THE MEANING AND CONSEQUENCES OF PERCEIVED DISCRIMINATION IN DISADVANTAGED AND PRIVILEGED SOCIAL GROUPS

We thank Monica Biernat, Jack Brehm, Chris Burris, Jean-Claude Croizet, Jake Harwood, Alex Haslam, Miles Hewstone, Jolanda Jetten, Nicolas N’gbala, Tom Postmes, Paul Silvia, Heather Smith, Joey Sprague and Wolfgang Stroebe for helpful comments on earlier versions of this manuscript.

Address correspondence to Nyla R. Branscombe, Department of Psychology, University of Kansas , Lawrence , KS 66045 . E-mail: nyla@ku.edu

Abstract

The subjective meaning and consequences of perceived discrimination depends on the position of one's group in the social structure. For members of disadvantaged groups, attributions to prejudice are likely to be internal, stable, uncontrollable, and convey widespread exclusion and devaluation of one’s group. For members of privileged groups, the meaning of attributions to prejudice is more localized. Because of such meaning differences, attributions to prejudice are considerably more harmful for the psychological well-being of members of disadvantaged groups than they are for members of privileged groups. According to the Rejection-Identification Model (Branscombe, Schmitt, & Harvey, 1999), members of disadvantaged groups cope with the pain of attributions to prejudice by increasing identification with their disadvantaged group. We conclude with an exploration of the social contextual factors that can affect how the disadvantaged cope, and a discussion of the challenges facing future research on attributions to prejudice.

THE MEANING AND CONSEQUENCES OF PERCEIVED DISCRIMINATION

IN DISADVANTAGED AND PRIVILEGED SOCIAL GROUPS

One of the most enduring ideas in social psychology is that the subjective experience of social life is moderated by attributions (Weiner, 1985). We consider the psychological consequences of attributing one’s negative outcomes to prejudice against one’s group membership. We review and integrate previous theoretical and empirical work on such attributions to prejudice, and consider how attributing negative events to prejudice influences psychological well-being. We first outline our theoretical reasoning for how and why attributions to prejudice have very different meanings for disadvantaged and privileged groups. We conclude that these attributions are more psychologically harmful for members of disadvantaged groups than for members of privileged groups and review evidence supporting this conclusion. We then discuss how the psychological benefits of identification with one’s disadvantaged group can counter some of the painful consequences of recognizing prejudice. Finally, we consider factors that limit disadvantaged group members’ ability to cope via group identification, and make suggestions for future research on attributions to prejudice.

The total relevance of attributions to prejudice

What an attribution to prejudice means to the person making it is central to understanding its emotional consequences. In its most basic sense, an attribution to prejudice explains a negative event as the result of someone’s bias against one’s category membership. While this description of an attribution to prejudice describes what the attribution implies for the immediate situation–its “local relevance,” it does not necessarily capture all of its larger potential meanings and implications–its “total relevance.” Heider (1958, p. 253) suggested that it is the total relevance of an attribution, not its local relevance, that guides our interpretation of and reaction to events. Thus, understanding the meaning and consequences of an attribution to prejudice requires a consideration of its total relevance for the person making it.

Because prejudice and discrimination are intergroup phenomena, an analysis of attributions to prejudice calls for an examination of how the state of intergroup relations determines the meaning of those attributions. Our central argument is that the total relevance of an attribution to prejudice is a function of the ingroup’s and the outgroup’s relative positions of power within the social structure. The meaning and consequences of an attribution to prejudice will depend on the nature of the existing intergroup relations at a particular historical juncture, and more importantly, the position of one’s own group within the social structure. The psychological experience of perceiving prejudice will be very different depending on whether one belongs to a group that is disadvantaged relative to other groups, or a group that is relatively privileged. For the reasons we outline below, we argue that because of their differing social realities, members of disadvantaged groups experience more harmful psychological consequences when making attributions to prejudice than do members of privileged groups.

Severity of discrimination experiences

Clearly, disadvantaged groups suffer because the outcomes they receive are poorer relative to higher status groups. We consider the psychological consequences of how those negative events are interpreted, rather than the consequences of the events themselves. However, the differential severity of events potentially labeled as discrimination is also likely to moderate the subjective experience of interpreting those events as due to prejudice. Because the magnitude of an event is likely to influence the magnitude of its effects, minor instances of discrimination will have minimal implications for well-being, while more severe discriminatory experiences will have a greater impact. Thus, one of the most important ways in which attributions to prejudice differ for disadvantaged and privileged groups is in the severity of the type of the events typically attributed to prejudice. For instance, Branscombe (1998) found that when men and women were asked to describe the disadvantages that they had experienced based upon their gender group membership, women described relatively severe events such as a lack of freedom, fear of sexual assault, and job discrimination. In contrast, men described less severe and more circumscribed events such as having to pay for dates, or being more likely to get a speeding ticket. In a related study (Kappen, Branscombe, Kobrynowicz, & Schmitt, 2000), women and men listed discriminatory events they had personally experienced, and coders who were blind to the gender of the participants rated the discriminatory events for their potential life consequences. The events generated by women were coded as having a substantially greater impact on the recipient’s life, compared to the events generated by men. Likewise, when White Americans were asked about their experiences with racial discrimination, they reported relatively minor types of discrimination such as not having exclusively White organizations and being seen as racially biased (Branscombe, Schiffhauer, Schmitt, & Valencia, 2000), while Black Americans report discrimination in a wide range of important life domains including housing, education, and employment (Sigelman & Welch, 1991). Thus, because of real differences in the severity of the discriminatory events experienced, attributions to prejudice are likely to have greater emotional consequences for disadvantaged than for privileged groups.

Attributional dimensions

An analysis of attributions to prejudice requires consideration of where these attributions fall on the three main dimensions of attribution: locus of the cause (Heider, 1958), stability (Weiner et al., 1971), and controllability (Weiner, 1979). Indeed, “knowing the dimensional locations of a given attribution allows accurate prediction of the consequences of that attribution” (Anderson, Krull, & Weiner, 1996, p. 280). In this section, we explore the total relevance of an attribution to prejudice by considering its locus (the extent to which it reflects something about the self), stability (the implications the attribution has for future situations), and controllability (the extent to which the outcome is perceived as controllable). We limit our analysis to these three aspects of attribution because they alone can account for a wide variety of effects, and the evidence for the existence or importance of other attributional dimensions (e.g., intentionality, globality) is inconclusive (Anderson et al., 1996).

Attributions to prejudice as both internal and external. Social psychologists often consider the causal locus of attributions as reflecting a single dimension ranging from internal to external, or even as a categorical distinction where attributions are either internal or external. However, assuming that external and internal attributions are inversely related may be an oversimplification (Kelley, 1983), and numerous empirical studies have found that, in fact, internal and external attributional ratings are often independent (McClure, 1998). Thus, the internal and external dimensions of attribution are best conceptualized and measured separately.

Most of the existing theoretical work on attributions to prejudice has assumed that the locus of such attributions is purely external (e.g., Crocker & Major, 1989; Crocker, Major, & Steele, 1998). This assumption lead to the “discounting hypothesis,” the idea that attributions to prejudice discount the self as a cause of negative events (Crocker & Major, 1989; Crocker et al., 1998; Crocker & Quinn, 1998). Accordingly, it is argued that “because prejudice against one’s group is an external attribution, making this attribution should protect the self-esteem of stigmatized people” (Crocker, Voelkl, Testa, & Major, 1991, p. 219).

Attributions to prejudice certainly do have an external component, with the prejudice of the source being outside of the self. On the other hand, we argue that they also have a substantial internal component–the target’s own group membership. Although attributions to prejudice are often described as attributions to “group membership” (e.g. Crocker & Major, p. 613-614), the fact that group membership is an internal aspect of the self has been largely ignored. Because the prejudice of the source (external) and the target’s group membership (internal) are both necessary but not sufficient causes of prejudicial treatment, neither can be discounted when making an attribution to prejudice (McClure, 1998).

The characterization of attributions to prejudice as partially internal is important because it leads to very different predictions about the consequences of attributions to prejudice than does the widely-held view that attributions to prejudice are external. We argue that because attributions to prejudice do not allow for the discounting of all aspects of the self as causes of the negative outcome, they will not necessarily result in self-protection. Because attributions to prejudice are partially internal, they involve the self and therefore have potential for negative self-evaluation (Stipek, 1983, Weiner, 1985; Weiner, Russell, & Lerman, 1979). Rather than shifting the locus of the cause away from the self, attributions to prejudice, in fact, implicate an aspect of the self–one’s group membership–that is often important (Tajfel & Turner, 1986), enduring, and uncontrollable (e.g., gender, ethnicity).

