Introduction

A dominant narrative in the United States (U.S.) is that American society has moved into a post-racial era, where the concept of race—and therefore racism—is obsolete. Consistent with this, many Americans strive to not “see” race and espouse racial attitudes that emphasize sameness and equal distribution of resources across racial lines (i.e., racial colorblind ideology). Despite the presumed benefits of endorsing racial colorblindness, there is growing evidence that racial colorblindness is associated with negative intergroup attitudes and behaviors (Neville et al. 2013). Given the invalidating nature of and negative consequences associated with racial colorblindness, it has thus been conceptualized as a form of “ultra-modern” racism (Neville et al. 2000). It is important to understand the correlates of racial colorblindness in order to advance theories about the development and maintenance of this ideology, with the long-term goal of informing interventions designed to reduce racial colorblindness and encourage the development of more effective racial ideologies. To this end, the goal of this paper is to address a gap in the literature by examining the personality correlates (i.e., Big 5, empathy, and aggression) of racial colorblindness.

Theory of Racial Colorblindness

Although many people who endorse racial colorblindness believe they are being anti-racist, there is evidence that racial colorblindness is associated with negative behaviors and attitudes toward racial and ethnic minorities. For example, racial colorblindness has been associated with less support for affirmative action policies (Awad et al. 2005), less support for confronting racism (Zou and Dickter 2013), and greater approval of racial insensitivity (e.g., racially themed parties; Tynes and Markoe 2010). Individuals who endorse racial colorblindness are also less aware of cultural diversity issues (Wang et al. 2014) and are less open to learning about other racial and ethnic groups (Neville et al. 2014). Interpersonally, there are data suggesting that individuals who avoid discussions of race tend to appear more hostile in interracial interactions (Apfelbaum et al. 2008), have less diverse social networks (Tawa et al. 2016), and may have lower empathy and concern toward African American clients in clinical (Burkard and Knox 2004) and community contexts (e.g., attitudes toward Hurricane Katrina victims; Warren 2013).

Moreover, endorsing racial colorblindness invalidates racial and ethnic minorities’ experiences of discrimination (Sue et al. 2008), suggesting that racial colorblindness is a form of modern prejudice itself. Along these lines, the construct of racial colorblindness is similar to what is known as modern racism (Neville et al. 2013) in that they both include the belief that racism is over. Importantly, however, the operationalization of modern racism includes explicit antipathy toward African Americans, such as the belief that African Americans’ success is unwarranted or that African Americans have gotten more than they deserve. Operationalizations of racial colorblindness, on the other hand, have not typically included explicit antipathy toward racial and ethnic minorities. Instead, the most widely used scale to assess racial colorblindness, the Colorblind Racial Attitudes Scale (CoBRAS), operationalizes the construct multidimensionally, with three components reflecting a lack of awareness about (or denial of) the existence of racial privilege (e.g., advantages associated with one’s skin color), implications of institutional discrimination (e.g., beliefs about affirmative action policies), and the pervasiveness of blatant racial discrimination (e.g., beliefs about how much of a problem racism is today; Neville et al. 2000). Subsequent revisions to this model of racial colorblind ideology expand on this, arguing that racial colorblindness dimensions might differ in terms of their functions (Neville et al. 2013). For example, based in part on work by Frankenberg (1993), Neville et al. (2013) argue that color evasion refers to the minimization of race (i.e., “I don’t see color”), whereas power evasion refers to the minimization of racism (i.e., “Racism doesn’t exist”). Consistent with the dual-process approach to understanding individual differences in prejudice [i.e., emphasizing different underlying mechanisms of prejudice, such as desire for a hierarchical social order (i.e., social dominance orientation) and traditionalism (i.e., right-wing authoritarianism)], it is possible that color evasion and power evasion may serve different psychological functions and are therefore espoused by different kinds of people (Sibley and Duckitt 2008). For example, a color evasion pathway (i.e., avoiding the acknowledgement of race) may be motivated by a genuine desire to not discriminate based on race, whereas power evasion (i.e., denying the existence of racism) may be more motivated to establish in-group dominance. Despite these advances in theory, no studies have looked at racial colorblindness components separately and examined whether there are unique correlates with relevant individual differences and personality dimensions.

