Expecting racial outgroups to view “us” as biased: A social projection explanation of Whites’ bias meta-stereotypes

Abstract

Meta-stereotypes, the stereotypes believed to be held about one’s ingroup by an outgroup, represent barriers to positive intergroup contact. Little is known, however, about factors accounting for meta-stereotypes. Although previous researchers have speculated on conceptual overlap between social projection (perceiving one’s personal attitudes to be commonly held) and meta-stereotypes, these constructs are typically studied separately. We propose the notion that meta-stereotypes can be explained by social projection processes. We examined Whites’ “bias meta-stereotypes” (perceptions that Blacks consider Whites biased) across two studies. Participants projected personal biases onto both their ingroup (Whites) and outgroup (Blacks); in turn, both ingroup and outgroup bias perceptions uniquely predicted bias meta-stereotypes. Overall, the positive relation between personal bias perceptions and bias meta-stereotypes was fully mediated (i.e., explained) by heightened perceptions of ingroup (White) and outgroup (Black) bias. Overall, there is considerable value in integrating basic social projection within intergroup domains, particularly with regard to meta-stereotyping.

Keywords

bias ingroup favoritism meta-stereotypes racial prejudice social projection

Understanding the psychological barriers to positive intergroup contact is critical given that contact remains “our best hope” for promoting positive intergroup relations (see Wright, Brody, & Aron, 2005). These obstacles can take the form of negative perceptions of the outgroup (e.g., they are prejudiced) or beliefs about the outgroup’s perception of the ingroup (e.g., they view us as prejudiced). Holding negative stereotypes about the outgroup often leads to tense, unfriendly, or avoided contact opportunities (Dovidio, Gaertner, Kawakami, & Hodson, 2002). Meta-stereotypes, the typically negative stereotypes that one believes an outgroup holds about one’s own ingroup (Vorauer, Main, & O’Connell, 1998), are likewise harmful to intergroup contact, creating intergroup anxiety, negative feelings, and/or disinterest in the interaction (e.g., see Finchilescu, 2010; Gordijn, Finchilescu, Brix, Wijnants, & Koomen, 2008; Méndez, Gómez, & Tropp, 2007; Vorauer et al., 1998; Vorauer & Turpie, 2004). Although the negative intergroup consequences of meta-stereotypes are well established, less is known about the factors explaining or accounting for meta-stereotypes. Understanding the psychological underpinnings of meta-stereotypes will provide insight into strategies to reduce meta-stereotypes, and hence promote positive intergroup relations.

Meta-stereotypes About Intergroup Bias

Meta-stereotypes regarding outgroup members’ views of the ingroup are shared among ingroup members, and are typically negative in valence (see Frey & Tropp, 2006; Vorauer et al., 1998). Although meta-stereotypes are held by both majority and minority group members (Gordijn et al., 2008; Shelton, Richeson, & Vorauer, 2006), majority group members are particularly concerned about appearing biased (see Dovidio & Gaertner, 1998; Plant & Devine, 1998; Vorauer, in press), reporting meta-stereotypes reflecting characteristics such as “arrogant”, “racist”, “selfish” (Finchilescu, 2010), “intolerant”, and/or “prejudiced” (Méndez et al., 2007). For instance, White Canadians believe that Aboriginals consider Whites to be prejudiced (Vorauer et al., 1998), and White Americans believe that Black Americans view Whites negatively (Judd, Park, Yzerbyt, Gordijn, & Muller, 2005). Because these bias-relevant meta-stereotypes are likely to be detrimental to positive intergroup relations, we examine Whites’ “bias meta-stereotypes” (i.e., beliefs that Blacks consider Whites to be racially biased), with particular focus on explaining the psychological underpinnings of meta-stereotypes about racial prejudice (Study 1) or racial ingroup favoritism (Study 2). In doing so we explore the possibility that meta-stereotypes are explained in part by social projection processes.

Are Meta-stereotypes Rooted in Social Projection Processes?

Social projection is the tendency to view others (ingroup and/or outgroup members, see Robbins & Krueger, 2005) as sharing one’s thoughts, feelings, or attitudes (see Krueger 1998, 2000, for reviews). Whereas the social projection of mundane, non-socially sensitive attitudes is well established (see meta-analysis by Mullen & Hu, 1988), recent evidence also suggests the projection of intergroup attitudes (e.g., biases). For instance, when Australian majority group members hold more negative attitudes toward Australian minority groups, they perceive greater ingroup consensus for these negative attitudes (Pederson, Griffiths, & Watt, 2008; Watt & Larkin, 2010). Judd et al. (2005) similarly found that French and American students projected “global evaluations of groups” (i.e., like/dislike of outgroup X) onto their respective ingroups. Specifically, French students who negatively evaluated Americans perceived that other French students also rate Americans negatively, and vice versa.

Judd et al. (2005) also found that French and American students projected group liking/dislikingonto their outgroups. Whereas the ingroup projection of intergroup bias is relatively straightforward (e.g., “I am prejudiced against Blacks, most Whites are prejudiced against Blacks as well”), outgroup projection of bias is more complex. Judd et al. (2005) distinguished between outgroup projection at the specific category level and the intergroup target category level (see Table 1 for glossary terms). In specific category projection, positive (or negative) ingroup ratings correspond to more positive (or negative) perceptions of the outgroup’s ratings of one’s specific ingroup (their outgroup). For example, the more that a White person likes Whites (ingroup), the stronger the perception that their outgroup (Blacks) likes that same target, Whites. This represents projection of an evaluation of one’s specific category (e.g., ‘Whites’). In contrast, with intergroup target category projection, the more positive (or negative) the ingroup evaluation, the stronger the belief that the outgroup rates their own ingroup positively (or negatively). For instance, a positive ingroup evaluation by a White person would correspond to a stronger belief that Blacks evaluate their respective ingroup (Blacks) positively. This represents projecting an evaluation of an intergroup target category (e.g., an ‘ingroup’ or ‘outgroup’), not a specific target (e.g., ‘Whites’ or ‘Blacks’). Although in the examples noted above, the “starting point” for these types of outgroup projection are ingroup evaluations, outgroup evaluations and other types of evaluations, beliefs, or perceptions may be projected in these ways as well. Judd et al. (2005) examined the outgroup projection of both ingroup (e.g., liking of one’s ingroup) and outgroup evaluations (e.g., liking of one’s outgroup) as well as several types of perceptions about groups (e.g., members of X group are emotionally reserved). In an example pertaining to outgroup evaluations, Judd et al. (2005) found that French students evaluating Americans negatively also thought that American students rated French students negatively, representing projection at the intergroup target category level. In keeping with this example, we examine both the projection of outgroup evaluations (i.e., biases toward the outgroup), as well as the projection of perceptions of ingroup bias and outgroup bias (comparable to Judd et al.’s examination of perceptions about groups).

