than those with high self-esteem to exhibit ingroup bias. Predictions from this perspective argue for a negative correlation between self-esteem and ingroup bias. Others argue (e.g., Crocker and Luhtanen 1990) that ingroup bias produces positive self-esteem so those who exhibit the most bias should have the highest self-esteem, indicating a positive correlation between self-esteem and ingroup bias.

Brown (1993) contended that all individuals experience a need to self-enhance; however, there also exists a need for self-consistency. As individuals with low self-esteem do not view themselves as superior to others, rating themselves or their ingroups as superior is inconsistent with experiences (Brown, Collins, and Schmidt 1988). Other definitions of ingroup bias, for example, rating similarity to successful ingroups, may be favored by those low in self-esteem as this type of measure does not require ratings of superiority (i.e., is not inconsistent with experience).

[226]
---------------
[227]

A meta-analysis suggested that personal self-esteem and identification with the ingroup correlate positively with ingroup bias but collective self-esteem does not. Individuals with higher personal self-esteem and those who identify more with the ingroup showed greater ingroup favoritism. The meta-analysis also tests a self-consistency hypothesis by distinguishing between direct and indirect bias strategies. The analysis defined direct bias as dimensions of evaluation requiring claims of ingroup superiority such as adjective ratings. Indirect bias strategies included measures of similarity or ratings of groups where the participant acts as an observer rather than a participant. Individuals with high personal but not collective self-esteem preferred direct bias strategies. Both low and high self-esteem individuals used indirect bias strategies (Aberson, Healy, and Romero 2000).

Though the meta-analysis did clarify the role of self-esteem measurement and bias strategy, there is room for other variables in explaining the relationship between ingroup bias and self-esteem. The current study extends this research through examination of status and collective self-esteem subscales as well as inclusion of group identification and personal self-esteem measures.

Status
Members of high status groups exhibit ingroup bias more consistently than do members of low status groups (Mullen, Brown, and Smith 1992). Often group status, defined as the relative standing of the group in relation to other groups, is equated with self-esteem. Several studies attempt to test social identity theory corollaries regarding self-esteem through manipulation of status (e.g., Sachdev and Bourhis 1987). However, others found that individuals who are members of low status or stigmatized groups are no more or less likely to have low self-esteem than members of high status or non-stigmatized groups (Crocker, Luhtanen, Blaine, and Broadnax 1994; Crocker and Major 1989). As such, it may be the case that negative aspects of the social identity (i.e., low status) disassociate from self-esteem (Steele 1997). Given these findings, we view status and self-esteem as independent constructs.

Hypothesis 1. More ingroup bias exists in evaluations involving high status ingroups compared to those involving low status ingroups.

Group Identification
Spears, Doosje, and Ellemers (1999) provide an extensive review of group identification research and conclude that identification is an important factor in determining commitment to group membership. Identification with the ingroup should correspond to a readiness to categorize one's self as a group member. Meta-analysis results indicate that group identification correlated positively with direct ingroup bias (Aberson, et. al. 2000). Though not directly supported by meta analysis results due to a small pool of studies, group identification is likely also associated with indirect ingroup bias, defined here as the tendency to enhance association with the ingroup.

[227]
---------------
[228]

Hypothesis 2: Group identification correlates positively with direct and indirect ingroup bias. Identification with the ingroup correlates positively with greater amounts of ingroup bias, regardless of group status.

Self-Esteem
Whereas the relationships between status and ingroup bias and self-esteem and ingroup bias are established, little is known about the relationship of status and self-esteem. Specifically, we are interested in how self-esteem influences evaluations of high and low status groups.

Personal Self-Esteem
Luhtanen and Crocker (1992) argued that the self-esteem referred to by Social Identity Theory reflects esteem gained from social group memberships. As such, measures of personal esteem may not be predictive of collective enhancement, as these measures do not refer to social group membership. Despite these claims, Aberson et. al. (2000) found that high personal self-esteem related positively to greater use of direct ingroup bias strategies.

Hypothesis 3: Personal self-esteem correlates positively with direct ingroup bias.

Collective Self-Esteem
The collective self-esteem scale consists of four subscales. These scales measure personal evaluations of the ingroup (private CSE), perceptions of other’s views of the ingroup (public CSE), perception of one’s contribution to the group (membership CSE), and importance of the group in one’s identity (identity CSE). The use of these scales is inconsistent. Aberson (1999) used the membership subscale, DeCremer (2001) used the public scale, and Crocker and Luhtanen (1993) used the private scale. We provide hypotheses for the impact of three of these subscales on ingroup bias.