Prejudice and Personality Dimensions

Early theories posited that examining personality—defined as habitual patterns of cognition, affect, and behavior (Zillig et al. 2002)—is an effective strategy for understanding prejudice (Allport 1954). The most common taxonomy of personality traits organizes traits into five dimensions, referred to as the Big 5 (i.e., neuroticism, conscientiousness, agreeableness, extraversion, and openness). Examining these traits is especially useful for studying prejudice because they are relatively stable over time (Roberts and DelVecchio 2000), which may explain observed stability in prejudicial attitudes (e.g., Aosved et al. 2009). In addition, personality has well-characterized developmental trajectories (Roberts and Mroczek 2008), which may help identify critical periods for intervention. Personality may also help elucidate the motivations (e.g., lack of openness) and maintenance mechanisms (e.g., lack of empathy) for endorsing a colorblind racial ideology.

There is a large literature on the personality correlates of prejudice, showing that on the one hand, general prejudice (i.e., related to gender, race, etc.) is related to lower levels of agreeableness and trait empathy (Ekehammar and Akrami 2007; Pettigrew and Tropp 2008; Sibley and Duckitt 2008), suggesting that people who tend to be prejudiced are generally disagreeable, uncaring, and antagonistic. On the other hand, prejudice is related to lower levels of openness to experience, suggesting that people who tend to be prejudiced are close-minded regarding learning about and interacting with people from different groups (Pettigrew and Tropp 2008). Conceptually similar to the agreeableness finding, research has found the individual differences in empathy are negatively associated with prejudice (Pettigrew and Tropp 2008). Moreover, empathy has been found to be an important mechanism for intergroup contact’s ability to reduce prejudice. Previous studies have generally not separated empathy into its two aspects (Decety and Jackson 2004), so it is unclear whether it is the cognitive component (i.e., the ability to take the perspective of others) or the emotional component (i.e., the ability to experience the feelings of others) that is driving these effects. Looking at these two separately is important because they have unique correlates (and are only moderately correlated with each other; Decety and Jackson 2004) and may elucidate different mechanisms in the maintenance of particular attitudes. Importantly, a downstream effect of low empathy is aggression (Vachon et al. 2014). Therefore, it is not surprising that prejudice is related to increased trait and laboratory aggression (e.g., Leonard and Taylor 1981; Nagoshi et al. 2008). Thus far, no research has examined whether racial colorblindness attitudes follow this pattern or whether this type of ultra-modern prejudice follows a unique pattern in relation to aggression.

In an attempt to integrate research from various fields, including counseling and personality psychology, the primary goal of the current study was to explore the associations between racial colorblindness and its three dimensions (unawareness of racial privilege, unawareness of institutional discrimination, and unawareness of blatant racism) and personality, empathy, and aggression. To do this, white undergraduates completed self-report scales of racial colorblindness, personality, empathy, and aggression in a laboratory study.

The current study adds to the literature in three ways. First, it helps place racial colorblindness in the larger literature of personality and prejudice, which can help advance theories of racial colorblindness by building connections to a larger body of work. Second, previous research looking at prejudice and empathy has not distinguished between empathic concern and perspective taking. Determining which factor plays a role may help inform development of interventions to reduce prejudice. Finally, previous studies have not parsed racial colorblindness into its component parts and therefore have not established unique correlates of the different aspects of racial colorblindness. This has the potential to inform theories of racial colorblindness, such as the validity of a color evasion and power evasion distinction.

Based on the current literature it is unclear whether the three dimensions of racial colorblindness are more similar to the disagreeable/lacking empathy, lack of openness pattern of associations, or both. Given that—unlike classic and modern forms of racism—racial colorblindness does not have explicit antipathy toward racial and ethnic minorities, it is possible that all three aspects of racial colorblindness would be related to lower openness to experience, as opposed to disagreeableness, lack of empathy, and aggression. However, because some aspects of racial colorblindness—namely, lack of understanding the implications of institutional discrimination and pervasiveness of blatant racial discrimination—involve some form of invalidation of the experiences of others and denial of systemic racism, it is possible that they may be associated with lower agreeableness and empathy. However, the other aspect of racial colorblindness focused on avoiding identification of race (i.e., unawareness of racial privilege) may be associated with agreeableness and empathy to a lesser degree. At the same time, it is possible that avoidance and unawareness of race and racial privilege may also involve some form of invalidation, and thus may be associated with empathy to a similar degree. By examining two separable components of empathy (i.e., empathic concern and perspective taking) that may be differentially associated with aspects of racial colorblindness, we hope to provide more nuanced answers to these exploratory questions.

