1 Colorblindness NORTON ET AL. and Diversity Social Cognition, Vol. 26, No. 1, 2008, pp COLORBLINDNESS AND DIVERSITY: CONFLICTING GOALS IN DECISIONS I...
Colorblindness NORTON ET AL.and Diversity
Social Cognition, Vol. 26, No. 1, 2008, pp. 102-111
COLORBLINDNESS AND DIVERSITY: CONFLICTING GOALS IN DECISIONS INFLUENCED BY RACE Michael I. Norton Harvard Business School Joseph A. Vandello and Andrew Biga University of South Florida John M. Darley Princeton University
The present studies address a conundrum in contemporary American society: While many Americans agree that increasing racial diversity is a worthy goal, they are reluctant to acknowledge the impact of race on individual decisions in an effort to honor norms of colorblindness. In two studies, participants made hypothetical college admissions decisions among sets of equally qualified Black and White candidates. Study 1 revealed that participants justified decisions in favor of Black candidates by citing nonracial criteria, and that Whites—more concerned with appearing colorblind—were more likely than non–Whites to continue to use these criteria when making subsequent decisions. Study 2 showed that a preferred strategy for Whites to appear colorblind across multiple selections between candidates of different races was to select diverse sets of candidates, regardless of their specific qualifications. These results demonstrate the conflict Whites experience in their effort to appear colorblind while simultaneously increasing diversity.
In the wake of a Fifth Circuit Court of Appeals decision calling into question the use of race as a criterion for admission to colleges and universities (Hopwood v. Texas, 1996), admissions officers in Texas who were motivated to favor minority applicants for admission to college to meet the important goal of increasing diversity
Michael I. Norton, Harvard Business School; Joseph A. Vandello and Andrew Biga, Department of Psychology, University of South Florida; John M. Darley, Department of Psychology, Princeton University. The authors thank Sam Sommers for his comments on an earlier version of this manuscript. Correspondence concerning this article may be addressed to Michael I. Norton, Harvard Business School, Soldiers Field Road, Boston, MA 02163. E-mail: [email protected]
COLORBLINDNESS AND DIVERSITY
were confronted with a difficult challenge: How could applicants be favored on the basis of their race when the use of such information was prohibited? The New York Times listened in on an admissions committee at Rice University in Texas faced with this dilemma, and found that admissions officers managed to admit minority applicants by searching through their resumes for nonracial information which would support that candidate’s admission, such as their ability to overcome obstacles (Steinberg, 2002). We conducted a survey of college students at the University of South Florida (N = 160) which demonstrates the tension between diversity as an abstract goal and implementing that goal in concrete instances: though fully 70% of participants felt that diversity was an important consideration when determining the overall composition of an incoming class, just 10% felt that race should factor into any specific decision between two individuals. How do individuals manage these competing goals, making decisions on the basis of race for the sake of diversity, while honoring norms of colorblindness in any given decision (Norton, Sommers, Apfelbaum, Pura, & Ariely, 2006; Wolsko, Park, Judd, & Wittenbrink, 2000)? A large body of research has demonstrated that individuals are skilled at explaining and justifying their decisions (e.g., Nisbett & Wilson, 1977; Scott & Lyman, 1968; Shafir, Simonson, & Tversky, 1993) by selectively interpreting—and sometimes distorting—information (e.g., Brehm, 1956; Festinger, 1957; Mather, Shafir & Johnson, 2000; Russo, Medvec, & Meloy, 1996; see Kunda, 1990; Pyszczynski, & Greenberg, 1987). In the domain of social decision–making, previous research has demonstrated that individuals use strategies similar to the admissions officers at Rice, explaining choices between two individuals made on the basis of race in nonracial terms (Hodson, Dovidio, & Gaertner, 2002; Norton, Vandello, & Darley, 2004; see Norton, Sommers, Vandello, & Darley, 2006), part of a larger body of work suggesting that individuals are willing to engage in biased behavior only to the extent that they are able to justify it (e.g., Gaertner & Dovidio, 1986; Snyder, Kleck, Strenta, & Mentzer, 1979). Many real–world decisions, of course, involve multiple decisions between candidates of different ethnicities and different qualifications; indeed, such multi–stage decisions are quite common when employers hire new groups of employees or admissions committees determine the overall composition of an incoming class (see Fryer & Loury, 2005). The present investigation explores how individuals navigate these treacherous decision contexts when they are forced to make multiple decisions between individuals of different backgrounds. Imagine an admissions officer who has just selected a Black candidate over a White candidate while citing that candidate’s community service record as the basis for her decision—what will our admissions officer do when confronted with two additional candidates who vary on their level of community service? If these next two candidates are White, we might imagine our admissions officer continuing to select based on community service, attempting to appear unbiased by perpetuating the use of the criteria cited in the first decision; we explore this scenario in Study 1. But what if the second two candidates are of different ethnicities? Will the admissions officer pick based on community service, will she pick another minority candidate, or might she even be motivated to pick a White candidate over a Black candidate to ensure the appearance of colorblindness? We explore this scenario in Study 2. Both studies document the strategies individuals use when trying to meet their goals of choosing based on race while appearing colorblind.