We introduce the previously untested proposition that the social projection of personal bias underlies bias meta-stereotypes (e.g., Whites’ belief that Blacks view Whites to be biased). That is, self-acknowledged personal biases (i.e., the extent to which one considers oneself to be relatively low or high in bias) may be projected onto both the ingroup and the outgroup, and these ingroup and outgroup bias perceptions may predict bias meta-stereotypes (see Figure 1).Theoretically, bias meta-stereotypes are expected to be systematically influenced by one’s subjective perceptions of one’s own bias levels, operating through perceptions of ingroup and outgroup bias. Although others have theoretically proposed that projection and meta-stereotyping are related processes (e.g., Ames, 2004a, 2004b; Frey & Tropp, 2006; Vorauer et al., 1998), the two domains are typically studied independently (for an exception see Judd et al., 2005). We posit that social projection and meta-stereotyping are not only related, but that meta-stereotypes are explained by social projection.

Figure 1.

Indirect-effect model representing the hypothesized relation between prejudice perceptions and meta-stereotypes.

Previous research has indirectly suggested that outgroup projection may influence meta-stereotypes. Krueger (1996) found that both Blacks and Whites believed that their ingroup was viewed more negatively by the outgroup than was actually the case. In Vorauer et al.’s (1998) terms, each group held a negative meta-stereotype. Krueger explained this finding through outgroup projection of ingroup favoritism, assuming that both groups favored their ingroup over the outgroup and that, as a result, assumed that their outgroups likewise favored their own ingroups over their outgroups. Thus, Krueger suggested that people expect their group to be viewed in terms of negative stereotypes by the outgroup because they themselves view the outgroup in terms of negative stereotypes (i.e., as biased, intergroup target category outgroup projection, see Judd et al., 2005). In addition to Krueger’s (1996) theorizing, we also suggest that people may additionally hold negative meta-stereotypes because they view their ingroup in terms of negative stereotypes (e.g., as biased, specific category outgroup projection, see Judd et al., 2005). In other words, the more one believes that a stereotype applies to one’s ingroup (e.g., Whites), the more one expects that an outgroup (e.g., Blacks) will believe that the stereotype applies the respondent’s ingroup (e.g., Whites) as well.

In the current investigation, we focus on Whites’ bias levels, perceptions of group-level bias, and meta-stereotypes. In the present context, Whites’ personal bias perceptions are expected to predict meta-stereotypes, but this relation is expected to be mediated through White and Black bias perceptions (see Figure 1). More specifically, Whites’ increased perceptions of their own personal bias are expected to inform their perceptions of greater White and Black bias, through ingroup projection and intergroup target category outgroup projection, respectively. Following from this, the more Whites view Whites to be biased, the stronger their bias meta-stereotype is expected to be (“I think Whites are biased, so Blacks presumably think Whites are biased as well”), representing outgroup projection at the level of specific category (Judd et al., 2005). Similarly, the more Whites view Blacks to be biased, the stronger their bias meta-stereotype is expected to be (“I think Blacks [my outgroup] are biased, so Blacks presumably think Whites [their outgroup] are biased”). This would reflect outgroup projection at the level of intergroup target category (see Judd et al., 2005) and is consistent with reasoning by Krueger (1996). After statistically accounting for White and Black bias perceptions, the proposed relation between self-bias perceptions and bias meta-stereotypes is expected to be significantly reduced. This pattern is expected regardless of the type of intergroup bias examined (i.e., prejudice or ingroup favoritism). Thus, bias meta-stereotypes were expected to be distally influenced by personal bias perceptions, and proximally explained by ingroup and outgroup social projection of bias.

Guided by the broader projection literature (see Robbins & Krueger, 2005), where relations between direct self-perceptions and perceptions about others are examined, we emphasize the projection of acknowledged, subjective personal bias perceptions (e.g., “I am biased against Blacks”), rather than the projection of one’s subtle biases as measured by prejudice scales (e.g., “Blacks are getting too pushy in their demands for equal rights”, as per the Modern Racism Scale [MRS], McConahay, Hardee, & Batts, 1981), as the starting point. The former reflects a subjective sense of one’s bias and thus is particularly relevant to social projection processes. In contrast, the latter reflects a racial attitude that may or may not be considered by a respondent as prejudicial (or reflecting aspects of the self). Our model proposes the former as a basis of meta-stereotyping (see Figure 1), such that acknowledged biases in the self are projected to the ingroup and outgroup, which foster bias meta-stereotypes. Whereas MRS is widely recognized as a subtle measure of racial biases (e.g., see Dovidio & Gaertner, 2000; McConahay et al., 1981), self-acknowledged bias perceptions are those that a person is aware of and recognizes in the self. The extent to which one perceives bias (or lack of bias) in others is likely to be influenced by the extent to which one perceives bias (or lack of bias) in oneself, rather than ones’ beliefs about Black-White relations (as tapped by the MRS). These latter beliefs may not even be construed by the individual as biased or unbiased (this is why they are considered measures of “modern prejudice”).