Methods

Participants

Three hundred and nineteen white undergraduates (63% women; .31% transgender) attending a large, public university in the Midwest completed the study for course credit. An additional three participants were missing data for all questionnaires (due to computer malfunction) and were not included in the analyses. No other participants had missing data. The majority (95%) of the participants were between the ages of 18–20. According to a power analysis performed in G*Power (Faul et al. 2009), this sample size gave us a priori power of .80 to detect effect sizes common in social and personality research (r = .21; Richard et al. 2003) in a multiple regression with five predictors.

Measures

Racial Colorblindness

Colorblind racial ideology was measured by the Colorblind Racial Attitudes Scale (CoBRAS; Neville et al. 2000). The CoBRAS is composed of three subscales: unawareness of racial privilege (e.g., “Everyone who works hard, no matter what race they are, has an equal chance to become rich.”), unawareness of institutional discrimination (e.g., “English should be the only official language in the U.S.”), and unawareness of blatant racial issues (e.g., “Racism may have been a problem in the past, it is not an important problem today.”). In addition to calculating scores for each subscale, we also created a score across all items. Participants rate items on a 6-point Likert-type scale (1 = strongly disagree; 6 = strongly agree). Table 1 displays the descriptive statistics. Cronbach’s alphas for the three scales were similar to those of previous research (Neville et al. 2000). As covered in the introduction, a large body of work supports the construct validity of the CoBRAS (e.g., Awad et al. 2005; Neville et al. 2000). The means for the full scale and subscales were similar to other college samples (e.g., Neville et al. 2000).

Table 1 Means and SD for study variables
Full size table

Big 5 Personality Traits

We used the Big 5 scales from the International Personality Item Pool (Goldberg 1999) to measure extraversion (e.g., “I am the life of the party”), neuroticism (e.g., “I get upset easily”), agreeableness (e.g., “I make people feel at ease”), conscientiousness (e.g., “I like order”), and openness to experience (e.g., “I have a vivid imagination”). In their review of Big 5 assessments, John and Srivastava (1999) suggest that the International Personality Item Pool has good reliability and converges with other measures of the Big 5 in terms of validity (also see McAbee and Oswald 2013). Our alphas were similar to other studies using this scale (e.g., Bresin et al. 2012). Participants rated items on a 5-point scale (1 = very inaccurate; 5 = very accurate) as to how they generally characterize them. We created a score for each subscale. The means (see Table 1) of each scale were similar to other college samples (Goldberg 1992).

While the Big 5 theory has been well replicated across US samples, there is evidence that it is not an optimal tool for assessing personality because it is not cross-culturally valid (i.e., structure may differ across cultural groups; McAdams and Pals 2006). Though early researchers argued that personality should be impervious to the influence of social and cultural factors (McCrae and Costa 1999), several researchers have provided evidence that this may not be the case (e.g., McAdams and Pals 2006) and have attempted to use more diverse samples. Using data gathered from culturally and linguistically diverse samples, other researchers have argued for a six-factor structure of personality to better capture these individual differences (e.g., HEXACO model; Lee and Ashton 2004). This model is similar to the Big 5, with the major difference being agreeableness being split into two factors. While the IPIP should be used with caution in research outside of the USA, there is considerable evidence that it is appropriate to use in US samples like the one described here (McAbee and Oswald 2013).