Recent empirical work supports the idea that attributions to prejudice have a substantial internal component. Schmitt and Branscombe (2000) asked male and female undergraduates to consider a situation in which a professor refused their request to enroll in a class that required the professor’s permission to do so. In one condition, participants learned that the professor had a very disagreeable disposition and refused everyone’s request–providing a highly plausible external explanation for the rejection. In the other condition, participants learned that the professor had often publicly expressed a strong bias against the participant’s gender and only admitted those of the other gender–making an attribution to prejudice highly plausible. A manipulation check confirmed that attributions to gender discrimination (e.g., “The professor’s actions were due to gender discrimination”) were made exclusively in the prejudice condition. In order to test the hypothesis that attributions to prejudice involve both internal and external causes, participants completed separate measures of the internal (e.g., “The professor refused to let me into the class because of something about me”) and external (e.g., “The professor refused to let me into the class because of something about him or her”) aspects of attribution. Participants who were asked about the cause of the rejection in the prejudice condition attributed the rejection to both internal and external causes, although in the condition in which the professor rejected everyone, attributions were predominately external. Furthermore, in the prejudice condition the internality of the attribution was significantly higher compared to the other condition, and above the midpoint of the scale. These results suggest that attributing negative treatment to another’s bigotry is not purely external to the self in the same way that an attribution to another’s personality might be.

In a similar study, Schmitt and Branscombe (2000) again asked undergraduates to consider a situation in which a professor denied their request to be let into a class. However, rather than manipulating information about the professor’s disposition or prejudices, we manipulated information about what factors covaried with the professor’s behavior (Kelley, 1967). In the “No one accepted” condition, the participant learned that the professor did not grant anyone’s request to be let into the class, making an attribution to one’s own qualities implausible. In the “Others accepted” condition, the professor granted the requests of about ten other students, making an attribution to the self more plausible. In the “Other gender accepted” condition, the professor only let in students who were not of the participant’s gender, making an attribution to gender prejudice highly plausible. Indeed, as revealed by a manipulation check, attributions to prejudice were much higher in this condition than in the other two. As predicted, the locus of the attributions made differed by condition. Attributions were significantly more internal in the Other Gender Accepted condition than in either of the other conditions, while externality of attributions did not differ by condition. Taken together, these two studies suggest that like attributions to other aspects of the self, attributions to prejudice against one’s group membership do have a substantial internal component.

Our findings are inconsistent with the fundamental assumption of the discounting hypothesis–that attributions to prejudice are exclusively external. Because these data illustrate that attributions to prejudice are in part internal, they undermine the theoretical basis for predicting that attributions to prejudice are self-protective in disadvantaged groups. By treating attributions to prejudice as purely external, the discounting hypothesis may greatly overestimate their self-protective properties. Consistent with this argument, in the two studies just described women reported less positive affect when rejection was attributable to sexism compared to when the rejection was attributable to an exclusively external cause, the professor’s disposition. This suggests that for members of disadvantaged groups, attributions to prejudice do not offer the level of self-protection provided by purely external attributions.

Stability. Although the internal aspects of attributions to prejudice call into question their potential for protecting the self, it is important to note that not all internal attributions for negative events are equally harmful (Rhodewalt, Morf, Hazlett, & Fairfield, 1991; Snyder & Higgins, 1985). Attributional stability is likely to moderate their effects. Attributions to prejudice are likely to be more stable for members of disadvantaged groups than for members of privileged groups. Disadvantaged groups, by definition, experience discrimination more frequently and across a wider variety of contexts than do privileged groups (Sigelman & Welch, 1991; Swim, Cohen, & Hyers, 1998). Women report experiencing more prejudice against their group and perceive prejudice against them as more pervasive across a wider variety of social contexts than do men (Kobrynowicz & Branscombe, 1997; Schmitt, Branscombe, Kobrynowicz, & Owen, 2000). Indeed, in the U.S. there is considerable consensus among men and women and among different ethnic groups regarding the relative status differences that exist between them (Sidanius & Pratto, 1999; Stewart, Vassar, Sanchez, & David, 2000). Precisely because members of disadvantaged are aware that they may face negative treatment on the basis of their stigma in many situations, attributions for negative events that implicate their group membership are likely to be stable and painful (Branscombe, Schmitt, & Harvey, 1999; Golin, Sweeney, & Shaeffer, 1981; Robins, 1988; Weiner,1985). In contrast, because privileged groups do not see prejudice against them as pervasive, they will see negative treatment based on their group membership as anomalies or localized occurrences. Thus, for privileged groups attributions to prejudice are more likely to be unstable and lack meaning beyond the immediate situation. In summary, the differential pervasiveness of discrimination faced by privileged and disadvantaged groups makes attributions to prejudice much more stable, and as a result, more psychologically costly for the disadvantaged.

Because the level of stability of attributions to prejudice reflects the degree to which prejudice will affect one’s outcomes in the future, the stability dimension is of critical importance for understanding the total relevance of an attribution to prejudice. Because an attribution to prejudice among disadvantaged groups reflects more stable treatment, the total relevance of the attribution will be much greater that its local relevance. However, among privileged groups, for whom an attribution to prejudice is more unstable, the total relevance of an attribution to prejudice is likely to be equivalent to its local relevance.

Controllability. The controllability dimension of attribution refers to the extent to which the cause of an outcome is under the volitional control of the self or others (Anderson et al., 1996). The subjective consequences of the perceived controllability of attributions to prejudice can best be captured by considering who, in the context of prejudicial treatment, has the power to inflict their prejudices on others. Because disadvantaged groups are afforded less power within the social structure than privileged groups, encounters with prejudicial treatment are likely to be seen by disadvantaged groups as relatively uncontrollable. Conversely, the outcomes of privileged group members are unlikely to be seen as generally under the control of members of disadvantaged groups, because their privileged status affords them a position of greater power. Consequently, attributions to prejudice made by privileged groups are less likely to harm feelings of control.

Empirical research supports the idea that attributions to prejudice have differential consequences for perceived control among privileged and disadvantaged groups. Ruggiero and Marx (1999) found that disadvantaged groups (women, African Americans, and members of a low status college) who attributed negative feedback to discrimination suffered a reduction in perceived control (see also Ruggiero & Taylor,1997), but among privileged groups (men, White Americans, and members of a high status college) perceived control was not affected by attributions to prejudice. The differential control experienced by disadvantaged and privileged groups is likely to have important implications for psychological well-being. Efficacy-based approaches to self-esteem (Bandura, 1997) suggest that the perception of control over one’s outcomes is an important component of well-being. According to Weiner (1985), negative, internal, uncontrollable events (as attributions to prejudice among disadvantaged groups are likely to be) result in depression (Brown & Siegel, 1988), and can even undermine cognitive ability (von Hecker & Sedek, 1999). Furthermore, when prejudice is recognized by the disadvantaged, making discrimination claims may be perceived as a futile effort because they could result in even more negative treatment from members of the privileged group (Kaiser & Miller, in press).

While members of disadvantaged groups may engage in a number of strategies to avoid encounters with prejudicial treatment (see Miller & Major, 2000), the use of such strategies may not leave one feeling very empowered. For instance, groups with “invisible” stigmas have the option of attempting to pass as members of the privileged group. Attempts to pass in order to achieve more positive outcomes clearly places the control over one’s life in the hands of the privileged group. For example, members of the U.S. military who are gay, lesbian, or bisexual are required to hide their sexual orientation or face exclusion. In such contexts, passing may give stigmatized group members some control in terms of whether they will avoid negative outcomes, but being forced to hide an important aspect of one’s identity severely limits one’s freedom and is likely to be experienced as a vital lack of control. Similarly, having to prove oneself as worthy of equal treatment is a severe disadvantage in itself, and implies that one has less control over one’s life than members of privileged groups, who are more likely to be assumed to be worthy of positive treatment even before they have the opportunity to demonstrate (or disconfirm) the validity of that assumption.