Previous research (Vorauer et al., 1998) has found that decreased actual racial bias (tapped by a measure comparable to the MRS, rather than subjective assessments of one’s own bias) is associated with more negative general meta-stereotypes (consisting of many different traits, not just bias). We adopt a slightly different approach in light of our focus on bias (rather than general) meta-stereotypes. Given our expectations that self-acknowledged personal bias will be projected, we expect increased self-acknowledged bias to be associated with more negative bias meta-stereotypes. In light of Vorauer et al.’s well known finding, however, we will also test whether the MRS correlates with bias meta-stereotypes, and whether individual differences in modern racism moderate our hypothesized model paths.

Table 1.

Glossary table.

Perceptions (from a White person’s perspective)
	Ingroup target	Outgroup target
Projection	Ingroup projection: Projecting one’s beliefs onto the ingroup, e.g.,“I favor Whites; other Whites favor Whites too”	Outgroup projection at the level of specific category: Projecting one’s beliefs about a specific group onto the outgroup,e.g., “I favor my ingroup (Whites); Blacks also favor my ingroup (Whites)”
		Outgroup projection at the level of intergroup target category: Projecting one’s beliefs about an intergroup category onto the outgroup,e.g., “I favor my ingroup (Whites); Blacks favor their ingroup (Blacks)”
Meta-stereotype		Meta-stereotype: Perceptions of stereotypes held by the outgroup about one’s ingroup, e.g., “Blacks think Whites are prejudiced”

Overview

In Study 1 we examine the extent to which the projection of prejudice explains “prejudice” meta-stereotypes; in Study 2 we examine the extent to which the projection of ingroup favoritism explains “ingroup favoritism” meta-stereotypes. This allows us to examine two distinct forms of bias: anti-outgroup bias (Study 1) and pro-ingroup bias (Study 2).

In each study, we first examine the extent to which Whites endorse the meta-stereotype that Blacks view Whites as biased. We then test whether White participants who recognize stronger personal biases in themselves perceive greater biases in Whites and Blacks. This would provide evidence of ingroup and outgroup social projection of bias, respectively. Next, as a preliminary examination of whether projection processes predict meta-stereotypes, we examine the relations between each bias perception (self, White, and Black) and the bias meta-stereotype.¹ We then test our proposed model (see Figure 1), where self-bias perceptions predict meta-stereotypes indirectly through White and Black bias perceptions. Finally, we test an alternative “social influence” or “introjection” model (see Krueger, 2002), whereby perceived group attitudes inform personal attitudes in explaining meta-stereotypes (see Figure 2). Past researchers have similarly pitted projection versus social influence models against one another in addressing other research questions (see Gatto & Dambrun, 2012).

Figure 2.

Indirect-effect model representing an alternative pattern of relations between prejudice perceptions and meta-stereotypes (an “introjection” account of meta-stereotyping).

Of course, bias exists on a continuum ranging from unbiased to strongly biased. Given our interest in understanding the nature of bias meta-stereotypes (e.g., Blacks view Whites as high in bias), we have emphasized bias at the negative end of this continuum. It is important to note however, that our model (see Figure 1) correspondingly dictates that perceptions of the self as unbiased will be projected onto the ingroup and outgroup and hence foster an unbiased meta-stereotype (i.e., Blacks view Whites as low in bias).

Study 1: Explaining Prejudice Meta-stereotypes

Method

Participants

Students at a Canadian university participated for course credit or $5. Three non-White participants were excluded from analyses, leaving a sample of 133 White participants (46 men, 87 women, M_age = 19.44, SD = 2.21).

Procedure and materials

The materials below were completed in groups of up to ten, as part of a larger project. Measures were completed in the order listed below but were interspersed with filler materials.

Modern Racism scale (MRS; McConahay et al.,1981)

Participants responded to seven items measuring subtle prejudice (e.g., “Discrimination against Blacks is no longer a problem in North America”) on a scale ranging from 0 (strongly disagree) to 4 (strongly agree), α = .77.

Prejudice perceptions

As a measure of self-acknowledged prejudice, participants reported the extent to which they consider themselves prejudiced against Blacks (0 = not at all, to 10 = very much).² To measure perceived ingroup (White) prejudice, participants indicated the percentage (0% to 100%) of Whites who are prejudiced against Blacks. To measure perceived outgroup (Black) prejudice participants indicated the percentage (0% to 100%) of Blacks who are prejudiced against Whites.

Prejudice meta-stereotype

Participants rated the extent to which the average Black person views Whites to be prejudiced against Blacks (0 = not at all to 10 = very much).³

Social dominance orientation (SDO; Pratto, Sidanius, Stallworth, & Malle, 1994)

SDO is characterized by the support of group inequalities and hierarchies. Inclusion of this measure can assist in ruling out the possibility that ideologies about group dominance alternatively explain our proposed pattern of relations.⁴ Participants responded to 16 items measuring SDO on scales ranging from 1 (strongly disagree) to 7 (strongly agree).

Results and Discussion

Prejudice meta-stereotype

Perceptions that Blacks view Whites as prejudiced (M = 6.41) were significantly greater than the scale midpoint (5), t (132) = 7.78, p < .001, demonstrating evidence of a shared prejudice meta-stereotype among Whites, of moderate-to-large magnitude, d = .68.

Projection of personal prejudice

Those acknowledging more personal prejudice perceived elevated anti-outgroup prejudice in Whites and Blacks (ps < .001, see Table 2). This finding provides evidence of ingroup projection and outgroup projection (intergroup target category level), respectively.

Table 2.

Descriptive statistics and correlations among variables (Studies 1 and 2).