Empathy

Individual differences in empathy were assessed by the Interpersonal Reactivity Index (IRI; Davis 1983). The IRI contains four subscales; however, we focused on the two most directly linked with empathy: perspective taking, which captures the cognitive aspects of empathy (e.g., “I sometimes try to understand my friends better by imagining how things look from their perspective.”), and empathic concern, which captures the affective aspects of empathy (e.g., “I often have tender, concerned feelings for people less fortunate than me.”). Each scale contains seven items, each of which is rated on a 5-point scale (1 = doesn’t describe me at all; 5 = describes me very well). The IRI has a literature supporting the reliability (with similar alphas to our sample) and validity of the scale. For example, the subscales are correlated with other self-report measures of empathy (Davis 1983), there is agreement between children’s self-report and parent report (Cliffordson 2001), and some subscales (e.g., perspective taking) predict future aggressive behavior (e.g., Lauterbach and Hosser 2007). We created a score for each subscale, and the means of the scales were similar to published norms (Davis 1983).

Aggression

To assess aggression, participants completed the short-form version of the Buss–Perry Aggression Questionnaire (Buss and Perry 1992) as developed by Diamond and Magaletta (2006). This measure consists of 12 items that participants rated on a Likert-type scale (1 = very unlike me; 5 = very like me). The BPAQ-SF is comprised of four subscales: Physical Aggression (e.g., “I have trouble controlling my temper”), Verbal Aggression (e.g., “I often find myself disagreeing with people”), Anger (e.g., “Sometimes I fly off the handle for no good reason”), and Hostility (e.g., “I wonder sometimes why I am so bitter about things”). The short form of the Aggression Questionnaire has a high correlation with the full scale and shows similar reliability and validity (Diamond and Magaletta 2006). We created a score for each subscale. Our means were comparable to the norms presented by Diamond and Magaletta (2006), aside from our sample reporting less physical aggression (1.81, SD = .88 versus 2.19).

Procedures

Participants first signed up for the study using the psychology department subject pool which listed several other available studies. The study was only visible to students who indicated in a prescreen survey that they identified as white/non-Hispanic, and participants were not given any information about the study prior to signing up. The informed consent process involved explaining to participants that they were participating in a study about attitudes and emotions and that they were welcome to withdraw from the study without any penalties if they felt uncomfortable. Participants completed the questionnaires on a computer following the completion of two computer tasks as part of a larger study (see Mekawi et al. 2016 for more details). Participants took approximately 30 min to complete the study, the majority of which was spent completing the questionnaires. The questionnaires were administered in the order they were presented in the measures section. At the end of the study, an experimenter debriefed with each participant by providing a more detailed explanation of the goals of the study and providing a comprehensive debriefing sheet. All procedures were approved by the Institutional Review Board at the University.

Results

Data Analysis

All variables were coded such that higher levels indicated higher levels of that variable. Our analysis proceeded in three steps. First, we examined the correlates of the three CoBRAS subscales and the overall scale by looking at the zero-order correlations (see Table 1). Second, we ran three multiple regressions for each CoBRAS subscale and the overall scale, where the CoBRAS or CoBRAS subscale was the criterion and the subscale (e.g., perspective taking and empathic concern) of individual difference measure (e.g., empathy) served as the predictor. This allowed us to adjust for the overlap among predictors within a scale. For all regression results, we report standardized betas (i.e., β) and 95% CI. We focus on confidence intervals as opposed to p values, in line with current recommendations for replicable science (e.g., Cumming 2014). We interpreted confidence intervals that did not include zero as statistically significant. Finally, we conducted some robustness tests. Given that men were significantly higher on overall colorblindness (d = .26, p = .029), unawareness of blatant racial issues (d = .37, p = .003), and unawareness of institutional discrimination (d = .34, p = .001), we performed analyses adjusting for gender. For the CoBRAS subscale analyses, we also performed analyses with the other two CoBRAS subscales as covariates. In line with the best practices, we thought it important to report results with and without the covariates (cf. Simmons et al. 2011). Given that there has not been research establishing the nomological network for colorblind racial ideology with gender or other CoBRAS subscales partialled out, we used these analyses to determine the robustness of our results but primarily interpret the non-partialled results (cf. Verona and Miller 2015).