Public Collective Self-Esteem
High group status reflects a general agreement that one group is better than another on some important dimension. Public collective self-esteem measures individual perceptions of how others view the group. Those individuals who see the group as positively viewed by others should show greater ingroup bias. However, this bias may only appear in situations wherein status considerations support these perceptions. That is, if the individual views the group as positively viewed by others, but the group is of low status, they will not show bias.

[228]
---------------
[229]

DeCremer (2001) found public collective self-esteem correlated positively to ingroup bias. We suggest that this result only applies to high status groups. Though not explicitly mentioned in the DeCremer paper, results indicate that the ingroup received higher ratings than the outgroup. Brewer (1979) suggests that differentiation of this type occurs when the ingroup is of high self-esteem. When the ingroup is of low status, ingroup bias may appear as a reduction in the differences between high status outgroups and low status ingroups.

In evaluating high status groups, public self-esteem may be of particular relevance. Public CSE refers to the individual’s perception of how others view the group. Those who are low in public self-esteem experience an inconsistency between group status and their perception of group status. That is, low public CSE may involve denial of the high status of the group. As such, public CSE should correlate positively with ingroup bias for high status groups, resulting in both more favorable evaluations (direct bias) and greater association with the group (indirect bias). We believe that the effects of high status and high public CSE are additive. The presence of both increases bias. Low status groups produce less ingroup bias, thus we predict no differences based on public CSE.

Hypothesis 4: Public collective self-esteem relates positively to direct and indirect ingroup bias exhibited towards high status groups.

Private Collective Self-Esteem
Private collective self-esteem reflects personal evaluations of the ingroup. Crocker and Luhtanen (1990) found participants higher in private CSE exhibited greater ingroup bias. They explain this result as indication that private CSE directly relates to social identity (e.g. Tajfel and Turner 1979) and thus should be associated with ingroup bias. We suggest that this form of self-esteem should be most relevant when group status is low, as no external reinforcement of group value exists. When group status is high, private CSE is less relevant as the group is already highly valued. As such, we believe that individuals with high private CSE are more likely to enhance the value of low status groups, as they are convinced of the group’s value. However, there is no reason to believe that this will lead to increased association with the low status group, as association is most appropriate when status is high.

Hypothesis 5: Private collective self-esteem relates positively to direct bias when status is low.

[229]
---------------
[230]

Membership Collective Self-Esteem
Membership collective self-esteem refers to the individual’s contribution to the group. Those individuals with low membership esteem do not feel that they contribute much to their groups. High membership self-esteem individuals are confident that they contributed to the group.

Aberson (1999) found that individuals lower in membership self-esteem tended to enhance their association with positively valued groups as a means of basking in the reflected glory of the ingroup. We predict this result for high status groups, as enhancing association with low status groups is likely an ineffective ingroup bias strategy.

Hypothesis 6: Membership collective self-esteem correlates negatively with indirect ingroup bias for high status groups.

METHOD

Participants
One hundred seventy-five undergraduate students at a private liberal arts college (n = 84) and a state university (n = 91) participated in a study of group memberships for extra credit, to satisfy research participation requirements, or for monetary compensation. Most participants were of traditional college age (89.1% 18 to 22 years old), white (83.25%), and women (74.3%).

Measures
Participants rated four target groups, college students, either people their age who are not college students or college faculty, students at their college, and students at another college.

Group Identification and Self-Esteem
Participants completed the Inclusion of the Ingroup in the Self measure of group identification (IIS; Tropp and Wright 1999). The Rosenberg's Self-Esteem Scale (RSE; Rosenberg 1965) assessed personal self-esteem. The Collective Self-Esteem Scale (CSES: Luhtanen and Crocker, 1992) measured group-level self-esteem. Appendix A presents the IIS. The CSES is accessible from http://rcgd.isr.umich.edu/crockerlab/cse.htm.

[230]
---------------
[231]

Direct Bias
A 16-item adjective rating scale including eight positive items and eight negative items, measured of direct ingroup bias. Participants rated how descriptive each item was for each target group. Responses range from (1) does not describe to (7) describes completely. Sample items included "friendly," "sincere," "trustworthy," "stupid," and "boring." Differences between ingroup and outgroup ratings comprise the direct ingroup bias measure. Appendix A presents this measure.