In summary, attributions to prejudice are not simply external attributions that permit a discounting of the self’s causal role in the production of negative outcomes. For disadvantaged groups, attributions to prejudice will be harmful because they implicate an aspect of the self that can result in pervasive negative treatment over which there may be little control across a wide variety of situations. In privileged groups, prejudicial treatment is attributed to an aspect of the self that only infrequently results in negative outcomes. Because of their instability, attributions to prejudice should be less harmful for privileged groups than for disadvantaged groups, and could even be beneficial if they serve to remind the privileged of the positive treatment that they normally receive. Despite these attributional reasons for expecting attributions to prejudice to have very different consequences for disadvantaged and privileged groups, the total relevance of an attribution to prejudice cannot be fully appreciated without looking beyond locus, stability, and controllability to other factors that are intimately intertwined with the power of one’s group in the social structure. Two other important ways in which attributions to prejudice might differ for disadvantaged and privileged groups are the degree of ambiguity about the actual causes of events, and the implications that these attributions have for acceptance of one’s group in society.

Attributional Ambiguity

Major and Crocker (1993) concluded that members of stigmatized groups “exist in a chronic state of attributional ambiguity with regard to the causes of others’ behavior toward them” (p. 346). Based on the assumption that attributions to prejudice for negative events are self-protective, they hypothesized that such attributional ambiguity could be beneficial because it provides the stigmatized with frequent opportunities for discounting the role of ability or performance as an explanation for negative outcomes. Crocker and Major (1989) argue “people who believe they personally are frequent victims of discrimination should be particularly likely to attribute negative outcomes or performance feedback to prejudice or discrimination against their group and, thus, may have high self-esteem” (p. 621). In other words, attributional ambiguity makes it possible for members of disadvantaged groups to blame their outcomes on prejudice when they encounter a variety of negative outcomes. Indeed, Crocker and Major (1989) write that making attributions to prejudice for negative outcomes “is particularly powerful because it may not only be used in response to negative evaluations or outcomes that do, in fact, stem from prejudice against the stigmatized group, but also in response to negative outcomes that do not stem from prejudice” (p. 612).

A number of empirical investigations, however, “strongly suggest that stigmatized individuals are relatively reluctant to blame their negative outcomes on prejudice or discrimination, even when there is good reason to suspect that the evaluator discriminates” (Crocker et al., 1998, p. 522). Indeed, when the likelihood that negative treatment is due to prejudice is even slightly ambiguous, members of disadvantaged groups prefer to attribute negative feedback to their own behavior (Ruggiero & Taylor, 1995, 1997). In Ruggiero and Taylor ’s studies, members of disadvantaged groups received negative feedback, and were presented with different base-rates (0%, 25%, 50%, 75%, 100%) for the probability that the raters were biased against their group. Only when participants were told that 100% of the raters were biased and discrimination was a virtual certainty, did the disadvantaged attribute their failure to prejudice more than to their own performance. More importantly, in all other conditions, participants attributed their failure to their own performance more than the prejudice of the raters. As Ruggiero and Taylor (1997) concluded, “under conditions of situational ambiguity, minority group members are likely to minimize discrimination in response to negative feedback” (p. 374).

As Crocker et al. (1998) note, these data are inconsistent with the idea that members of disadvantaged groups are motivated to make attributions to prejudice, and suggest that attributional ambiguity and attributions to prejudice may have some costs (p. 522). However, they also suggest that the clear evidence of minimization in Ruggiero’s studies is not generalizeable to other disadvantaged groups (p. 522). To reconcile the evidence of minimization of attributions to prejudice with the attributional ambiguity perspective, Crocker et al. (1998) offer a revised perspective on the potential costs and benefits of attributional ambiguity. As Crocker et al. (1991) found, attributional ambiguity is costly when it leads to the discounting of one’s personal role in bringing about positive outcomes. For negative events, they suggest that the self-protective properties of attributional ambiguity “may be limited to circumstances in which negative outcomes are unrelated to one’s negative social identity” (Crocker et al., 1998, p. 520). In other words, in their revised view, attributional ambiguity is threatening when “negative events that are caused by prejudice cannot be easily attributed to prejudice” (p.521). This statement suggests that while attributional ambiguity for stigmatized group members “affords them a degree of latitude” to make attributions to prejudice (p. 521), such ambiguity is threatening when it interferes with reaping the benefits of making an accurate attribution to prejudice. Furthermore, they suggest that while attributions to prejudice for personal outcomes might be minimized because of their costs, attributions to prejudice for group outcomes are not as likely to be minimized. However, the presumed difference between personal and group discrimination perceptions has been shown to be due to the two ratings involving different comparisons and does not reflect distancing one’s own experience from that of the ingroup (see Kessler, Mummendey, & Leisse, 2000; Postmes, Branscombe, Spears, & Young, 1999). Indeed, as Crocker et al. (1998) note, its not clear theoretically why the potential costs and benefits of making attributions to prejudice would differ for judgments made at the personal and group levels (p. 524).

Our perspective on attributional ambiguity is much simpler. We argue that chronic attributional ambiguity—having to regularly consider prejudice as an explanation for one’s negative outcomes—will be detrimental to the well‑being of disadvantaged groups. Such ambiguity about the role of prejudice in one’s outcomes is stressful because it is an ongoing reminder of the rejection and devaluation in the broader culture. Indeed, the fear of confirming other people’s negative views of the ingroup undermines the performance of Black Americans (Steele & Aronson, 1995), women (Spencer, Steele, & Quinn, 1999), and members of the working class (Croizet & Claire, 1998). The mental work required to navigate through social contexts in which one may be a target of prejudice is likely to be extremely taxing (Frable, Blackstone, & Scherbaum, 1990; Goffman, 1963; Kramer & Wei, 1999). Furthermore, not having to consider the role of group membership as a cause of one’s outcomes is one of the major advantages of belonging to a privileged social group (McIntosh, 1998). From our perspective, attributional ambiguity is generally costly to the psychological well-being of disadvantaged groups, and in contrast to the original attributional ambiguity perspective, it is only self-protective in that it provides opportunities to deny that prejudice might actually be the true cause of negative treatment.

Implications for acceptance into society

Among disadvantaged groups, a pattern of stable attributions to prejudice reflects perceived systematic exclusion by the privileged group. Because privileged groups have the power to define who is and who is not fully accepted, rejection by privileged groups implies that one’s group is excluded from the most valued positions in society. Many theoretical approaches predict that feeling excluded in this way will harm self-esteem (Baumeister & Leary, 1995; Cooley, 1956; Mead, 1934; Rosenberg, 1979; Tajfel & Turner, 1986), and empirical research has supported the contention that such exclusion is painful, resulting in anxiety, depression, feeling a lack of control, and lowered self-esteem (Baumeister & Tice, 1990; Bowlby, 1973; Cozzarelli & Karafa, 1998; Frable, 1993; Leary, Tambor, Terdal, & Downs, 1995; Solomon, Greenberg, & Pyszczynski, 1991; Williams, Shore, & Grahe, 1998). In contrast, when privileged group members are rejected by the disadvantaged, it carries no implications for exclusion from the wider culture. Thus, because attributions to prejudice among privileged groups do not represent exclusion of one’s social identity from the dominant society as a whole, they are less painful than when the same attributions are made by disadvantaged group members.

Recognizing disadvantage is painful

As we have argued, the attributions to prejudice made by disadvantaged groups differ on a number of dimensions from those made by privileged groups. Attributions to prejudice made by privileged groups are less stable, and are less likely to reflect a general lack of control or rejection by the dominant culture. In contrast, attributions to prejudice among members of disadvantaged groups will be more stable, more likely to reflect a lack of control over one’s life, and more likely to represent exclusion from the culture. Stated another way, in disadvantaged groups, the total relevance of an attribution to prejudice extends far beyond the immediate situation; however, for members of privileged groups, the total relevance of an attribution to prejudice is virtually identical to its local relevance. Thus, attributions to prejudice should be more harmful to the well-being of disadvantaged groups than for privileged groups.

Empirical support

Research on a variety of disadvantaged groups has demonstrated that recognizing prejudice against one’s social group is negatively related to psychological well-being. For instance, women (Kobrynowicz & Branscombe, 1997; Landrine, Klonoff, Gibbs, Manning, & Lund, 1995) and African‑Americans (Branscombe et al., 1999; Cross & Strauss, 1998; Klonoff & Landrine, 1999; Williams, Yu, Jackson, & Anderson, 1997) who perceive pervasive discrimination against their group are more likely to exhibit debilitating psychiatric and physical health symptoms compared to those who perceive less discrimination. In a recent review, Clark, Anderson, Clark, and Williams (1999) concluded that perceived racism among African Americans results in psychological and physiological stress responses which have a number of long-term negative health effects. Such costs of making attributions to prejudice are not limited to women or ethnic minority groups. In an study of lesbian and gay crime victims, those who attributed the crime to prejudice against their sexual orientation experienced more symptoms of depression, anxiety, vulnerability, and posttraumatic stress than did those who attributed the crime to other causes (Herek, Gillis, & Cogan, 1999). Consistent with our perspective, Herek et al. concluded that the greater costs associated with attributing victimization to prejudice occur because they link the negative outcome with one’s stable social identity.

