Threatened Selves and Differential Prejudice Expression by White and Black Perceivers

Abstract

Previous theorizing suggests that often-stigmatized individuals may be just as likely, if not more likely, than infrequently stigmatized individuals to protect self-regard by derogating members of low-status groups after receiving negative feedback from high-status others. Often-stigmatized individuals, however, can discount criticism from these high-status others as reflecting prejudice, thereby making outgroup derogation unnecessary as an esteem-protective strategy. Replicating past research, White participants in Experiment 1 expressed prejudices after receiving negative feedback from a White evaluator; as predicted, however, Black participants did not. In Experiment 2, participants instead received negative feedback from Black evaluators (evaluators more likely to threaten Black participants’ self-regard). Here, contrary to previous theorizing, Black participants expressed prejudices, not toward another low-status group, but toward high-status Whites. In all, findings reveal flaws in previous assumptions that frequently stigmatized individuals may be especially prone to devalue lower-status others after rejection or negative feedback from members of higher-status groups.

Expressing prejudice against members of another group can buffer one’s self-esteem against failure or other self-image threats. For example, Fein and Spencer (1997) found that, after receiving false negative feedback on an intelligence test, participants were more likely to stereotype a gay target and respond negatively toward an ethnic minority target—actions that protected participants against feedback-based decrements in self-esteem (see also Collange, Fiske, & Sanitioso, 2009).

This esteem-buffering process has been studied primarily as a strategy employed by infrequently stigmatized individuals (e.g., Whites), although some have suggested it may be employed, perhaps to an even greater degree, by frequently stigmatized individuals. In his influential text, Gordon Allport (1954) hypothesized: “Victims of prejudice may, of course, inflict on others what they themselves receive…Pecked at by those higher in the pecking order, one may, like a fowl in the barnyard, peck at those seen as weaker and lower than oneself…” (p. 153). More recent research raises doubts about Allport’s hypothesis. For chronic targets of prejudice, the meaning of negative feedback from higher-status groups is ambiguous. It can reflect either an honest evaluation of one’s skills and abilities (thereby threatening one’s self-image) or evaluator prejudice (thereby enabling one to discount the criticism and reducing self-image implications) (Crocker & Major, 1989; Crocker, Voelkl, Testa, & Major, 1991; Mendes, Major, McCoy, & Blascovich, 2008).

Thus, the literature is unclear regarding how frequently-stigmatized targets will respond to other groups following negative feedback. Allport’s line of reasoning suggests that after receiving negative feedback from high-status (e.g., White) evaluators, members of lower-status groups (e.g., Blacks) should be just as likely, if not more likely, as members of high-status groups (e.g., Whites) to derogate other low-status groups (e.g., Native Americans). In contrast, research on how frequently-stigmatized groups construe negative feedback from higher-status groups suggests an alternative prediction (e.g., Crocker et al., 1991): Because criticism from Whites can be attributed to prejudices rather than personal failings, negative feedback from Whites may elicit relatively little self-image threat and, thus, little derogation of other low-status groups (Experiment 1).

The reasoning underlying this alternative prediction also implies an ancillary hypothesis—that, if self-threatening criticism from another cannot be readily discounted, as would be the case when it comes from a member of one’s own group, members of often-stigmatized groups might indeed derogate outgroup members. Consistent with Allport’s theorizing, some evidence suggests that such derogation would be focused toward members of other low-status groups (e.g., Cadinu & Reggiori, 2002). However, a closer consideration of the processes underlying esteem-enhancing derogation suggests an alternative prediction. Esteem-buffering derogation tends to be directed, generally, toward groups about whom one holds negative stereotypes (Spencer, Fein, Wolfe, Fong, & Dunn, 1998). Black Americans tend to hold more negative views of higher-status Whites than of members of other low-status groups (Cottrell, Richards, & Neuberg, 2009; Niemann, Jennings, Rozelle, Baxter, & Sullivan, 1994; Shapiro & Neuberg, 2008). Thus, we expected that, contrary to Allport’s assessment, when Blacks received criticism from an ingroup member—thereby eliciting self-image threat—they would be especially likely to derogate high-status (e.g., White), compared to low-status (e.g., Native American), targets (Experiment 2).