Indirect Bias
Indirect bias measures focused on perceived similarity between the participant and members of the ingroup and outgroup (cf. Aberson 1999; Aberson, Healy, and Romero 2000). Participants indicated the extent that they shared common interests, common experiences, and their overall similarity with the group. For questions about common interests and experiences, responses ranged from (1) share none to (7) share many. For the overall similarity question, responses ranged from (1) not at all to (7) very. The differences between ingroup and outgroup ratings of similarity comprise the indirect ingroup bias measure. Appendix A presents this measure.

Procedure
Instructions informed participants that they were taking part in a study of attitudes towards themselves and members of specific social groups. Participants first completed the Rosenberg and the Collective Self-Esteem then rated two sets of groups, comprising the low and high status comparisons. For half of the participants the low status ingroup vs. high status outgroup comparisons was students at their college (IG) vs. students at a more prestigious rival college in the same region (OG) and the high status ingroup vs. low status outgroup comparison was college students (IG) vs. people your age who are not college students (OG). The other group rated college students (IG) vs. college faculty as the low status ingroup-high status outgroup condition and students at their college (IG) vs. students at a local community college (OG) as the high status ingroup vs. low status outgroup comparison. We examined interactions between group composition (i.e., the ingroup-outgroup pairs rated) and found no interactions between the pairs and any of the predictor variables. Given this result we collapse across the two conditions.

For both sets of comparisons, participants first listed three characteristics descriptive of each group. The purpose of this task was to focus attention on group identity prior to rating groups. After listing traits, participants indicated identification with the ingroup on the Inclusion of the Ingroup in the Self Scale. For each group, participants completed the direct and indirect bias measures. Low/high status target presentation and direct/indirect bias measures were counterbalanced, however, ingroup ratings always preceded outgroup ratings within the pair.

[231]
---------------
[232]

RESULTS

Tables 1 and 2 summarizes means, standard deviations, reliability statistics (where appropriate), and correlations for each measure. Table 3 presents a regression analyses predicting ingroup bias from ingroup identification, personal self-esteem, and collective self-esteem. Analyses predict ingroup bias from group identification, personal self-esteem, and the four collective self-esteem subscales. Though specific hypotheses do not exist for all collective self-esteem subscales for each analysis, we include these variables to be better able to draw conclusions as to the effects of each subscale.

As shown in Table 2, correlations between some of the self-esteem measures were larger than ideal for inclusion as predictors in regression analysis, suggesting the presence of multicollinearity. To address these issues, we examined tolerances, condition indices, and variance proportions for each analysis following criteria suggested by Tabachnick and Fidell (2001). None of the four analyses produced low tolerances. Further, no analysis found large fit indices coupled with two or more high variances proportions for any single predictor. These results suggest that multicollinearity is not a serious problem for the analyses that follow.

[232]
---------------
[233]

Table 1. Means, Standard Deviations, and Reliabilities

Note. Inclusion of Ingroup in Self does not include reliability as it is a single item measure. Positive scores on ingroup bias measures indicate greater bias toward ingroup.

[233]
---------------
[234]

Table 2. Correlation Matrix

Note. * p < .05, ** p < .01. n for each pair in parentheses.

[234]
---------------
[235]

Partially supporting the first hypothesis, participants favored the high status ingroup (M = 7.1) over the low status ingroup (M = -3.3), t(158) = 9.44, p < .001, eta-squared = .36. Indirect bias results did not support this hypothesis. Similarity ratings of high status (M = 3.3) and low status (M = 3.1) were comparable, t(158) = 0.43, p = .67, eta-squared = .01. Self-esteem results below qualify these findings.

Supporting Hypothesis 2 and shown in Table 2, greater identification with the ingroup related to greater ingroup bias in all status and bias dimension conditions. Contrary to the third hypothesis, personal self-esteem did not correlate with ingroup bias.

Consistent with Hypothesis 4, public collective self-esteem related positively to direct bias and indirect ingroup bias exhibited towards high, but not low, status groups. Supporting Hypothesis 5, private collective self-esteem related positively to direct bias when the ingroup was low status. Membership collective self-esteem results supported Hypotheses 6. Membership CSE correlated negatively with direct ingroup bias and indirect bias for high status groups. Surprisingly, membership CSE also correlated negatively with direct ingroup bias.

[235]
---------------
[236]

Table 3. Regression Analysis Predicting Ingroup Bias from Group Identification, Personal Self-Esteem, and Collective Self-Esteem Subscales

Note. * p < .05, ** p < .01, *** p < .001.