NORTON ET AL.
STUDY 1 In Study 1, we used a scenario similar to the one with which we opened this paper, in which a college admissions officer is motivated to favor a Black candidate over a White candidate for admission to college. Previous research has demonstrated that given two equally qualified candidates—one Black candidate who has a higher GPA and one White candidate who has taken more challenging classes—individuals will select the Black candidate but refuse to admit that race played a factor, instead inflating the value of GPA and claiming it as the basis for their decision. If the White candidate has the higher GPA, however, participants will select the Black candidate and devalue the importance of GPA (see Norton et al., 2004). Extending this previous work, we explored whether making decisions under these social constraints would compel people to lock in their stated preferences when faced with a second choice between two new candidates, both White: One with a higher GPA, one with more AP classes. In addition, we explored a moderating role for the ethnicity of our participants in Study 1. In the survey we reported earlier, while just 10% of all participants felt that using race was appropriate in deciding between two specific candidates, over twice as many Black and Hispanic participants (17%) felt that using race was appropriate than Whites (7%), χ2 (1) = 4.03, p < .05. At the same time, both groups felt that diversity was an important goal, though non–Whites1 (78%) were somewhat more supportive than Whites (66%), χ2 (1) = 2.35, p = .13. Thus, both Whites and non–Whites endorse diversity as an important overarching goal, but are less comfortable with using race in a given decision—and it is Whites who are particularly uncomfortable with the idea. The two–decision design of Study 1 provides two measures of whether Whites and non–Whites might differ in the extent to which their different levels of comfort might impact their subsequent decision–making: Given preference for Black candidates in the first decision, we can explore whether participants will inflate the value of the criterion that favors that candidate when explaining their decision, but can also explore if they will continue to select based on this criterion in the second decision. We expected both groups to engage in justificatory strategies, but expected that their greater concern with appearing colorblind would cause Whites to go to greater lengths—by carrying their justifications further forward in time—than non–Whites.
METHOD Participants Sixty University of South Florida undergraduates (48 females) received course credit for their participation: 55% were White, 23% were Hispanic, and 22% were Black.
1. Both here and in Study 1, we compare White participants to Blacks and Hispanics combined (Non–Whites), as in other investigations such as the Boston Federal Reserve Mortgage Lending Study (Munnell, Tootell, Brown, & McEneaney, 1996).