	Study 1		Study 2		(1)	(2)	(3)	(4)	(5)
	M	SD	M	SD
(1) Self-acknowledged bias (0–7/ 0–10)	1.73	1.70	2.28	2.64	–	.34***	.33***	.33***	.33***
(2) Whites’ bias (0–95/ 10–99)	49.17	21.04	54.64	19.74	.42***	–	.60***	.49***	.01
(3) Blacks’ bias (2–100/ 10–100)	50.79	22.54	64.65	20.12	.32***	.74***	–	.55***	.16
(4) Meta-stereotype (1–10/ 1–10)	6.41	2.08	7.00	2.00	.27***	.67***	.68***	–	−.01
(5) MRS (0–2.29/ 0–2.86)	.86	.61	.82	.59	.49***	.20**	.25**	.09	–

Notes. MRS = modern Racism Scale. Study 1 (i.e., “prejudice”) correlations are displayed above the diagonal; Study 2 (i.e., “ingroup favoritism”) correlations are displayed below the diagonal. Study 1 N = 132–133; Study 2 N = 216. *p < .05; **p < .01; ***p < .001. Values in parentheses following variable names represent observed variable ranges in Study 1/Study 2, respectively.

Relations among variables

The relations among the variables in Study 1 are displayed in Table 2 (above the diagonal). Notably, prejudice perceptions were positively intercorrelated, with each prejudice perception positively related to the prejudice meta-stereotype. Interestingly, increased MRS was significantly related to greater self-acknowledged prejudice, but not to group prejudice perceptions or to prejudice meta-stereotypes. This finding supports our assertion that self-acknowledged bias is more relevant to predicting bias meta-stereotypes than is one’s racial attitude.

Explaining the prejudice meta-stereotype through projection processes

As illustrated in Table 2 (above the diagonal), self-acknowledged prejudice, perceived White prejudice, and perceived Black prejudice were each positively associated with the prejudice meta-stereotype. To explore the hypothesis that a prejudice meta-stereotype can be explained by ingroup and outgroup projection of intergroup evaluations, we tested a path model using AMOS 18.0. We tested whether self-acknowledged prejudice predicted the prejudice meta-stereotype through White and Black prejudice perceptions.⁵ A fully-saturated model was tested initially, with all possible paths accounted for (i.e., df = 0). Bootstrapping (N = 1000) was employed to estimate the significance of the indirect effect (Kline, 2011). As expected, self-acknowledged prejudice positively predicted perceptions of both White and Black group-level prejudice (see Figure 3). In turn, perceptions of both White and Black prejudice uniquely positively predicted the prejudice meta-stereotype. Importantly, the direct path from self-acknowledged prejudice to the prejudice meta-stereotype was not significant in the model (β = .13, p = .09), despite a significant zero-order correlation outside of the model (r = .33, p < .001). Self-acknowledged prejudice exerted a significant indirect effect (IE = .20, p = .002, 61% of the total effect) on the prejudice meta-stereotype. Self-acknowledged prejudice therefore impacted the prejudice meta-stereotype through both ingroup and outgroup prejudice perceptions simultaneously. This pattern represents full mediation, such that self-acknowledged prejudice would not predict the prejudice meta-stereotype if not for group-level prejudice perceptions. In a follow-up analysis, participants were categorized as low or high in MRS based on a median split, and a multiple-groups analysis was conducted to determine if the model paths were moderated by MRS. No paths in the saturated model were moderated by MRS, meaning that the model held equally for those scoring low or high in racial bias.

Figure 3.

Results of path model testing the hypothesized relation between prejudice perceptions and meta-stereotypes (Study 1).

We then tested the fit of our proposed model (as per Figure 1). Recognized fit criteria include non-significant chi-squared values, chi-squared/df values < 2, comparative fit index [CFI] values > .95, root-mean-square-error of approximation [RMSEA] values < .06, and standardized root-mean-squared residual [SRMR] values < .08 (Kline, 2011). The model demonstrated very good fit, χ²(1) = 3.73, p = .285, χ²/df = 1.44, CFI = .995, RMSEA = .045, and SRMR = .037. These results are consistent with predictions: self-ratings of prejudice impacted the meta-stereotype about ingroup prejudice indirectly through both ingroup and outgroup prejudice perceptions. This suggests therefore, that ingroup and outgroup projection processes explain and account for prejudice meta-stereotypes. To rule out the competing possibility that social dominance orientation might explain these findings, we re-ran the model statistically controlling for SDO; our model continued to be supported, ruling out this potential confound.⁶

We also tested an alternative model presented in Figure 2 (i.e., reversing the positions of self prejudice and group prejudice perceptions in Figure 3). A fully-saturated model was tested (i.e., df = 0), and bootstrapping (N = 1000) was employed to estimate the significance of the indirect effects (Kline, 2011). Both White and Black prejudice perceptions predicted self-acknowledged prejudice, as well as the prejudice meta-stereotype (bs .20–.41=, ps < .05). However, the path from self-acknowledged prejudice to the prejudice meta-stereotype was not significant(β = .13, p = .093), nor were the indirect paths from White or Black prejudice perceptions to the prejudice meta-stereotype (IEs = .03, ps > .154). Thus, the relation between group prejudice perceptions and the prejudice meta-stereotype was not mediated by self prejudice perceptions. In terms of explaining meta-stereotypes, the self-to-group projection model was better supported than the group-to-self ‘introjection’ model.

Study 2: Explaining Ingroup Favoritism Meta-stereotype

Study 2 departs from Study 1 by examining whether the projection of ingroup favoritism influences ingroup favoritism meta-stereotypes. Ingroup favoritism represents preference for one’s own group over others (Brewer & Brown, 1998) rather than anti-outgroup prejudice, and thus represents a related but conceptually distinct construct. With ingroup favoritism the most common form of intergroup bias (Brewer, 1999), it is critical to examine whether meta-stereotypes are explained by social projection in this domain also. Study 2 improves upon Study 1 by including multiple-item measures of group prejudice perceptions and meta-stereotypes to increase reliability of these measures.