[236]
---------------
[237]

DISCUSSION

The results of this correlational study partially supported our hypotheses and suggest a complex relationship between ingroup bias dimension, status, self-esteem, and group identification. Our results suggest that different types of self-esteem operate differently depending on status and bias dimension.

Group identification predicted both direct and indirect ingroup bias. The effects of group identification appear to be independent of status. Regardless of the status, those who identified more showed greater ingroup bias on both dimensions. This result suggests a central role for group identification in predicting ingroup bias.

Whereas previous research finds an overall effect for personal self-esteem wherein individuals high in personal self-esteem exhibited more ingroup bias (e.g., Aberson, Healy, and Romero 2000), the correlational study failed to replicate this result. We are reluctant to discard personal self-esteem as a useful predictor of ingroup bias based on this result. One explanation for the lack of a personal self-esteem effect is range restriction on Rosenberg Self-Esteem Scale scores. For our sample, the average item score was 3.29 (out of a possible 4.0). Further, only 11 participants (6.3%) produced scores that would fall in the lower half of the scale. As such, we failed to obtain a large sample of individuals who were truly low in personal self-esteem. To resolve this problem, future research should include populations of individuals with markedly depressed self-esteem.

Collective self-esteem results supported most hypotheses and suggest that different forms of collective self-esteem influence ingroup bias in several manners. Public collective self-esteem related positively to direct bias and indirect bias for high but not low status ingroup comparisons. This supports our proposition that public self-esteem influences depend on group status. Public CSE reflects the individual’s perceptions of how others view the ingroup. If the individual believes that others view the ingroup positively, then public CSE is high. Our findings suggest that public CSE relates to ingroup bias only when group status confirms the legitimacy of high public CSE. When group status is low, it may indicate that high public CSE is not justifiable and reduce ingroup bias. Additionally, we found that high public CSE participants enhanced their association with the high status ingroup.

Private CSE related positively to favoritism for the low status ingroup. This suggests that individuals with high private collective self-esteem believe in the value of the ingroup regardless of status. This is, however, limited to direct bias. As others negatively value the ingroup, enhancing association with a low status ingroup may not constitute an effective ingroup bias strategy.

[237]
---------------
[238]

Membership self-esteem results suggest that individuals who do not feel they are contributors to the ingroup show greater ingroup bias in both direct and indirect dimensions. We predicted that those who felt their contribution to the ingroup was minimal would enhance their association with the ingroup. However, the direct bias result was not expected. These findings suggest that those who do not view their contributions to the group as important will exhibit more ingroup bias when members of high but not low status groups. This result makes the most sense for indirect bias measures as enhancing association with a positively valued ingroup should result in enhancement though increased association.

Taken as a whole, collective self-esteem results provide some clarification as to when each form of self-esteem affects ingroup bias. Further, the fact that private and public CSE may related positively to ingroup bias, whereas membership CSE relates negatively to ingroup bias suggests that research collapsing across these subscales (i.e., using a total scale score; e.g., Ruttenberg, Zea, and Sigelman 1996) may ignore important relationships and possibly wash out effects.

Limitations
As mentioned earlier, the lack of variability of scores on the Rosenberg self-esteem scale severely limits conclusions about personal self-esteem. Measurement of collective self-esteem is also problematic. The version of the collective self-esteem scale used in this study does not specify group membership, referring only to "my social groups." As such, it is unclear as to which groups the individual is responding. Future research may benefit from use of versions of this scale focused on reactions to specific groups.

APPENDIX A

All examples below use college students as ingroup.

Ingroup in the Self measure of group identification (Tropp and Wright 1999)
Using the pictures below, indicate the pair of circles that best represents your level of identification with College Students. Note: "Self refers to you and "Ingroup" refers to your group (college students).

[238]
---------------
[239]

Direct Bias Measures

Please rate College students on the following items using the scale below:

[239]
---------------
[240]

Indirect Bias Measures
How much do you share common interests with college students?

How much do you share common experiences with college students?

Overall, how similar would you rate yourself to college students?

REFERENCES

Aberson, C. L. (1999). "Low self-esteem and ingroup bias." Social Behavior and Personality, 27: 17-28.

Aberson, C. L., M. R. Healy and V. L. Romero (2000). "Ingroup bias and self-esteem: A meta-analysis." Personality and Social Psychology Review, 4: 157-173.