COLORBLINDNESS AND DIVERSITY
Procedure Participants were instructed to imagine that they were on the Admissions Board at a top–ranked university with one final slot to fill in the incoming class. They reviewed resumes from two male high school seniors, and selected one candidate for admission. Each resume included a photograph, information about GPA, SAT scores, number of Advanced Placement (AP) classes, letters of recommendation and essays received, and extracurricular activities. The only substantive difference between the two was that one candidate had a superior GPA (4.0 vs. 3.6) whereas the other candidate had taken more AP classes (9 vs. 6). In one condition, the Black candidate had the higher GPA (and therefore fewer AP classes) while in the other the qualifications were flipped such that the Black candidate had the lower GPA (and therefore more AP classes). After selecting their preferred candidate, participants were asked to rank nine criteria (letters of recommendation, GPA, student government, number of AP classes, SAT verbal, SAT math, athletic participation, essays, and race) in order of their importance in their decisions. We were most interested in the relative rankings given to GPA and number of AP classes; in addition, including race as a criterion allows us to demonstrate that White participants were less willing to cite race as a criterion of importance than non–Whites. To test whether participants continued to use the same criteria in further selections, we then gave participants a surprise second decision between two additional White male candidates: One had a higher GPA (3.58) but fewer AP classes (2), while the other had a lower GPA (3.32) but more AP classes (5). We could thus test whether participants would be consistent in selecting the candidate whose qualifications matched that of the candidate they chose in the first selection.
RESULTS AND DISCUSSION First Selection As we expected, participants showed an overwhelming preference for the Black candidate, selecting him 92% of the time, χ2 (1) = 49.03, p < .001, a preference that held true for both White (88%) and non–White (93%) participants. Most importantly, selection was moderated by our manipulation of which criteria was associated with which candidate. When the Black candidate had taken more AP classes, fully 100% (of both White and non–White participants) selected this candidate; in the other condition, on the other hand, just 17% of participants (27% of Whites, 7% of non–Whites) selected the (White) candidate who had taken more AP classes, meaning that 83% suddenly preferred the candidate with the higher GPA, the Black candidate, χ2(1) = 42.86, p < .001. Regression analyses supported this account: The criterion assigned to the Black candidate was a significant predictor of choice (β = –.85, p < .001); because Whites and non–Whites engaged in the same strategies, there was no effect of participant race and no interaction (βs < –.11, ps > .15). This overwhelming preference for the Black candidate—regardless of qualifications—suggests the power of race in influencing preferences.
NORTON ET AL.
Justification We next checked to see if participants altered their rankings of nonracial qualifications to justify their choices. We created a dichotomous variable by coding whether participants had ranked GPA or number of AP classes higher. Mirroring the selection results, the importance of these criteria was moderated by our manipulation: When the Black candidate had taken more AP classes, some 50% ranked AP as more important than GPA; when the White candidate had taken more AP classes, however, just 7% ranked AP higher, meaning that 93% now believed that GPA—the criteria now associated with the Black candidate—was the more important criteria, χ2 (1) = 13.87, p < .01. Again, results were similar for White (44% and 7%) and non–White (58% and 7%) participants, and once again regression analyses supported this account: the criterion assigned to the Black candidate was a significant predictor of which criterion was ranked as more important (β = .47, p < .001) and there was no effect of participant race and no interaction (βs < .08, ps > .51). As expected, participants inflated the value of whichever qualification favored the Black candidate. Use of Race Despite the fact that participants showed a strong preference for the Black candidate, participants ranked race as the least important (M = 8.19, SD = 1.70) of the 9 criteria, significantly lower than athletic participation, the next lowest ranked criterion (M = 6.63, SD = 2.01), paired t(59) = 4.98, p < .001. White participants (M = 8.56, SD = .93) ranked race marginally lower than non–Whites (M = 7.74, SD = 2.26), t(58) = 1.90, p = .063, again suggesting that Whites were less comfortable admitting that race was a factor in their decision–making. Thus, while results for Whites and non–Whites were similar for both the first selection and justification, these results begin to suggest that the two groups may begin to diverge in their approach to these multi–stage decisions. Second Selection These differences between Whites and non–Whites were revealed when we examined choices in the second selection, between two additional White candidates. Unlike with the first selection, we observed no overall effect of our manipulation on the second selection: When the Black candidate had taken more AP classes in the first decision, 67% of participants selected the candidate that had taken more AP classes in the second decision, a number not significantly different from the 57% who selected the candidate who had taken more AP classes in the second decision who had seen a White candidate with more AP classes at time 1, χ2 <1, ns. Unlike results for the first selection, however, these results were impacted by the ethnicity of our participants. When the Black candidate had taken more AP classes at Time 1, Whites continued to select based on this criteria in the second decision, selecting this candidate 78% of the time; those Whites who had seen a White candidate with more AP classes in the first decision, on the other hand, selected the candidate with more AP classes in the second just 47% of the time, χ2 (1) = 3.42, p = .06. For non–White participants, on the other hand, there was no effect of the manipulation in the first decision on their second decisions (50% and 66%, respectively), χ2 <1, ns. Because our manipulation impacted the second decisions of only Whites (and not
COLORBLINDNESS AND DIVERSITY
non–Whites), we observed an interaction between condition and participant ethnicity (β = .25, p = .053), though this difference was only marginally significant; the simple effects of condition and participant ethnicity were not significant (βs < .08, ps > .56).