Method

Participants

Students at a Canadian university participated for course credit or $5. Twenty-six non-White participants were excluded from analyses, leaving a final sample of 216 White participants (60 men, 156 women, M_age = 20.08, SD = 4.45).

Procedure and materials

The materials below were completed individually (up to five people per session). As in Study 1, measures were completed in the order listed below, but were interspersed by filler items.

Modern Racism Scale

The MRS was completed as in Study 1, α = .80.

Ingroup favoritism perceptions

As a measure of self-acknowledged ingroup favoritism, participants indicated the extent to which they personally favor Whites over Blacks (0 = not at all, to 10 = very much). To measure perceived ingroup (White) ingroup favoritism, participants indicated both the percentage (0% to 100%) of Whites who favor Whites over Blacks, as well as the extent to which Whites favor Whites over Blacks (0 = not at all, to 10 = very much). After converting each ingroup-favoring perception to the same 0–100 scale, these measures were averaged (White ingroup favoritism α = .79). Similarly, to measure perceived outgroup (Black) ingroup favoritism participants indicated the percentage (0% to 100%) of Blacks who favor Blacks over Whites, as well as the extent to which Blacks favor Blacks over Whites (0 = not at all, to 10 = very much). After converting each perception to the same 0–100 scale, these measures were averaged (Black ingroup favoritism α = .83).

Ingroup favoritism meta-stereotype

Participants rated both the degree (0 = not at all to 10 = very much) and percentage (0% to 100%) of Blacks viewing Whites to be ingroup-favoring; these items were averaged after converting each to the same scale (α = .88).

Social dominance orientation (SDO; Pratto et al.,1994).

SDO was measured as in Study 1 for use as a covariate (i.e., to rule out an alternative explanation of our proposed model).

Results and Discussion

Ingroup favoritism meta-stereotype

Consistent with Study 1, Blacks were perceived to view Whites as ingroup-favoring, with the meta-stereotype mean (M = 7.00) significantly exceeding the scale midpoint (5), t (215) = 14.68, p < .001, representing a strong effect, d = 1.00.

Projection of personal ingroup favoritism

Those acknowledging more personal ingroup favoritism perceived elevated ingroup favoritism among Whites and Blacks (ps < .001, see Table 2). As in Study 1, this provides evidence of ingroup projection and outgroup projection (intergroup target category level), respectively.

Relations among variables

Again, prejudice perceptions were positively intercorrelated (see Table 2, below the diagonal), with each prejudice perception positively related to the ingroup favoritism meta-stereotype. MRS was significantly related to self-acknowledged ingroup favoritism and White and Black ingroup favoritism perceptions, but, consistent with Study 1, not to the prejudice meta-stereotype. As in Study 1, therefore, bias meta-stereotypes were predicted by personally recognized bias, but not by individual differences in MRS (relevant to whether Blacks are victims of discrimination). This finding is in keeping with our theoretical account that meta-stereotypes are rooted in the projection of a direct self-perception (“I am biased”).

Explaining the ingroup-favoring meta-stereotype through projection processes

As illustrated in Table 2 (below the diagonal), self-acknowledged ingroup favoritism, as well as perceived White and Black ingroup favoritism, were each positively associated with the ingroup favoritism meta-stereotype. A path model tested whether self-acknowledged ingroup favoritism predicted the ingroup favoritism meta-stereotype through White and Black ingroup favoritism perceptions, to determine whether this meta-stereotype can be explained through projection processes. A fully saturated model (df = 0) was tested in AMOS 18.0, with bootstrapping procedures (N = 1000) to estimate the indirect effect. As predicted, self-acknowledged ingroup favoritism positively predicted both White and Black ingroup favoritism perceptions, with each in turn positively predicting the ingroup favoritism meta-stereotype (see Figure 4). The direct path from self-acknowledged ingroup favoritism to meta-stereotype was not significant in the model (β = −.02, p = .685), despite being significant at the zero-order level (r = .27, p < .001). Indeed, a significant indirect effect (IE = .29, p = .002) explained the entire effect of self-acknowledged ingroup favoritism on the ingroup favoritism meta-stereotype, representing full mediation. Therefore, acknowledging the self as ingroup-favoring would not predict a corresponding meta-stereotype if not for perceptions of group biases (Whites and Blacks) derived from this self-bias. As in Study 1, a multiple-groups analysis revealed that no paths in the fully saturated model were moderated by MRS. That is, the model structure was consistent across individual differences in MRS.⁷ Additionally, as in Study 1, the model was supported after statistically controlling for SDO.⁸

Figure 4.

Results of path model testing the hypothesized relation between prejudice perceptions and meta-stereotypes (Study 2).

The hypothesized indirect-effect model was tested next (following Figure 1). The model demonstrated strong fit, χ²(1) = 1.49, p = .684, χ²/df = .498, CFI = 1.00, RMSEA = .000, and SRMR = .013 (see Figure 4). These results conceptually replicate Study 1 but with regard to ingroup favoritism as opposed to outgroup prejudice. Consistent across both studies, self-bias perceptions predicted meta-stereotypes indirectly through group-level bias perceptions (see Figure 4). Again, the alternative model tested in Study 1 was not well supported: a fully-saturated alternative model was tested (i.e., df = 0), and bootstrapping (N = 1000) was employed. White ingroup favoritism perceptions predicted self-acknowledged ingroup favoritism (β = .41, p < .001), but Black ingroup favoritism perceptions did not predict self-acknowledged ingroup favoritism (β = .02, p = .866).White and Black ingroup favoritism perceptions predicted the ingroup favoritism meta-stereotype (bs = .39–.40, ps < .001). As in Study 1, the path from self-acknowledged ingroup favoritism to the ingroup favoritism meta-stereotype was not significant (b = −.02, p = .685), nor were the indirect paths from White or Black prejudice perceptions to the prejudice meta-stereotype (IEs = .00, ps > .732). Thus, the relation between group prejudice perceptions and the prejudice meta-stereotype was not mediated by self prejudice perceptions. We find more support for our proposed projection account than for the alternative introjection account. Overall, Study 2 tapped a different type of intergroup bias and employed more reliable multiple-item measures of group prejudice perceptions and meta-stereotypes, producing results consistent with Study 1.