Brewer, M. B. (1979). "Ingroup bias in the minimal intergroup situation: A cognitive-motivational analysis." Psychological Bulletin, 86: 307-324.

Brown, J. D. (1993). "Motivational conflict and the self: The double-bind on low self-esteem." Pp. 117-130 in Self-Esteem: The puzzle of low self-regard, edited by R. F. Baumeister. New York: Plenum Press.

Brown, J. D., R. L. Collins and G. W. Schmidt (1988). "Self-Esteem and direct versus indirect forms of self-enhancement." Journal of Personality and Social Psychology, 55: 445-453.

Crocker, J. and R. Luhtanen (1990). "Collective self-esteem and ingroup bias." Journal of Personality and Social Psychology, 58: 60-67.

[240]
---------------
[241]

Crocker, J. and B. Major (1989). "Social stigma and self-esteem: The self-protective properties of stigma." Psychological Review, 96, 608-630.

Crocker, J., R. Luhtanen, B. Blaine and S. Broadnax (1994). "Collective self-esteem and psychological well-being among White, Black, and Asian college students." Personality and Social Psychology Bulletin, 20: 503-513.

DeCremer, D. (2001). "Relations of self-esteem concerns, group identification, and self-stereotyping to in-group favoritism." The Journal of Social Psychology, 141: 389-400.

Hogg, M. A. and D. Abrams (1990). "Social motivation, self-esteem, and social identity." Pp. 28-47 in Social identity theory: Constructive and critical advances, edited by D. Abrams and M. A. Hogg. New York: Springer-Verlag.

Hunter, J. A., J. M. Reid, N. M. Stokell and M. J. Platow (2000). "Social attribution, self-esteem, and social identity." Current Research in Psychology, 5: 97-125. http://www.uiowa.edu/~grpproc.

Luhtanen, R. and J. Crocker (1992). "A collective self-esteem scale: Self-evaluation of one’s social identity." Personality and Social Psychology Bulletin, 18: 302-318.

Mullen, B., R. Brown and C. Smith (1992). "Ingroup bias as a function of salience, relevance, and status: An integration." European Journal of Social Psychology, 22: 103-122.

Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

Rubin, M and M. Hewstone (1998). "Social identity theory's self-esteem hypothesis: A review and some suggestions for clarification." Personality and Social Psychology Review, 2: 40-62.

Ruttenberg, J., M. C. Zea and C. K. Sigelman (1996). "Collective identity and intergroup prejudice among Jewish and Arab students in the United States." The Journal of Social Psychology, 136: 209-220.

Sachdev, I. and R. Y. Bourhis (1987). "Status differentials and intergroup behavior." European Journal of Social Psychology, 17: 277-293.

Spears, R., B. Doosje and N. Ellemers (1999). "Commitment and the context of social perceptions." Pp. 59-83 in Social Identity, edited by N. Ellemers, R. Spears, B. Doosje. Malden, MA: Blackwell Publishers.

[241]
---------------
[242]

Steele, C. M. (1997). "A threat in the air: How stereotypes shape intellectual identity and performance." American Psychologist, 52: 613-629.

Tabachnick, B. G. and L .S. Fidell (2001). Using multivariate statistics (4^th ed.). Boston: Allyn and Bacon.

Tajfel, H. (1982). "Social psychology of intergroup relations." Annual Review of Psychology, 33: 1-30.

Tajfel, H. and J. C. Turner (1979). "An integrative theory of intergroup conflict." Pp. 33-47 in The social psychology of intergroup relations, edited by W. G. Austin and S. Worchel. Monterey, CA: Brooks/Cole.

Tropp, L. R. and S. C. Wright, S. C. (2001). "Assessing the inclusion of the ingroup in the self." Personality and Social Psychology Bulletin, 27: 585-600.

AUTHORS’ NOTE

Funds from the Franklin and Marshall College committee grants supported this project. Portions of this research were presented at the 2000 American Psychological Society conference, Miami, FL.

AUTHORS’ BIOGRAPHIES

Chris Aberson is an Assistant Professor of Psychology at Humboldt State University, Arcata, CA 95521, USA. Previously, Dr. Aberson worked as a Visiting Assistant Professor at Franklin and Marshall College. He received his Ph.D. in psychology, social and quantitative focus, from Claremont Graduate University in 1999. E-mail: cla18@humboldt.edu.

Lauren Howanski is an undergraduate at Franklin and Marshall College.

[242]
---------------
[243]