Experiment 1

Method

Seventy-three Black (43 female) and 64 White (31 female) students from Arizona State University (ASU) participated in exchange for course credit or ten dollars. Participants evaluated a job candidate after receiving negative feedback or no feedback from a White evaluator. The overall experimental design was a 2[Candidate Race: Native American (low-status)/White (high-status)] X 2(Negative Feedback: Yes/No) X 2(Participant Race: Black/White), fully between-participants with random assignment to the Candidate Race and Feedback conditions.

Participants were led to believe they were participating in two unrelated studies. In the first, participants completed a demographics questionnaire and their photograph was taped to the bottom. Next, participants were given five minutes to write an essay to be critiqued by a student “partner” ostensibly in another room. Experimenters stapled the essay and demographics questionnaire together in front of participants so there was no doubt that the partner would know the participant’s race when evaluating the essay.

The experimenter then left—ostensibly to bring the essay to the partner—and returned in five minutes. The experimenter gave all participants their partner’s demographics questionnaire. The partner was a White, same-gender senior English major. In the no-feedback condition, participants learned they would receive the critique at the end of the study. In the feedback condition, participants received a 9-item (e.g., extent to which the essay demonstrated intelligence) handwritten critique. The ostensible evaluator circled only 2s and 3s on 9-point response scales (1=most negative). The last question was an overall evaluation using a 1-10 scale (1=lowest score); the evaluator scored the essay a 2 and wrote “I thought the essay was very poor.”

The second study was described as an assessment of a managerial training program that would take place either after or before a set of managerial exercises. All participants were purportedly assigned randomly to complete the managerial exercises first. The first exercise was a 3-minute managerial decision task designed to bolster the cover story. Participants then received a job description along with a candidate’s application and resume. The materials were identical across participants with one exception: The applicant’s race was either low-status (Native American) or high-status (White). To evaluate the candidate, participants completed an 11-item measure of competence (α=.96) used by Shapiro and Neuberg (2008); all questions employed 9-point Likert-type scales (1=Not at all to 9=Extremely).

Results and Discussion

An Analysis of Variance (ANOVA) on competence evaluations revealed a main effect of participant race, F(1,129)=4.33, p=.04, η_p²=.03, qualified by the predicted three-way interaction, F(1,129)=5.72, p=.02, η_p²=.04 (Figure 1).

Figure 1.

Evaluations of job applicant competence as a function of job applicant race/ethnicity, participant race/ethnicity, and negative feedback from a White evaluator (Study 1). Error bars show standard errors.

For White participants, there were no main effects of candidate race or feedback (F<1). However, as anticipated, a significant interaction emerged, F(1,129)=6.76, p=.01, η_p²=.05: Replicating previous research, White participants evaluated the minority candidate more negatively after receiving negative feedback compared to no feedback, F(1,129)=4.91, p=.03, η_p²=.04. In contrast, the White candidate was not evaluated differentially as a function of receiving feedback, F(1,129)=2.69, p=.10. The overall pattern of data was primarily driven by receiving negative feedback: In the absence of feedback, White participants did not differentially evaluate the two candidates, F(1,129)=2.19, p=.14. After receiving feedback, however, White participants evaluated the minority candidate more negatively than the White candidate, F(1,129)=6.52, p=.01, η_p²=.05.

For Black participants a very different pattern emerged: There were no main effects of candidate race or feedback, and no interaction between these variables (all Fs<1). Thus, whereas negative feedback from a White evaluator significantly influenced White participants’ evaluations of the minority (compared to White) candidate, this feedback did not influence Black participants’ evaluations. Counter to Allport’s hypothesis, after being “pecked at” by high-status group members, chronically stigmatized individuals were unlikely to turn to outgroup derogation.