General Discussion

We demonstrate that Whites project their personal biases onto both their ingroup (Whites) and outgroup (Blacks), simultaneously. These specific findings are novel but are in keeping with previous meta-stereotype (e.g., Vorauer et al., 1998) and projection (e.g., Judd et al., 2005; Krueger, 1996) research, respectively. Our particular contribution to this literature is to establish meaningful relations between these phenomena, such that self-acknowledged personal biases predict bias meta-stereotypes through ingroup and outgroup bias perceptions. In other words, we establish that social projection processes can explain meta-stereotypes. This represents the first empirical examination of the association between these related constructs.

Studies 1 and 2 examined related but distinct different forms of bias (anti-outgroup vs. pro-ingroup bias) and documented virtually identical findings. Regardless of how bias is conceptualized, personal levels of bias predict bias meta-stereotypes almost entirely through ingroup and outgroup bias perceptions. It is worth noting, however, that the mediated effect was stronger in Study 2. That is, the indirect effect of self-bias perceptions on the bias meta-stereotype (through ingroup and outgroup bias perceptions) represented 61% of the total effect in Study 1, but 100% of the total effect in Study 2. Additionally, 35% of the variance in the prejudice meta-stereotype was accounted for in Study 1, whereas 52% of the variance in the ingroup-favoritism meta-stereotype was accounted for in Study 2. These differences may reflect differences between the examination of pro-ingroup relative or anti-outgroup bias, or the more reliable bias perception measures used in Study 2. To further examine this issue, future research can examine both outgroup prejudice and ingroup favoritism and their corresponding meta-stereotypes simultaneously. Regardless, the overall pattern replicates across the studies.

It is important to note that our findings reflect relative rather than absolute levels of bias, as is common in the intergroup bias literature. Our self-acknowledged bias measure tapped blatant, self-acknowledged attitudes, the type of self-attitudes typically assessed in general social projection studies (see Robbins & Krueger, 2005). Few participants endorsed extremely high levels of self-acknowledged bias (see Table 2), which is not surprising given prevailing egalitarian social norms (see Dovidio & Gaertner, 1998). Importantly, our primary interest did not concern examining mean levels of bias, but rather associations among relative levels of self-acknowledged bias, ingroup and outgroup bias, and bias meta-stereotypes. From this examination, we uncovered associations between self-acknowledged bias and group bias perceptions and bias meta-stereotypes. This suggests that relatively high levels of self-acknowledged bias are associated with perceptions of relatively high levels of ingroup bias, outgroup bias, and bias meta-stereotypes. By the same token, relatively low levels of self-acknowledged bias are associated with perceptions of relatively low levels of ingroup bias, outgroup bias, and bias meta-stereotypes. Thus, self-acknowledged bias was indeed projected, with participants perceiving their ingroup and outgroup to hold bias levels similar to their own, whether low, mid-range, or high.

Interestingly, in Studies 1 and 2, personal prejudice perceptions were projected onto both the ingroup and the outgroup, and both ingroup and outgroup bias perceptions predicted bias meta-stereotypes. Clearly, both types of perceptions uniquely influence inform bias meta-stereotypes. This represents an important consequence of outgroup projection: at both the specific category and intergroup target category level, outgroup projection plays a key role in the formation of bias meta-stereotypes. Therefore, bias meta-stereotypes are not only informed by perceptions of the ingroup as biased (specific category outgroup projection), but also by perceptions of the outgroup as biased (intergroup target category outgroup projection), simultaneously.

In some ways our findings appear to depart from past meta-stereotype findings. Vorauer et al. (1998) found that those low in prejudice held more negative meta-stereotypes than those high in prejudice. In both of our studies, higher self-acknowledged bias (i.e., subjective perceptions of greater bias) was associated with more negative bias meta-stereotypes, but MRS (i.e., actual bias level, conceptually similar to the prejudice variable tapped by Vorauer et al.) was unassociated with bias meta-stereotypes (see Table 2). Importantly however, our focus was specifically and narrowly focused on bias meta-stereotypes per se. In contrast, Vorauer and colleagues examined the overall meta-stereotype Whites hold regarding Aboriginal’s views of Whites (assessing which, among a large number of traits such as arrogant and disrespectful represented Whites’ meta-stereotype about Aboriginals views of Whites, with one of the traits being “prejudiced”). Thus, Vorauer et al.’s research does not address whether the meta-stereotype examined by the current research, the prejudiced meta-stereotype, is correlated with individual differences in prejudice. Rather, Vorauer et al. find a positive association between lower prejudice and an overall general meta-stereotype consisting of many traits (including prejudice). Our research is the first to examine the specific association between prejudice (both self-acknowledged bias, and that measured by the MRS) and bias meta-stereotypes. Our findings are consistent with our proposition that negative meta-stereotyping is rooted in self-perceptions of personal bias, regardless of whether one is low or high in prejudice relative to others in the sample. In fact, individual differences in MRS did not moderate model paths in either study, meaning that our projection-based account of meta-stereotyping holds among those with low or high racial biases. Overall these findings place considerable importance on the acknowledgement of racial bias (“I am racially biased”) in explaining meta-stereotypes, relative to modern biases as tapped by the MRS (“Blacks are getting too demanding; discrimination is no longer a problem”). This finding highlights a unique and previously unrecognized emphasis on intrapersonal, self-relevant psychological processes in meta-stereotyping.