Although the repeatedly observed negative relationship between perceived discrimination and well-being is compelling, one might argue that this finding results from the reverse causal direction. This alternative explanation would suggest that those who generally see the world as “out to get them” are consequently more likely to see themselves as the target of prejudice (e.g., Crocker, Luhtanen, Broadnax, & Blaine, 1999; Kramer, 1998). However, this “paranoia” explanation is less plausible as an account of the relationship between physical health outcomes and perceived discrimination (Clark et al., 1999). Further, experimental studies have found that, in fact, attributions to prejudice do harm well-being among members of disadvantaged groups. In one of the earliest experimental investigations of this issue, Dion and Earn (1975) found that Jewish participants who could attribute their failure to a Gentile’s anti-Semitism felt more stress and negative affect than those in experimental conditions where an attribution to prejudice was implausible.

In a recent study from our own lab (Schmitt, Branscombe, & Postmes, 2000), women read an article that argued that sexism and discrimination against women was either pervasive or rare. The manipulation was very successful. In the rare condition participants reported that 27% of men were biased against women, while women in the pervasive condition believed that 51% of men were sexist. Participants also reported expecting to encounter gender discrimination more often in the pervasive condition than in the rare condition. Women in the pervasive condition had significantly lower personal self‑esteem and collective self‑esteem compared to those in the rare condition. Thus, the more that women believed that future negative events in their lives might be explained by prejudice, the worse they felt about themselves and their group membership.

In a second study (Schmitt, Branscombe, & Postmes, 2000) we found that for a single negative performance outcome, attributions to prejudice were most harmful when prejudice was seen as pervasive. In the context of a mock job interview, women received extremely negative evaluations from a man, who was ostensibly a businessperson who made real hiring decisions. Just before the participant received the negative feedback, a female experimenter confided in the participant, telling her about how the evaluator had evaluated past participants. In the “Jerk” condition, the female experimenter suggested that the interviewer was a “jerk” who evaluated almost everyone negatively. In the “Lone Sexist” condition, the experimenter told the participant that the interviewer was sexist, but that the other 19 interviewers helping with the study were not. In the “Pervasive Sexism” condition, participants learned that all 20 interviewers involved in the study were biased against women. Manipulation checks revealed that participants were more likely to attribute the negative feedback to prejudice in the Lone Sexist and Pervasive conditions compared to the jerk condition. Furthermore, participants saw sexism as more pervasive in the pervasive condition than in the other two conditions. As predicted, women reported the most depressed affect in the condition in which the negative feedback came from a sexist evaluator and sexism was pervasive, and the least depressed affect when the feedback was not perceived as due to sexism. Depression in the Lone Sexist condition fell in between the other two conditions. Thus, attributions to pervasive prejudice are particularly harmful to the well-being of members of disadvantaged groups.

Although perceiving prejudice harms well-being among disadvantaged group members, it does little harm to, and sometimes even benefits, members of privileged groups. In a correlational study, Schmitt, Branscombe, Kobyrnowicz and Owen (2000) found that perceptions of discrimination against one’s gender were unrelated to well-being in men, but they were negatively correlated in women. In an experimental study, Branscombe (1998) asked participants to think about and list the consequences of their group membership. Women and men were randomly assigned either to list the disadvantages that they had received because of their gender group membership, or to list the privileges they received because of their gender. Men considering their gender group’s disadvantages exhibited higher personal self-esteem than men who thought of their gender group’s advantages. Women, however, showed the reverse trend.

In an important set of experiments examining the differential costs and benefits of attributions to prejudice for privileged and disadvantaged groups, Ruggiero and Marx (1999) manipulated the plausibility of attributions to prejudice for a negative evaluation among women and men (Study 1), and Blacks and Whites (Study 2). In the condition in which an attribution to prejudice was plausible, a confederate posing as another participant whispered to the participant that she or he had heard that one of the two evaluators was biased and rated the groups in the study differently. In the non-suspicious condition, the confederate simply whispered that she or he hoped that the experiment would be over quickly. A measure of attributions to prejudice revealed that the manipulation of the plausibility of such attributions was successful in both studies. Examination of the effects of attributional condition separately for privileged and disadvantaged groups revealed that attributions to prejudice were costly for disadvantaged groups, but protective for privileged groups (K. M. Ruggiero, May 3, 2000 ). Disadvantaged groups felt better about themselves and more in control when the negative evaluation was attributable to their own individual performance compared to when it was attributable to prejudice. Among privileged groups, attributions to prejudice did not harm self-esteem or perceived control. In addition, attributions to prejudice protected perceptions of one’s performance among men and Whites, but not among women and Blacks. These results provide the strongest evidence thus far that attributions to prejudice are not self-protective for members of disadvantaged groups, even when the alternative explanation for a negative outcome is one’s own performance.

Reconsidering the evidence for the discounting hypothesis

The discounting hypothesis suggests that disadvantaged groups may be motivated to make attributions to prejudice because such attributions are self-protective (Crocker & Major, 1989; Crocker et al., 1998). However, studies of disadvantaged groups have not provided strong support for the self-protection hypothesis. Dion’s (1975) early study of attributions to prejudice manipulated the perceived likelihood that sexism could explain the negative treatment that female participants received by varying the gender of the source of the treatment. Contrary to the self-protection hypothesis, participants felt worse about themselves when attributions to prejudice were most plausible (i.e., when the source of the negative treatment was a man), compared to when it was less plausible (the negative treatment came from a woman). In favor of the discounting perspective, Crocker et al. (1991, Study 1) found that women who received negative evaluations from a sexist male evaluator reported less depressed affect compared to those who received negative feedback from a non-sexist evaluator, although no support for the discounting hypothesis was obtained on a measure of self-esteem. In Study 2, the discounting hypothesis was not supported. Black participants who received negative feedback did not significantly differ in terms of mood or self-esteem as a function of whether the feedback was attributable to prejudice or not.

If attributions to prejudice do have self-protective consequences for members of disadvantaged groups, it is logical to suspect that members of disadvantaged groups might be motivated to make attributions to prejudice when those attributions are at least somewhat plausible. Any evidence that stigmatized individuals “overuse” attributions to prejudice and discrimination could be therefore interpreted as support for the discounting hypothesis. However, the evidence of vigilance on the part of disadvantaged groups is especially dubious. Not only is this hypothesis inconsistent with a large body of work finding that the disadvantaged groups minimize the extent to which they perceive the painful reality of their own disadvantage (Crosby, Pufall, Snyder, O’Connell, & Whalen, 1989), when it has been carefully tested experimentally it has not been supported. Ruggiero and Taylor (1995, 1997) found that women, Asian Americans, and Black Americans who had received negative feedback were quite reluctant to attribute their failure to an evaluator’s prejudice, and did so only when prejudice was presented as a virtual certainty.

Kleck and Strenta (1980) are often cited as demonstrating that the self-protective properties of an attribution to prejudice lead the disadvantaged to make such attributions even when they are a logical impossibility. In their study, the experimenter applied a fake facial scar on participants, but then secretly removed it while ostensibly applying moisturizer to the make up. After interacting with another person, participants who believed that they had been given the facial disfigurement perceived a strong bias against them. However, attributions to prejudice against a fake, temporary scar would neither be internal nor stable, and would not raise concerns of potential devaluation or rejection in the future. For that reason, the meaning of the attributions to prejudice made in the context of Kleck and Strenta’s experiment are very different from those made by members of historically disadvantaged groups. The aversive qualities of attributions to prejudice that reflect the experience of real disadvantaged groups is absent.