Overall Colorblindness

At the zero-order level, the full racial colorblindness scale was significantly correlated to lower agreeableness, perspective taking, and empathic concern, and more physical aggression with small effect sizes according to Cohen (1992; see Table 1 for values). In the regression analyses (see the first column of Table 2), the full colorblindness scale was significantly associated with lower agreeableness (β = −.17, 95% CI [−.28, −.05]) and perspective taking (β = −.15, 95% CI [−.28, −.02]); however, the confidence intervals for empathic concern (β = −.05, 95% CI [−.18, .07]) and physical aggression (β = .10, 95% CI [−.01, .23]) contained zero. Together, these results suggest that colorblindness in general, as assessed by the CoBRAS, is characterized by low agreeableness and the cognitive aspects of empathy and perhaps to some extent the emotional aspects of empathy and physical aggression. These results offer a useful comparison for the individual scales.

Table 2 Correlations among the measures
Full size table

Unawareness of Racial Privilege

At the zero-order level, unawareness of racial privilege was related positively to empathic concern, and negatively to neuroticism, both at the marginal significance level (p < .10). When adjusting for overlap among the Big 5 (see the second column in Table 2), unawareness of racial privilege was significantly negatively related to openness to experience (β = −.12, 95% CI [−.57, −.01]), as was predicted. When adjusting for empathic abilities, unawareness of racial privilege was significantly negatively related to perspective taking (β = −.17, 95% CI [−.29, −.05]) and significantly positively related to empathic concern (β = .18, 95% CI [.06, .33]). These results suggest that the lack of awareness of racial privilege aspect of racial colorblindness fits with the lack of openness pattern more than the disagreeable pattern. Aside from perspective taking, unawareness of racial privileged had a unique set of correlations compared to the overall scale.

Unawareness of Institutional Discrimination

At the zero-order level, unawareness of institutional discrimination was significantly negatively related to agreeableness and perspective taking and significantly positively related to anger and hostility. In the regression results, unawareness of institutional discrimination was still significantly correlated to agreeableness (β = −.20, 95% CI [−.32, −.09]) and hostility (β = −.17, 95% CI [.04, .29]). The other correlations were reduced to below statistical significance. Thus, as predicted, unawareness of institutional discrimination was more consistent with the disagreeable/lack of empathy pattern, with a similar pattern of correlations as the full scale.

Unawareness of Blatant Racism

At the zero-order level, unawareness of blatant racism was significantly negatively related to agreeableness, perspective taking, and empathic concern, and significantly positively related to physical aggression, verbal aggression, and anger. When adjusting for overlap between the other Big 5 scales, unawareness of blatant racism was still significantly negatively related to agreeableness (β = −.24, 95% CI [−.83 −.30]). Unawareness of blatant racism was also still significantly negatively related to perspective taking (β = −.14, 95% CI [−.26, −.01]) and empathic concern (β = −.17, 95% CI [−.29, −.04]), when adjusting for overlap in empathic abilities. When adjusting for the other subscales of the Aggression Questionnaire, unawareness of blatant racism was only significantly related to physical aggression (β = .13, 95% CI [.01, .26]), with the other two associations being reduced to close to zero (see Table 3). Hence, the results for unawareness of blatant racism were similar to unawareness of institutional discrimination and the full scale.

Table 3 Standardized regression coefficients (β) and 95% CI for regression results
Full size table

Robustness

To ensure the robustness of our results, we performed two sets of follow-up analyses. Our goal was to see whether our results might be explained by important covariates, namely gender and the other CoBRAS. With gender included as a covariate, the results were largely the same. The full colorblindness scale was still significantly related to agreeableness (β = −.14, 95% CI [−.28, −.02]) and perspective taking (β = −.17, 95% CI [−.31, −.05]). Unawareness of racial privilege was still significantly related to perspective taking (β = −.17, 95% CI [−.30, −.05]) and empathic concern (β = .18, 95% CI [.04, .31]), but was marginally (p = .082) related to openness to experience (β = −.10, 95% CI [−.24, .02]), although the point estimate and confidence interval were similar as without the covariate. Unawareness of institutional discrimination was still significantly related to agreeableness (β = −.14, 95% CI [−.26, −.02]) and hostility (β = .20, 95% CI [.07, .32]). Unawareness of blatant racial issues was still significantly related to agreeableness (β = −.20, 95% CI [−.75, −.17]) and perspective taking (β = −.16, 95% CI [−.28, −.03]) but was no longer significantly related to empathic concern (β = −.13, 95% CI [−.26, .01]) or physical aggression (β = .08, 95% CI [−.05, .21]). Thus, although gender may explain some of our results, the bulk of associations were robust to gender as a covariate.