Our findings also appear to depart from those of Lammers, Gordijn, and Otten (2008), who find that powerful groups are less likely to activate and apply meta-stereotypes than powerless groups. In our context, where Whites may be considered a more powerful group than Blacks, we find that White participants hold the meta-stereotype that Blacks view Whites as biased. Our finding that Whites hold this biased meta-stereotype is consistent with other work finding that powerful, advantaged groups hold meta-stereotypes (e.g., Finchilescu, 2010; Vorauer et al., 1998). Even with regard to activation, Lammers et al. found baseline meta-stereotype activation across all participants (both powerless and powerful). Without having Black participants, our data do not allow for the comparison of Whites’ and Blacks’ meta-stereotypes, but we recognize the possibility that Blacks’ meta-stereotypes may be stronger than Whites. However, given that bias meta-stereotypes are likely to be of primary concern to powerful rather than powerless group members (i.e., those concerned about appearing biased, Dovidio & Gaertner, 1998; Plant & Devine, 1998; Vorauer, in press), this possibility is unlikely.

To be clear, we do not believe that all meta-stereotypes are derived from projection processes, particularly given the past literature on meta-stereotypes stressing intergroup dynamics (e.g., Gordijn et al., 2008; Vorauer et al., 1998). Indeed, our own results did not account for all of the variance in bias meta-stereotypes. Nonetheless, our findings establish that projection processes have a significant role in forming bias meta-stereotypes. Future research can examine why racial bias is projected in this manner. It is possible that these projection processes serve as a means to release or justify socially undesirable attitudes (see Crandall & Eshleman, 2003). First, projecting one’s own bias onto the ingroup and outgroup may allow one to feel justified in holding a bias (i.e., “Others are biased, so it is okay for me to be biased”). Second, a negative bias meta-stereotype, held as a result of projecting personal and group biases, may further justify viewing the outgroup negatively (i.e., “They view us negatively, so it is acceptable to dislike them”). Future research can explore the viability of these possibilities.

Future research might also examine whether it is possible to tap direct self-perceptions of bias with multiple-item measures. The broader social projection literature has widely employed single-item measures (e.g., see Ames, 2004b; Mullen et al., 1985; Mullen & Hu, 1988), and we sought to stay within this paradigm, but this issue remains an open question. Our single-item measure, however, was high in face, construct, and criterion validity, given that it directly tapped the perceptions of interest, and was associated (repeatedly) with other model-relevant variables as expected (see Table 2).Future research can examine this question using multiple-item measures of personal bias perception.

We focused on bias meta-stereotypes in light of their detrimental effects on intergroup contact (see Finchilescu, 2010; Méndez, Gómez, & Tropp, 2007; Vorauer et al., 1998; Vorauer & Turpie, 2004). Given that intergroup contact is often avoided, the real challenge for researchers is to understand the factors underpinning resistance to contact (Hodson, 2011). The present investigation sheds some light on the effects of personal biases and their social projection on a key factor deterring contact (i.e., meta-stereotypes), with contact being one of our clearest means to reduce prejudice. Future research can consider whether other types of meta-stereotypes held by majority group members are similarly explained through projection processes (e.g., the meta-stereotype that Whites are arrogant, see Vorauer et al., 1998) and whether bias meta-stereotypes in non-racial contexts (e.g., concerning sexual orientation) are similarly characterized. The current investigation also provides direction for interventions that can focus on the reduction of bias meta-stereotypes. Attenuating bias meta-stereotypes may promote more positive intergroup contact, which itself is likely to improve attitudes. We demonstrate that perceptions of low personal bias predict perceptions of low ingroup and outgroup bias, and in turn, less negative bias meta-stereotypes. Thus, keeping bias and bias perceptions low is critical for fostering positive contact intentions and intergroup relations. Future work can examine strategies to reduce personal bias in those who perceive themselves to be especially high, with the end goal of reducing bias meta-stereotypes of the nature detrimental to intergroup contact.

Footnotes

Notes

References

Ames

D. R.

(2004a). Inside the mind reader’s tool kit: Projection and stereotyping in mental state inference. Journal of Personality and Social Psychology, 87, 340–353. doi: 10.1037/0022-3514.87.3.340

Ames

D. R.

(2004b). Strategies for social inference: A similarity contingency model of projection and stereotyping in attribute prevalence estimates. Journal of Personality & Social Psychology, 87, 573–585. doi: 10.1037/0022-3514.87.5.573

Brewer

M. B.

(1999). The psychology of prejudice: Ingroup love or outgroup hate? Journal of Social Issues, 55, 429–444. doi: 10.1111/0022-4537.00126

Brewer

M. B.

Brown

R. J.

(1998). Intergroup relations. The handbook of social psychology, 2, 554–594.

Byrne

B. M.

(2001). Structural equation modeling with AMOS: Basic concepts, applications, and programming. Mahwah, NJ: Lawrence Erlbaum Associates.

Crandall

C.S.

Eshelman

(2003). A justification-suppression model of the expression and experience of prejudice. Psychological Bulletin, 129, 414–446. doi: 10.1037/0033-2909.129.3.414

Dovidio

J. F.

Gaertner

S. L.

(1998). On the nature of contemporary prejudice: The causes, consequences, and challenges of aversive racism. In Eberhardt

J. L.

Fiske

S.T.

(Eds.), Confronting racism: The problem and the response (pp. 3–32). Thousand Oaks, CA: Sage.

Dovidio

J. F.

Gaertner

S. L.

(2000). Aversive racism and selection decisions: 1989 and 1999. Psychological Science, 11, 315–319. doi: 10.1111/1467-9280.00262

Dovidio

J.F.

Gaertner

S.E.

Kawakami

Hodson

(2002). Why can’t we just get along? Interpersonal biases and interracial distrust. Cultural Diversity and Ethnic Minority Psychology, 8, 88–102. doi: 10.1037/1099-9809.8.2.88

10.