In a study comparing the attributions of Black and White participants, Crocker et al. (1991) found a racial group main effect where “Black subjects were more likely than White subjects to attribute the feedback they received to prejudice when they received negative rather than positive feedback and when the evaluator could see them, hence they were more aware of their race, than when they were not seen” (p. 225), and interpreted this as evidence that disadvantaged groups might be overly vigilant of prejudice against them. However, because of a confound of the design, there is a more plausible explanation for these data. In their research, both Black and White participants received negative feedback from a White evaluator–an outgroup member in the former case, but an ingroup member in the latter. While the Black participants’ racial prejudice attribution ratings were indeed higher than those of the White participants, as the authors themselves note, this is likely to have resulted from a confound between participant race and whether the evaluator was an ingroup or outgroup member. In order to compare Whites and Blacks in terms of the degree of attributional ambiguity and willingness to make attributions to prejudice, attributions to prejudice must be at least somewhat plausible for both groups. For Whites, being evaluated by a White evaluator clearly creates a context in which attributions to prejudice are implausible, and for that reason White participants were less likely to make attributions to prejudice compared to Blacks for whom an attribution to prejudice was plausible. Despite this alternative explanation for the observed racial difference, social psychologists continue to cite this study as evidence that disadvantaged groups are more motivated to make attributions to prejudice than are privileged groups. However, this racial difference in attributions really indicates that participants realistically assume that Whites are more likely to be prejudiced against Blacks than Whites are likely to be prejudiced against Whites.

Recent empirical work which does not contain the confound present in Crocker et al. (1991), has found that the differences between privileged and disadvantaged groups in willingness to make attributions to prejudice show a quite different pattern than Crocker et al. observed. When Ruggiero and Major (1998) controlled for the plausibility of discrimination for a particular event by manipulating the probability of discrimination from an outgroup member orthogonally to group membership, disadvantaged groups (Blacks and women) minimized perceptions of discrimination against them, but privileged groups (Whites and men) appeared to be vigilant in detecting it. Thus, the motivation to exaggerate the possibility of discrimination (given the opportunity) appears to be present in privileged groups, while the opposite–the motivation to minimize the possibility of discrimination–appears to be present in disadvantaged groups.

In another study with interpretational problems, Crocker et al. (1999) concluded that because Blacks Americans rated anti-Black conspiracies as more plausible than did Whites, Blacks are vigilant in detecting systematic discrimination against them. However, this racial difference could also be explained by Whites’ lack of awareness or denial of the existence of real anti-Black conspiracies. Recently, we conducted a study (Nelson, Branscombe, & Schmitt, 2000) in which Black and White participants evaluated the plausibility of a bogus newspaper article describing an alleged White conspiracy against Blacks. As in the Crocker et al. (1999) study, Blacks did rate the anti-Black conspiracy as more plausible than did Whites. However, we found that this racial difference was eliminated when we controlled for Blacks’ greater knowledge of real past conspiracies against Blacks (e.g., the Tuskegee experiment). These results suggest that Crocker et al.’s finding that Blacks see anti-Black conspiracies as more plausible than do Whites is not due to Black’s motivation to protect personal or group-based self-esteem, but rather that Whites are unaware of or deny the reality of past conspiracies by Whites against Blacks.

Are attributions to prejudice ever self-protective?

As Crocker and Major (1989) suggested, attributions to prejudice can under some circumstances protect perceptions of one’s performance or ability by discounting them as causes of the negative outcome (Britt & Crandall, 2000; Ruggiero & Taylor, 1997). However, for disadvantaged groups, attributions to prejudice also imply stable exclusion making them mostly harmful, as the research we reviewed indicates. In contrast, when members of privileged groups make attributions to prejudice they endure less harmful consequences and reap more psychological benefits than do members of disadvantaged groups. Privileged groups can discount their own personal qualities as causes of the specific negative outcome without simultaneously feeling systematically devalued and disadvantaged.

Research on disadvantaged groups provides strong evidence that attributions to prejudice are even more costly than attributions to one’s performance (Ruggiero & Marx, 1999; Ruggiero & Taylor, 1997). Perhaps this is because one’s performance, qualifications, and even ability in a domain are perceived as less stable than prejudice against one’s group membership. We can, however imagine a possible situation where an attribution to personal qualities is more stable, and therefore, more threatening, than an attribution to prejudice. Attributing a lack of long-term success in a highly valued domain to a stable personal characteristic might well be a worse alternative to an attribution to prejudice. In such situations, discrimination may be seen as a less stable impediment to future success than one’s personal traits. In that context, an attribution to prejudice, while still being an unpleasant attribution to make, might be somewhat less painful that the alternative.

That said, we suggest that such situations are relatively rare, in part because in most situations many other more psychologically comforting explanations for negative outcomes exist. The discounting perspective considers only two attributional alternatives for negative events: that they “could be due to one’s lack of merit, inferior qualifications, poor performance, or other shortcomings,” or “due to prejudice and discrimination against one’s social identity” (Crocker et al., 1998, p. 520). However, in many situations where people face a negative outcome these are not the only two types of attributions that might be plausible. For instance, one could attribute a negative outcome to truly external factors (e.g., another’s negative disposition toward everyone, aspects of the situation that could interfere with performance, task difficulty, or chance) or transient internal factors (e.g., fatigue or mood). Furthermore, in many situations in which members of disadvantaged groups regularly face negative treatment, an attribution to personal qualities, ability, or performance are highly implausible, because the negative treatment occurs before there is even an opportunity to demonstrate one’s deservingness or lack thereof. In future research, to fully gauge the costs or benefits of an attributions to prejudice, we must consider such attributions within a wider range of plausible attributions.

The preponderance of the data suggests that attributions to prejudice among disadvantaged groups are painful. Thus, our review is inconsistent with Crocker et al.’s (1991) speculation that “members of stigmatized groups who believe that they are discriminated against or that others are racist should be more likely to attribute negative feedback to prejudice and therefore may be higher in self-esteem” (p. 226). In contrast, our perspective suggests that the relative frequency of disadvantaged groups’ encounters with prejudice make such attributions especially harmful, while the self-protective properties of such attributions are mainly limited to members of privileged groups, who are likely to encounter prejudice relatively infrequently.

Coping with prejudice in disadvantaged groups

Both correlational and experimental evidence attests to the psychological costs of attributions to prejudice among members of disadvantaged groups. Given these costs, we now consider how members of disadvantaged groups cope with the pain of recognizing their disadvantaged status. Crocker and Major (1989) argued that members of disadvantaged groups do not simply passively accept the dominant culture’s devaluation of their social identity, but resist the internalization of that devaluation. Our perspective suggests that group identification is one important means by which disadvantaged groups cope with the pain of recognizing their devaluation.

The Rejection-Identification Model

The “Rejection‑Identification Model” of attributions to prejudice (Branscombe et al., 1999) suggests that psychological identification with one’s disadvantaged group attenuates the negative effects of attributions to prejudice on well-being. As shown in Figure 1, the Rejection-Identification Model predicts that perceiving prejudice directed against one’s disadvantaged group harms psychological well-being. However, in response to prejudice, disadvantaged group members increasingly identify with their disadvantaged group which in turn alleviates some of these harmful consequences. Despite the fact that recognizing group-based disadvantage harms well-being, responding to that threat by identifying with that group tends to offer psychological benefits to counter that harm. Indeed, minority group identification is associated with less depression (Munford, 1994), more positive self‑esteem, (Bat‑Chava, 1994; Grossman, Wirt, & Davids, 1985; Phinney, 1990), and other measures of psychological adjustment (Rowley, Sellers, Chavous & Smith, 1998). For persons with bulimia and gay, bisexual, and lesbian individuals, the mere presence of similarly stigmatized others raises self‑esteem and lowers depression and anxiety (Frable, Pratt, & Hoey, 1998; McKenna & Bargh, 1998). In the context of failure on a verbal ability test (a domain in which Black Americans are likely to see themselves as disadvantaged), Black Americans who were reminded of their racial identity felt better than those who were not reminded (Major, Spencer, Schmader, Wolfe, & Crocker, 1998).

Threats to the ingroup generally encourage group identification (Turner, Hogg, Turner, & Smith, 1984). As the Rejection-Identification model predicts, recognizing prejudice and minority group identification are correlated among Jews (Rollins, 1973), women (Gurin & Townsend, 1986), African‑Americans (Gurin, Gurin, Lao, & Beattie, 1969), lesbians (Crosby et al., 1989), and non‑mainstream college groups (e.g., punks, hippies, nerds; Cozzarelli & Karafa, 1998). In order to experimentally test the causal direction of this relationship, we conducted a study in which we manipulated, rather than measured, perceived discrimination against one’s group (Jetten, Branscombe, Schmitt, & Spears, in press). In a sample of people with body piercings, we provided participants with bogus information about the pervasiveness of anti-piercing prejudice among members of the mainstream. In one condition, participants were told that the majority of the mainstream disliked them and would discriminate against them. In another condition, participants were lead to believe that treatment by the mainstream was more positive. Consistent with predictions, participants in the pervasive discrimination condition reported significantly more identification with other group members than did participants in the positive treatment condition.