With the other CoBRAS subscales as covariates, the results were also generally robust. Unawareness of racial privilege was still negatively associated with openness to experience (β = −.11, 95% CI [−.21, −.01]) and perspective taking (β = −.11, 95% CI [−.23, −.01]), but positively related to empathic concern (β = .25, 95% CI [.13, .37]). Together with the above results, this suggests that unawareness of racial privilege has the most unique correlates of the three subscales. For unawareness of institutional discrimination, the positive association with hostility (β = .19, 95% CI [.08, .29]) was still significant, but the association with agreeableness was only marginally significant (β = −.10, 95% CI [−.21, .01]). This suggests that some of the relation between agreeableness and unawareness of institutional discrimination is accounted for by overlap among the scales. For unawareness of blatant racial issues, the associations with agreeableness (β = −.17, 95% CI [−.27, −.07]) and empathic concern (β = −.18, 95% CI [−.29, −.07]) were still significant, but those for perspective taking and physical aggression were not. This implies that unawareness of blatant racial issues has unique associations with agreeableness and empathic concern, but not perspective taking and physical aggression.

Discussion

The goal of this study was to establish the personality correlates of racial colorblindness, which has been conceptualized as a form of ultra-modern prejudice. Consistent with the previous literature on personality and prejudice (Sibley and Duckitt 2008), we found two divergent patterns of results. Unawareness of racial privilege was associated with lower openness to experience and perspective taking, whereas unawareness of institutional racism, unawareness of blatant racism, and the full racial colorblindness scale were related to lower agreeableness and empathic concern. Thus, different aspects of racial colorblindness were differentially associated with aspects of personality and empathy. These results help locate racial colorblindness in the larger literature on personality and prejudice. Moreover, they add to the literature by suggesting that different aspects of empathy may have unique associations with prejudice. Finally, they establish unique correlates to the different aspects of racial colorblindness that diverge in some ways from the full scale.

Implications

Our results suggest that at least in terms of personality, there may be two different types of racial colorblindness that are somewhat in line with the color evasion (i.e., avoidance of race, similar to unawareness of racial privilege) and power evasion (i.e., avoidance of systemic racism, similar to unawareness of institutional and blatant racism). The fact that unawareness of racial privilege was associated with lower openness to experience and perspective taking—but not significantly related to agreeableness—suggests that individuals who are unaware of the privilege afforded to them for being white are not necessarily disagreeable, uncaring people. In fact, they reported a greater tendency to feel other’s emotions. It appears that their unawareness of their privilege may have to do with deficits in the cognitive aspects of empathy, namely their inability to see things from perspectives other than their own, perhaps due to limited experiences. It is also possible that being unaware of racial privilege buffers the distress of knowing that others are suffering because of the status quo. In the context of the dual-process model, it is possible that this form of prejudice may be parallel to the right-wing authoritarianism pathway, such that people who value conformity and are deferential to authority figures are more motivated to avoid “seeing color,” including their own whiteness and associated privilege. For example, endorsing the unawareness of privilege item, “Everyone who works hard, no matter what race they are, has an equal chance to become rich,” might be motivated by a desire to maintain traditional beliefs (e.g., Protestant work ethic) and avoid feeling threatened. Moreover, endorsing the idea that race does not play a role in who gets sent to prison may be motivated by a desire to see authority figures and traditional institutions as fair and just. It is possible that people who espouse these views, given the positive association with empathic concern, may be amenable to interventions focusing on increasing experiences with diverse groups and providing education about oppression in an attempt to build perspective taking. Along these lines, there is evidence that interventions taking an educational route (e.g., diversity training) are particularly effective at reducing unawareness of privilege, relative to the other dimensions (Neville et al. 2000). Furthermore, Neville et al. (2014) found that diversity experiences in college were associated with decreases in total CoBRAS scores over four years. This study, however, did not report results for each subscale, so it is unclear whether diversity experiences made specific changes to unawareness of racial privilege and whether changes in perspective taking mediate this effect.