Our theorizing suggests that ambiguity as to whether negative evaluations from Whites constitute an accurate assessment of personal abilities may have buffered Blacks from experiencing self-image threat, thereby reducing the likelihood that Blacks will engage in outgroup derogation. However, an alternative explanation is simply that chronic targets of prejudice do not employ prejudice expression as a self-image maintenance strategy. Experiment 2 pits these alternatives against one another by having Whites and Blacks confront negative feedback from Black evaluators.

Experiment 2

Experiment 2 tested the hypothesis that Black participants receiving negative feedback from a Black evaluator would be especially likely to express prejudices. Unlike Study 1, because White participants can attribute negative feedback from a Black evaluator to anti-White prejudice, White participants should consequently escape self-image threat (Mendes et al., 2008) and thus have no need to express prejudices. However, because Black participants should not be able to attribute a Black evaluator’s negative feedback to anti-Black prejudice (Mendes et al., 2008), such feedback may threaten self-image and facilitate prejudice expression.

If Blacks receiving negative feedback from other Blacks become more inclined to derogate, whom do they target? Allport’s hypothesis was that stigmatized individuals should derogate other low- (or lower-) status targets (see also Cadinu & Reggiori, 2002). There are compelling reasons to believe, however, that the focus of esteem-enhancing derogation within chronically stigmatized individuals would not be members of lower-status groups per se. Self-image threats increase the salience of negative stereotypes, which in turn leads to outgroup derogation (Spencer et al., 1998). Previous research indicates that chronically stigmatized groups (e.g., Blacks) do not privately hold the same negative stereotypes and biases as Whites (Alexander, Brewer, & Livingston, 2005; Cottrell et al., 2009; Monteith & Spicer, 2000; Niemann et al., 1994; Shelton & Richeson, 2006): With respect to the current studies, whereas Whites tend to have a private preference for White over Native American targets, Blacks tend to prefer Native American over White targets (Shapiro & Neuberg, 2008). Thus, we anticipated that Black participants receiving negative feedback from Black evaluators would rate high-status (White) candidates more negatively than low-status (Native American) candidates.

Method

Thirty-four Black (15 female) and 75 White (33 female) ASU students participated in exchange for course credit or ten dollars. The procedure was identical to Experiment 1 with the exception that the “partner” was Black. Hence, the overall experimental design was a 2[(Candidate Race: Native American (low-status)/White (high-status)] X 2(Negative Feedback: Yes/No) X 2(Participant Race: Black/White), fully between-participants with random assignment.

Results and Discussion

An ANOVA on competence evaluations (α=.96) revealed a main effect of the candidate’s race, F(1,103)=9.13, p=.003, η_p²=.08, qualified by the predicted three-way interaction, F(1,103)=6.80, p=.01, η_p²=.06 (Figure 2).

Figure 2.

Evaluations of job applicant competence as a function of job applicant race/ethnicity, participant race/ethnicity, and negative feedback from a Black evaluator (Study 2). Error bars show standard errors.

For White participants, there was a main effect of candidate race, F(1,103)=4.99, p=.03, η_p²=.05; the Native American candidate was evaluated more positively than the White candidate.¹ More important, and unlike Study 1, the feedback manipulation did not interact with candidate race (F=.02); as predicted, White participants evaluated the candidates similarly regardless of the feedback condition.

For Black participants, a main effect emerged for candidate race, F(1,103)=5.47, p=.02, η_p²=.05, qualified by the predicted race by feedback interaction, F(1,103)=8.42, p=.01, η_p²=.08: When Black participants received negative feedback (compared to no feedback) from the Black evaluator, evaluations of the White candidate significantly dropped [F(1,103)= 7.83, p=.01, η_p²=.07] whereas evaluations of the minority candidate did not differ [F(1,103)=1.39, p=.24]. This overall pattern was driven primarily by receiving negative feedback, which, contrary to Allport’s expectation, led Black participants to evaluate the White candidate more negatively than the minority candidate, F(1,103)=18.66, p<.001, η_p²=.15. In sum, Experiment 2 revealed that frequently stigmatized individuals, like those less frequently stigmatized, might derogate out-group members following negative feedback.