Although the reverse effect of group identification on perceptions of discrimination might also be possible (as hypothesized by Crocker & Major, 1989; Crocker et al., 1998), we are aware of no experimental evidence supporting this causal direction. Furthermore, a number of other researchers have found experimental evidence of the Rejection-Identification Model’s prediction that attributions to prejudice encourage minority group identification. Dion (1975) found that women who had attributed their failure to sexism subsequently described themselves more favorably on positive aspects of the stereotype of women. Similarly, Dion and Earn (1975) found that Jewish participants who were lead to believe that their failure at a task was attributable to a Gentile’s anti-Semitism described themselves more in terms of the positive aspects of the Jewish stereotype compared to participants in the non-prejudice condition. Dion, Earn, and Yee (1978) argued that these two studies suggest that attributions to prejudice elicit a heightened identification with the positive aspects of one’s minority group. Similarly, when women are made aware of their stigmatized status, they are more likely to self-stereotype in terms of their gender (Foster & Matheson, 1999; Hogg & Turner, 1987). Reminding gay men of anti-gay prejudice increases their identification with the gay movement (Simon et al., 1998). More generally, when power relationships among experimentally created groups were manipulated, Dépret and Fiske (1993) found that the lower power group showed increased allegiance to their ingroup compared to the more powerful group. All of these studies provide support for the hypothesis that perceiving prejudice leads members of disadvantaged groups to see themselves in more group terms, and to feel greater emotional attachment to that group.

Empirical tests of the Rejection-Identification Model

Thus far, we have reviewed evidence that is consistent with the three relationships predicted by the Rejection-Identification Model: 1) Perceptions of pervasive prejudice will harm psychological well-being in disadvantaged groups; 2) Perceptions of pervasive prejudice will encourage identification with one’s disadvantaged group; 3) Identification with one’s disadvantaged group will positively affect psychological well-being. However, examination of the Rejection-Identification Model is incomplete without assessing these relationships simultaneously in disadvantaged group members. Such tests of the model are crucial because the direct negative effect of perceiving prejudice on well-being is expected to be suppressed by the positive effects of coping via group identification. Below we describe research testing the full Rejection-Identification Model.

African Americans. The Rejection-Identification Model was first tested with a sample of African Americans (Branscombe et al., 1999). Participants completed measures concerning past experiences with racial discrimination, attributions to prejudice for hypothetical instances of negative treatment across a variety of life domains, minority group identification, hostility toward Whites, and well-being. Personal well-being was measured with the Rosenberg Personal Self-esteem Inventory ( Rosenberg , 1979), a measure of the frequency of negative emotions, and collective self-esteem. Using structural equation modeling, we found that the hypothesized Rejection-Identification Model fit the data extremely well. Furthermore, as shown in Figure 2, all of the hypothesized effects were significant and in the predicted direction. Perceiving prejudice encouraged hostility toward Whites and directly harmed psychological well-being. However, African Americans coped with this harm to well-being to some extent by increasing identification with their racial group. Thus, consistent with the Rejection-Identification Model, by turning toward the ingroup in response to prejudice, the psychological benefits of group identification counter some of the psychological costs of recognizing the disadvantaged status of one’s group.

Women and men. In order to test the model with a non-ethnic disadvantaged social group, and to assess how the model fits differently for disadvantaged and privileged groups, we asked White women and men to complete measures of perceived prejudice against their gender group, gender group identification, the Rosenberg Self-Esteem Inventory, the Beck Depression scale, the Taylor Manifest Anxiety Inventory, and general life satisfaction (Schmitt, Branscombe, Kobrynowicz, & Owen, 2000). Consistent with the findings obtained with African Americans, the more women perceived discrimination based on their gender group membership, the poorer psychological adjustment they exhibited. Moreover, identification with women attenuated the negative effects of perceived prejudice. In other words, perceiving prejudice increased gender group identification, which in turn enhanced well-being in women. Thus, the Rejection-Identification Model fit the data for women extremely well.

Because of our basic assumption that attributions to prejudice will have different effects for privileged than for disadvantaged group members, in this research we compared the model fit for men as well as women. Consistent with our theoretical perspective, we found that the model did not fit the data for men. Unlike for women, perceived discrimination against their group had no significant impact on men’s psychological well-being. Thus, in a privileged group, attributions to prejudice were not harmful as they were in disadvantaged groups. In addition, perceived discrimination did not encourage identification with their gender group, suggesting that because privileged groups do not experience attributions to prejudice as threatening, there is no need to turn toward their group as a means of coping. The differential pattern of effects observed for women and men implies that increased identification in response to prejudice is serving defensive purposes in the disadvantaged group that are not needed by privileged groups.

Future directions for the Rejection-Identification Model

The Rejection-Identification Model provides a basis for studying the subjective experience of disadvantaged groups from a social identity perspective. Social identity theory guided the creation of the model, and offers a number of ways in which the model can be expanded in the future. Below we describe how the Rejection-Identification Model can be used as a theoretical vehicle for examining the processes by which group identification protects well-being, and the factors that might moderate how well the disadvantaged cope.

The psychological benefits of group identification

Although a number of empirical studies have documented that identification with one’s disadvantaged group is associated with psychological well-being, more work needs to be done to examine the processes by which group identification protects the well-being of members of disadvantaged groups. At first it may seem paradoxical to suggest, as the Rejection-Identification Model does, that people can make a threatening intergroup comparison in which they recognize their group’s disadvantage relative to other more privileged groups, and simultaneously benefit psychologically by identifying with that group. Social identity theory suggests that group identification can lead to two general strategies for constructing a positive social identity despite threatening comparisons with more privileged groups–that is, social creativity and social competition (Tajfel, 1978). When adopting the social creativity strategy, the disadvantaged shift their intergroup comparisons to dimensions on which their group fares more favorably, and increase the relevance of those dimension to the group’s identity. Crocker and Major (1989) also suggest that such “selective devaluing” is one way that disadvantaged groups cope with a negative social identity. Research supports the idea that disadvantaged groups acknowledge reality and admit that the outgroup is superior on status defining dimensions (e.g. wealth), but see the ingroup as superior on non-status defining dimensions (Ellemers, van Rijswijk, Roefs, & Simons, 1997; Lemaine, 1974). In political terms, social creativity amounts to rejecting the norms and standards of the dominant culture, which disadvantages and devalues the ingroup. In fact, rejection of mainstream norms–which serve the interests of privileged groups–is probably a necessary condition for building a positive minority identity. In Tajfel’s (1978) words, “differentiation [from the mainstream] often represents, socially and politically, a rejection of the status quo by groups which perceive themselves as separate and socially disadvantaged” (p. 7).

Tajfel (1978) identified social competition as another outcome of group identification which will be adopted by disadvantaged groups who perceive alternatives to their group’s disadvantage as possible. In support of this idea, a number of empirical investigations have found that group identification is an important predictor of willingness to engage in collective action on behalf of the group (Kelly & Breinlinger, 1996; Kelly & Kelly, 1994; Veenstra & Haslam, 2000). By engaging in attempts at social change, members of disadvantaged groups can actively alter the relative status of their group, and create a more positive group identity based in a positive future for the group (De Vos & Suarez-Orozco, 1990). While recognizing prejudice harms perceptions of personal control, group identification may provide an alternative basis for a sense of control–that of collective efficacy (Bandura, 1997). Indeed, social change movements typically acknowledge that individual control over the social system is impossible, but believe that collective action can bring about a more positive outcome for the group as a whole (see Kelly & Breinlinger, 1996; Klandemans, 1997).

Group identification might also protect well-being by providing greater opportunities for social support from other ingroup members. Such social support can be provided in the form of direct emotional support or in terms of shared time and resources which can help counter some of the material costs of discrimination. Group identification might also counter the psychological costs of feeling rejected and devalued by the dominant culture by providing disadvantaged groups with a sense of belonging and acceptance. Therefore, when members of disadvantaged groups perceive that acceptance and fair treatment by the more powerful group is unlikely, increasing psychological investment in their ingroup may be the best possible strategy for feeling accepted and enhancing psychological well-being.