In contrast to unawareness of racial privilege, both unawareness of institutional discrimination and unawareness of blatant racism were related to lower levels of agreeableness and empathic concern. Additionally, unawareness of institutional discrimination was related to hostility and unawareness of blatant racism was related to physical aggression. The results for these two scales were the most similar to the full scale, suggesting that the full scale largely represents these scales and not unawareness of racial privilege. We interpret these results to suggest that the second type of racial colorblindness is more akin to a generally hostile, disagreeable, un-empathic interpersonal style. Individuals high in these aspects of racial colorblindness are likely hostile and aggressive toward most people, not just racial and ethnic minority group members. This is in line with previous research that has established a link between aggression and prejudice (e.g., Leonard and Taylor 1981). In addition, it is possible that, unawareness of the problems caused by institutional discrimination and blatant racism may make it psychologically easier to keep resources for one’s own group. Related to the dual-process model, this pathway may be akin to the social dominance pathway, such that individuals who support maintaining a hierarchy and power over others may be more likely to deny the existence of systemic racism. For example, endorsing the item, “Racial and ethnic minorities in the U.S. have certain advantages because of the color of their skin,” likely delegitimizes the accomplishments of people of color as a way to maintain the current racial order. If one believes that advancement of racial and ethnic minority is legitimate, it may threaten the status quo by questioning whites’ position in society. In comparison with interventions targeted at individuals high in unawareness of racial privilege, interventions targeted at individuals high in unawareness of institutional discrimination and/or unawareness of blatant racism would likely require more than just experiences with members from outgroups. These interventions might also require aspects that challenge hostile attributions and relaxation training common to treatments for chronic aggression (e.g., Deffenbacher 2011).

Limitations and Future Directions

This study has a few limitations worth noting. First, due to time constraints, we only collected basic demographic data and left out several demographic variables that may co-vary with racial colorblindness (e.g., political orientation and social economic status; Neville et al. 2000). It is possible that adjusting for some of these variables would reduce the size of the associations between colorblindness and personality. This should be explored in future research. Second, this study was cross-sectional, therefore our results do not speak to the possible causal order of personality and colorblind ideology. Further theorizing is necessary to clarify what role personality and other factors (e.g., political orientation) play in the development and maintenance of a colorblind ideology. Now that this study has established cross-sectional associations, future longitudinal studies will have a better idea of what aspects to measure to understand these relations and rule out third variables. Another limitation is that we only measured the Big 5 personality traits at the domain level (e.g., agreeableness). A growing body of research suggests that the more specific facet level (i.e., the factors that make up agreeableness such as trust, sympathy) has more predictive value (Paunonen and Ashton 2001). Still, our study helps identify which traits to focus on in future research. While the CoBRAS remains the most validated and use measure of racial colorblindness in the literature (Neville et al. 2013), it is no longer up to date with current theories of racial colorblindness (e.g., specific delineation between color evasion and power evasion). Nevertheless, our findings suggesting that the full scale is mostly indexing unawareness of institutional discrimination and blatant racism and that the subscales have unique correlates have implications for using the CoBRAS in the future. Specifically, it may be important for researchers who use the full CoBRAS to also include analyses by subscale to ensure that the results are the same across all subscales. In addition, the divergent pattern we found brings into question whether the unawareness of privilege subscale of the CoBRAS—which in retrospect has been theorized to tap into power evasion (Neville et al. 2013)—might actually be more reflective of color evasion. Further psychometric work is needed to more clearly test the two-factor model of racial colorblindness and delineate their unique correlates. Finally, all of our assessments were self-report. Future studies could use ability measures of empathy and/or behavioral measures of aggression. It would also be possible to reduce method variance by using informant reports of personality and/or racial colorblindness, though the multiple predictors may diminish the extent of this issue.

Even with these limitations, these novel findings expand our understanding of how racial colorblindness fits into a larger nomological network of personality traits, and present new avenues for future research.