General Discussion

There is long-standing interest in how often-stigmatized individuals feel about and respond to members of other potentially stigmatizable groups (Major & Vick, 2005; Shelton, 2000), although very little research has addressed this question. In the present studies, we explored one situation theorized to elicit stigmatization of low-status groups by other low-status groups—contexts in which one receives negative feedback from members of high-status groups (Allport, 1954; Cadinu & Reggiori, 2002).

We anticipated and found that Black—unlike White—participants did not differentially evaluate White (high-status) and minority (low-status) job candidates after being criticized by a White evaluator. Yet, we did find evidence of outgroup derogation by Black participants when the negative feedback came from a Black evaluator. Research on the implications of self-image threats suggests that individuals are especially likely to derogate members of groups they, more generally, negatively stereotype (e.g., Spencer et al., 1998). Consistent with this research, we predicted and found that Black participants derogated the high-status White targets and not the low-status Native American targets. This finding is inconsistent with theorizing by Allport (1954). It is also inconsistent with recent findings by Cadinu and Reggiori (2002). We speculate these differences likely derive from the different groups studied—Cadinu and Reggiori explored views of psychologists towards ostensibly higher-status physicians and lower-status social workers—and the likely influences of variables such as group type (racial versus occupational), chronicity of group stigmatization (greater for American Blacks than professional psychologists), and strength and accessibility of the group’s negative attitudes (greater toward other races versus occupations).

A surface interpretation of the findings in the present research reveals diverging outcomes between Black and White participants. A deeper consideration, however, suggests that Whites and Blacks may possess the same underlying psychology of criticism-fed outgroup derogation—a psychology based on attributional inference and self-image maintenance. Shapiro and Neuberg (2008) provide another example in which apparently diverging race-based outcomes emerge from a similar underlying (in that case, norm-focused) process.

We note two limitations of this work. First, we have suggested a set of mechanisms that underlie our findings: Because Blacks can interpret criticism from Whites, and not Blacks, as motivated by prejudice, it can be discounted (and is therefore less personally threatening), thereby eliminating any utility of derogating others. Such mechanisms, however, await direct test. A second limitation results from the lack of research on racial minority group member’s attitudes toward, or stereotypes of, other racial minority groups (for exceptions see Cottrell et al., 2009; Niemann et al., 1994). Our reasoning about likely targets of Black outgroup derogation builds upon recent findings that Blacks view Native Americans more favorably than they view Whites (Shapiro & Neuberg, 2008), although these findings are limited to data gathered in the American Southwest. Future research exploring interethnic attitudes will increase the generalizability of the present findings and contribute to our understanding of intergroup relations.

Implications for Understanding and Reducing Prejudice

The current studies are among some of the first to examine prejudice expression in frequently stigmatized individuals. Recent research has been critical of the asymmetry that characterizes prejudice research (e.g., Hebl & Dovidio, 2005; Major & Vick, 2005; Shelton, 2000; Shelton & Richeson, 2006): Most research focuses on behaviors of the dominant group aimed at a minority group and has little to say about how minority racial prejudices manifest or the consequences of these prejudices.

The present research suggests that, to the extent that prejudice-reduction interventions are based on majority group behavior, they may prove to be of questionable utility for reducing the ubiquity of specific prejudices within the broader multi-group network that characterizes contemporary Western society. For example, interventions should continue to identify situations that can yield self-image threats, as prejudice expression will likely emerge in these contexts. However, the present findings suggest that different situations may differentially yield self-image threats for members of different races, and therefore yield divergent prejudice expression outcomes. Interventionists will need to carefully assess whether converging processes for minority and majority groups are likely to emerge and, if so, whether they are likely to manifest similarly across identical contexts. More generally, the present findings highlight that an understanding of intergroup relations will be problematically restrictive if researchers limit their study to the prejudices directed from those holding power to those holding less power.