Future work examining the psychological benefits of group identification among the disadvantaged should consider the mediating processes by which identification affects subjective well-being. Social creativity, social competition, increased social support, and feelings of acceptance are all potential routes by which group identification could protect psychological well-being.

Factors that block group identification

Although the Rejection-Identification Model predicts that perceptions of pervasive discrimination lead members of disadvantaged groups to increase their identification with the ingroup, the more general theory on which the model is based–social identity theory–suggests that under certain social conditions, group identification will not increase in response to perceived disadvantage. As Tajfel (1978) proposed and a number of empirical studies have confirmed, group identification is less likely when the boundary between the minority and the majority is blurred by the apparent potential for individual upward social mobility (Ellemers, van Knippenberg, de Vries, & Wilke, 1988; Ellemers, van Knippenberg, & Wilke, 1990; Mummendey, Kessler, Klink, & Mielke, 1999), when alternatives to the current social structure are difficult to imagine (Ellemers et al., 1990; Martin ,1986), and when the status quo is legitimized (Ellemers, Wilke, & van Knippenberg, 1993; Commins & Lockwood, 1979).

Crocker and Major (1994) have also considered the importance of the perceived legitimacy of prejudice and discrimination as a moderator of the costs of making attributions to prejudice. They argued that while an attribution to illegitimately held prejudice can be self-protective, making an attribution to prejudice that is seen as justifiable is costly. Research on “overweight” women supports the idea that attributions to legitimately held prejudice are harmful to well-being (Crocker, Cornwell, & Major, 1993), and that believing in or being reminded of ideologies which legitimize prejudice against the overweight harms psychological well-being (Quinn & Crocker, 1999). While we do agree that attributions to prejudice are more harmful when prejudice is seen as legitimate, compared to when it is seen as illegitimate, we reach this conclusion for a different reason. Social identity theory suggests that the perceived legitimacy of prejudice blocks the possibility of coping with disadvantage through group identification, making perceptions of prejudice and discrimination especially harmful. We (Garza, Branscombe, Schmitt, & Zarate, 2000) found support for this idea in a study of Mexican Americans. Participants completed measures of the perceived legitimacy of discrimination, perceived pervasiveness of prejudice, group identification, and psychological well-being. As expected, we replicated our findings with African-Americans and women among those who saw prejudice against Mexican-Americans as illegitimate. Among those participants, perceived prejudice was positively related to group identification, and group identification partially attenuated the negative consequences of perceiving discrimination on well-being. However, among Mexican Americans who saw prejudice against them as legitimate, the relationship between perceived discrimination and minority group identification was non-significant. As a result, the overall effect of perceived discrimination was more harmful among those who saw it as legitimate, compared to those who saw it as illegitimate, even though the direct negative effect was equivalent in both cases.

Like legitimacy, other factors, such as the perceived mutability of the social structure and perceived opportunities for individual upward social mobility, are likely to moderate when perceived rejection from the mainstream will encourage coping via group identification. Factors that block coping via group identification can arise from a number of sources, but of particular interest are aspects of culture that are primarily controlled by the powerful. For example, disadvantage can be legitimized by beliefs in the inferiority of disadvantaged groups (e.g., “old-fashioned” racism), or the controllability of their stigma (e.g. weight, sexual orientation). Similarly, group identification might be discouraged by ideologies that make individual upward mobility appear possible despite the group’s overall disadvantage (e.g., Protestant Work Ethic, the “American Dream” or through the practice of “tokenism”; Ellemers et al., 1993; Taylor & McKirnan, 1984; Wright, 1997; Wright & Taylor, 1998; Wright, Taylor, & Moghaddam, 1990). In addition, ideologies that make inequality appear immutable or inevitable (e.g., those that explain inequality in terms of its functionality or evolutionary adaptiveness) may also affect how the disadvantaged cope by discouraging group identification. More generally, ideologies that endorse an individualistic perspective on human nature (political perspectives advocating “color-blindness”; Ferdman, 1989) may discourage perceptions of discrimination and thus, minority group identification.

Not only might these ideologies harm members of disadvantaged groups by blocking group identification as a means of psychological well-being protection, but they simultaneously discourage collective action and social change. Social psychologists have noted that the ideologies promoted by privileged groups often protect their position in the social structure by justifying prejudice and discrimination against the disadvantaged (Biernat, Vescio, & Theno, 1996; Jackman, 1994; Katz & Hass, 1988); however, these ideologies also protect the status quo by affecting and limiting how the disadvantaged cope with and resist their structural position in society (Jost, 1995; Marx & Engels, 1864). Future research should investigate the role of dominant ideologies in maintaining the status quo by reducing the likelihood of group identification and collective action by the disadvantaged.

Challenges for future research on attributions to prejudice

As social psychologists continue to research this important and socially meaningful area of study, they are likely to encounter a variety of methodological and theoretical challenges. These difficulties arise, in part, because the attributional processes underlying the meaning of perceived discrimination do not occur in a “social vacuum.” They operate in relation to the larger social context. In order to capture the total relevance or meaning of an attribution to prejudice, researchers must consider the individual’s position within the larger social context of the existing intergroup relations.

Attributions to prejudice can vary along a number of dimensions which have important implications for their consequences. For this reason, generalization beyond the specific laboratory context requires close attention to where the attributions made in that context fall on those dimensions. First and foremost, researchers must ensure that the stigma employed in their studies is an actual quality of the participants. Perceived discrimination on the basis of an attribute that the individual knows he/she does not in fact possess, will be attributed externally, while attributions to prejudice made by real members of disadvantaged groups are, in part, internal. Secondly, because stability and perceived legitimacy are important moderators of the effects of attributions to prejudice, researchers should be cognizant of what the laboratory context suggests to participants about the pervasiveness and legitimacy of prejudice. For example, discrimination coming from a single individual might imply that discrimination is less pervasive than if the discrimination comes from the outgroup as a whole (see Abelson, Dasgupta, Park, & Banaji, 1998). In addition, when groups are created ostensibly on the basis of performance, coping via group identification is less likely because discrimination against the low performance group is likely to be perceived as legitimate. In sum, all forms of discrimination are not experienced identically. Thus, researchers must create in their labs or examine in the field the kinds of psychological contexts that actually reflect the psychological dilemmas faced by the disadvantaged groups of interest. In addition, researchers should consider the important role of identification with one’s social group as a means of coping with perceived discrimination. As we have shown, group identification suppresses some of the harmful effects of attributions to prejudice; consequently, the harmful effects of such attributions are more evident when identification is included in the analysis. Thus, researchers who do not include group identification in their analyses are likely to underestimate the direct negative effects of attributions to prejudice.

Conclusions

Crocker and Major (1989) brought the issue of attributions to prejudice in disadvantaged groups to the forefront of mainstream social psychology. That being said, there are important differences between our theoretical perspective and that of our colleagues. Crocker and Major conceptualized attributions to prejudice in comparison to attributions to personal aspects of the self, such as ability, performance, and qualifications. Accordingly, they hypothesized that attributions to prejudice are self-protective because they avoid the alternative attributional possibility–that negative outcomes are the result of a personal lack of deservingness. Approaching this topic from a social identity perspective, we conceptualized attributions to prejudice in disadvantaged groups using a different point of comparison–attributions to prejudice made by members of privileged groups. Attributions to prejudice do not just discount personal aspects of the self, but actually implicate another important aspect of the self–one’s group membership. Because the consequences of attributions to prejudice are likely to depend on the meaning of its internal component, we considered how attributions to prejudice have different meanings depending on the position of the ingroup within the social structure. For disadvantaged groups, attributions to prejudice are likely to be relatively stable, uncontrollable, and reflective of widespread rejection and devaluation. For privileged groups, however, attributions to prejudice are likely to be far more unstable and controllable, with very localized implications. For this reason, we hypothesized that while attributions to prejudice are less harmful and potentially self-protective for members of privileged groups, they are detrimental to the psychological well-being of the disadvantaged.

We perceive social psychologists as having generally accepted the idea that disadvantaged groups protect their psychological well-being by making attributions to prejudice for negative events, with very little empirical evidence to support this conclusion. Our review of the empirical research revealed strong support for the hypothesis that for disadvantaged groups, recognizing prejudice is not self-protective, and is, in fact painful. In a variety of disadvantaged social groups, we found that the disadvantaged cope with the psychological pain of perceiving prejudice by increasing identification with their ingroup. However, this coping strategy only partially alleviates the harm to psychological well-being caused by attributions to prejudice. We suggest that future work on attributions to prejudice examine how aspects of the culture discourage group identification as a means of coping with the recognition of devaluation.