Finchilescu

(2010). Intergroup anxiety in interracial interaction: The role of prejudice and metastereotypes. Journal of Social Issues, 66, 334–351. doi: 10.1111/j.1540-4560.2010.01648.x

11.

Frey

F. E.

Tropp

L. R.

(2006). Being seen as individuals versus as group members: Extending research on metaperception to intergroup contexts. Personality and Social Psychology Review, 10, 265–280. doi: 10.1207/s15327957pspr1003_5

12.

Gatto

Dambrun

(2012). Authoritarianism, social dominance, and prejudice among junior police officers. Social Psychology, 43, 61–66. doi: 10.1027/1864-9335/a000081

13.

Gordijn

Finchilescu

Brix

Wijnants

Koomen

(2008). The influence of prejudice and stereotypes on anticipated affect: Feelings about a potentially negative interaction with another ethnic group. South African Journal of Psychology, 38, 589–601.

14.

Hodson

(2011). Do ideologically intolerant people benefit from intergroup contact? Current Directions in Psychological Science, 20, 154–159. doi: 10.1177/0963721411409025

15.

Judd

C. M.

Park

Yzerbyt

Gordijn

E.H.

Muller

(2005). Attributions of intergroup bias and outgroup homogeneity to ingroup and outgroup others. European Journal of Social Psychology, 35, 677–704. doi: 10.1002/ejsp.281

16.

Kline

R. B.

(2011). Principles and practice of structural equation modeling (3rd ed.). New York, NY: Guilford Press.

17.

Krueger

(1996). Personal beliefs and cultural stereotypes about racial characteristics. Journal of Personality and Social Psychology, 71, 536–548. doi: 10.1037/0022-3514.71.3.536

18.

Krueger

(1998). On the perception of social consensus. In Zanna

M. P.

(Ed.), Advances in experimental social psychology (Vol. 30, pp. 163–240). San Diego, CA: Academic Press.

19.

Krueger

(2000). The projective perception of the social world. In Suls

Wheeler

(Eds.), Handbook of social comparison: Theory and research (pp. 323–351). New York, NY: Kluwer Academic/Plenum.

20.

Krueger

J. I.

(2002). On the reduction of self-other asymmetries: Benefits, pitfalls, and other correlates of social projection. Psychologica Belgica, 42, 23–41.

21.

Lammers

Gordijn

Otten

(2008). Looking through the eyes of the powerful. Journal of Experimental Social Psychology, 44, 1229–1238. doi: 10.1016/j.jesp.2008.03.015

22.

McConahay

J. G.

Hardee

B. B.

Batts

(1981). Has racism declined? It depends on who is asking and what is asked. Journal of Conflict Resolution, 25, 563–579.

23.

Martin

C. L.

(1987). A ratio measure of sex stereotyping. Journal of Personality and Social Psychology, 52, 489–499.

24.

Méndez

Gómez

Tropp

L. R.

(2007). When metaperceptions are affected by intergroup processes. International Journal of Psychology and Psychological Therapy, 7, 237–250.

25.

Mullen

Atkins

J. L.

Champion

D. S.

Edwards

Hardy

Story

J. E.

Vanderklok

(1985). The false consensus effect: A meta-analysis of 115 hypothesis tests. Journal of Experimental Social Psychology, 21, 262–283.

26.

Mullen

(1988). Social projection as a function of cognitive mechanisms: Two meta-analytic integrations. British Journal of Social Psychology, 27, 333–356. doi: 10.1111/j.2044-8309.1988.tb00836.x

27.

Pedersen

Griffiths

Watt

S. E.

(2008). Attitudes toward out-groups and the perception of consensus: All feet do not wear one shoe. Journal of Community and Applied Social Psychology, 18, 543–557. doi: 10.1002/casp.964

28.

Plant

E. A.

Devine

P. G.

(1998). Internal and external motivation to respond without prejudice. Journal of Personality and Social Psychology, 75, 811–832. doi: 10.1037/0022-3514.75.3.811

29.

Pratto

Sidanius

Stallworth

L. M.

Malle

B. F.

(1994). Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67, 741–763. doi: 1995-09457-00110.1037/0022-3514.67.6.998

30.

Robbins

J. M.

Krueger

(2005). Social projection to ingroups and outgroups. Personality and Social Psychology Review, 9, 32–47. doi: 10.1207/s15327957pspr0901_3

31.

Shelton

J. N.

Richeson

J.A.

Vorauer

J. D.

(2006). Threatened identities and interethnic interactions. European Review of Social Psychology, 17, 321–358. doi: 10.1080/10463280601095240

32.

Vorauer

(in press). Getting past the self: Understanding and removing evaluative concerns as an obstacle to positive intergroup contact effects. In Hodson

Hewstone

(Eds.), Advances in intergroup contact. London: Psychology Press.

33.

Vorauer

J. D.

Main

K. J.

O’Connell

G. B.

(1998). How do individuals expect to be viewed by members of lower status groups? Content and implications of meta-stereotypes. Journal of Personality and Social Psychology, 75, 917–937. doi: 10.1037/0022-3514.75.4.917

34.

Vorauer

J. D.

Turpie

(2004). Disruptive effects of vigilance on dominant group members’ treatment of outgroup members: Choking versus shining under pressure. Journal of Personality and Social Psychology, 87, 384–399. doi: 10.1037/0022-3514.87.3.384

35.

Watt

S. E.

Larkin

(2010). Prejudiced people perceive more community support for their views: The role of own, media, and peer attitudes in perceived consensus. Journal of pplied Social Psychology, 40, 710–731. doi: 10.1111/j.1559-1816.2010.00594.x

36.

Wright

S. C.

Brody

S. A.

Aron

(2005). Intergroup contact: Still our best hope for reducing prejudice. In Crandall

C. S.

Schaller

(Eds.), The social psychology of prejudice: Historical perspectives (pp. 115–142). Seattle, WA: Lewinian Press.