References

Abelson, R. P., Dasgupta, N., Park, J., & Banaji, M. R. (1998). Perceptions of the collective other. Personality and Social Psychology Review, 2, 243-250.

Anderson, C. A., Krull, D. S., & Weiner, B. (1996). Explanations: Processes and consequences. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 271-296). New York : Guilford .

Bandura, A. (1997). Self-efficacy: The exercise of control. New York : W. H. Freeman.

Bat-Chava, Y. (1994). Group identification and self-esteem of deaf adults. Personality and Social Psychology Bulletin, 20, 494-502.

Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497-529.

Baumeister, R. F., & Tice, D. M. (1990). Anxiety and social exclusion. Journal of Social and Clinical Psychology, 9, 165-195.

Biernat, M., Vescio, T. K., Theno, S. A. (1996). Violating American values: A "value congruence" approach to understanding outgroup attitudes. Journal of Experimental Social Psychology, 32, 387‑410.

Bowlby (1973). Attachment and loss, Vol. 2: Separation anxiety and anger. New York : Basic Books.

Branscombe, N. R. (1998). Thinking about one's gender group's privileges or disadvantages: Consequences for well-being in women and men. British Journal of Social Psychology, 37,167‑184.

Branscombe, N. R., Schiffhauer K., Schmitt, M. T., & Valencia , L. (2000). Effects of thinking about White privilege or disadvantage and degree of White racial identification for feelings about the ingroup and outgroup. Unpublished manuscript, University of Kansas .

Branscombe, N. R., Schmitt, M. T., and Harvey, R. D. (1999). Perceiving pervasive discrimination among African‑Americans: Implications for group identification and well‑being. Journal of Personality and Social Psychology, 77, 135‑149.

Britt T. W., & Crandall, C. S. (2000). Acceptance of feedback by the stigmatized and the non-stigmatized: The mediating role of the motive of the evaluator. Group Processes and Intergroup Relations, 3, 79-95.

Brown, J. D., & Siegel, J. M. (1988). Attributions for negative life events and depression: The role of perceived control. Journal of Personality and Social Psychology, 54, 316-322.

Clark, R., Anderson, N. B., Clark, V. R., Williams, D. R. (1999). Racism as a stressor for African Americans: A biopsychosocial model. American Psychologist, 54, 805-816.

Commins, B., & Lockwood, J. (1979). The effects of status differences, favoured treatment and equity on intergroup comparisons. European Journal of Social Psychology, 9, 281-189.

Cooley, C. H. (1956). Human nature and the social order. New York : Free Press.

Cozzarelli, C., & Karafa, J. A. (1998). Cultural estrangement and terror management theory. Personality and Social Psychology Bulletin, 24, 253-267.

Crocker, J., Cornwell, B., & Major, B. (1993). The stigma of overweight: Affective consequences of attributional ambiguity. Journal of Personality and Social Psychology, 64, 60-70.

Crocker, J., Luhtanen, R., Broadnex, S., & Blaine, B. E. (1999). Belief in U.S. government conspiracies against Blacks among Black and White college students: Powerlessness or system blame? Personality and Social Psychology Bulletin, 25, 941-953.

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

Crocker, J., & Major, B. (1994). Reactions to stigma: The moderating role of justifications. In M. P. Zanna & J. M. Olson (Eds.) The psychology of prejudice: The Ontario symposium, Vol. 7. (pp. 289‑314). Hillsdale , NJ : Lawrence Erlbaum.

Crocker, J., Major, B., & Steele, (1998). Social stigma. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.) The handbook of social psychology, Vol. 2 (4th ed.) (pp. 504‑553). Boston , MA : McGraw‑Hill.

Crocker, J., & Quinn, D. (1998). Racism and self‑esteem. In J. L. Eberhardt, & S. T. Fiske, (Eds.), Confronting racism: The problem and the response (pp. 169‑187). Thousand Oaks , CA : Sage

Crocker, J., Voelkl, K., Testa, M., & Major, B. (1991). Social stigma: The affective consequences of attributional ambiguity. Journal of Personality and Social Psychology, 60, 218-228.

Croizet, J. C., & Claire, T. (1998). Extending the concept of stereotype threat to social class: The intellectual underperformance of students from low socioeconomic backgrounds. Personality and Social Psychology Bulletin, 24, 588-594.

Crosby , F. J., Pufall, A., Snyder, R. C., O'Connell, M., & Whalen, P. (1989). The denial of personal disadvantage among you, me, and all the other ostriches. In M. Crawford & M. Gentry (Eds.), Gender and Thought (pp. 79-99). New York : Springer-Verlag.

Cross, W., Jr., & Stauss, L. (1998). The everyday functions of African-American identity. In J. K. Swim & C. Stangor (Eds.), Prejudice: The target’s perspective (pp. 268-279). New York : Academic Press.

De Vos, G. A., & Suarez-Orozco, M. M. (1990). Ethnic belonging and status mobility. In G. A. De Vos & M. M. Suarez-Orozco (Eds.), Status Inequality: The Self in Culture (pp. 246-263). London : Sage.

Dépret, E., & Fiske, S. T. (1993). Social cognition and power: Some cognitive consequences of social structure as a source of control deprivation. In G. Weary, F. Gleicher, & K. L. Marsh (Eds.) Control Motivation and Social Cognition (pp. 176-202). New York : Springer-Verlag.

Dion, K. L. (1975). Women's reactions to discrimination from members of the same or opposite sex. Journal of Research in Personality, 9, 294-306.

Dion, K. L., & Earn, B. M. (1975). The phenomenology of being a target of prejudice. Journal of Personality and Social Psychology, 32, 944-950.

Dion, K. L., Earn, B. M., & Yee, P. H. N. (1978). The experience of being a victim of prejudice: An experimental approach. International Journal of Psychology, 13, 197-214.

Ellemers, N., van Knippenberg, A., de Vries, N., & Wilke, H. (1988). Social identification and permeability of group boundaries. European Journal of Social Psychology, 18, 497-513.

Ellemers, N., van Knippenberg, A., & Wilke, H. (1990). The influence of permeability of group boundaries and stability of group status on strategies of individual mobility and social change. British Journal of Social Psychology, 29, 233-246.

Ellemers, N., van Rijswijk , W., Roefs, M., & Simons, C. (1997). Bias in intergroup perceptions: Balancing group identity with social reality. Personality and Social Psychology Bulletin, 23, 186-198.

Ellemers, N., Wilke, H., & van Knippenberg, A. (1993). Effects of legitimacy of low group or individual status on individual and collective identity enhancement strategies. Journal of Personality and Social Psychology, 64, 766-778.

Ferdman, B. M. (1989). Affirmative action and the challenge of the color-blind perspective, In F. Blanchard & F. Crosby, (Eds.), Affirmative Action in Perspective (pp. 169-176). New York : Springer-Verlag.

Foster, M. D, & Matheson, K. (1998). Perceiving and feeling discrimination: Motivation or inhibition for collective action. Group Processes and Intergroup Relations, 1, 165-174.

Frable, D. E. S. (1993). Being and feeling unique: Statistical deviance and psychological marginality. Journal of Personality, 61, 85-110.

Frable, D. E. S., Blackstone, T., & Scherbaum, C. (1990). Marginal and mindful: Deviants in social interactions. Journal of Personality and Social Psychology, 59, 140-149.

Frable, D. E. S., Pratt, L., & Hoey, S. (1998). Concealable stigmas and positive self-perceptions: Feeling better around similar others. Journal of Personality and Social Psychology, 74, 909-922.

Garza, A. A., Branscombe, N. R., Schmitt, M. T., & Zarate, M. A. (2000). Coping With Prejudice Among Mexican Americans. Manuscript in progress.

Goffman, I. (1963). Stigma: Notes on the management of a spoiled identity. Englewood Cliffs, NJ: Prentice Hall.

Golin, S., Sweeney, P. D., & Shaeffer, D. E. (1981). The causality of causal attributions in depression: A cross-lagged panel correlational analysis. Journal of Abnormal Psychology, 90, 14-22.

Grossman, B., Wirt, R., & Davids, A. (1985). Self-esteem, ethnic identity, and behavioral adjustment among Anglo and Chicano adolescents in West Texas . Journal of Adolescence, 8, 57-68.

Gurin, P., Gurin, G., Loa, R. C., & Beattie, M. (1969). Internal-external c