STUDY 2 In Study 1, while both White and non–White participants favored Black candidates and inflated the value of criteria that favored these candidates, only White participants—who were even less willing than non–Whites to admit that race played any role in their decision—continued to carry these criteria forward when making subsequent decisions. It thus appears as though White participants believe that using consistent (albeit arbitrary) criteria in subsequent decisions is an effective means of appearing colorblind when an initial decision is made on the basis of race. In Study 1, however, only the first decision involved candidates of different ethnicities, while the second decision was a “safer” one between two White candidates. In Study 2, we complicated participants’ task further by asking them to make consecutive decisions between two sets of candidates of different ethnicities. Two potential solutions to this problem follow from Study 1: First, if Whites are motivated to select Black candidates over equally qualified White candidates in every instance, they could simply select Black candidates in both decisions. On the other hand, if they maintain their consistency strategy from Study 1, they should select whichever candidate is better on the criterion they used in their first decision, whether that second candidate is White or Black. We suggest, however, that they may choose a third alternative to accomplish their goal of appearing colorblind: Forgoing either of the strategies above, and instead making an effort to select one candidate of each ethnicity (see also Drolet, 2002). Why might this be so? Clearly selecting two White candidates opens one up to accusations of racism, an extremely damaging label (see Sommers & Norton, 2006). But choosing two Black candidates might also lead one to be perceived as biased; indeed, accusations of “reverse discrimination” can also be damaging, as evidenced by lawsuits over the legality of affirmative action in college admissions (e.g., Gratz v. Bollinger, 2003; Grutter v. Bollinger, 2003). As an initial test of this hypothesis, we presented Harvard University undergraduates and graduate students (N = 118) with the photos and names of two sets of candidates, each set containing one White and one Black candidate. Participants were then told that a previous participant had selected either both Black candidates, both White candidates, or one White and one Black candidate. We asked participants to rate the extent to which they thought race played a role in this person’s decisions, on a five–point scale (1: no role to 5: a big role). As expected, participants reported that race played a larger role in the decisions of targets who chose two White (M = 2.84, SD = 1.28) or two Black (M = 3.03, SD = 1.23) candidates than those who chose one White and one Black candidate (M = 2.15, SD = .96), F(2, 112) = 5.85, p < .01. We also asked participants to rate how biased they thought the decision–maker was on a five–point scale (1: not at all to 5: very); not surprisingly, these ratings closely paralleled ratings of the role of race. Targets who chose two White (M = 2.72, SD = 1.28) or two Black (M = 2.66, SD = 1.15) candidates were seen as more biased than those who chose one White and one Black candidate (M = 2.05, SD = .92), F(2, 112) = 4.42, p < .02. Perhaps most interestingly, our sample was 77% White, 15% Black, and 7% Hispanic, and when we compared Whites and non–Whites, there was no interaction on either measure (Fs < 1). Both
NORTON ET AL.
Whites and non–Whites agreed that the best way to appear colorblind was to select one candidate of each ethnicity. In Study 2, we explore whether decision–makers forced to make two such decisions themselves use this effective strategy for appearing unbiased.
METHOD Participants Twenty–nine White University of South Florida undergraduates (20 female) received course credit for their participation. Procedure We used the same college admissions task as in Study 2. In this study, however, the Black candidate always had the higher GPA (4.0 vs. 3.6) and the White applicant always had more AP classes (9 vs. 6) in the first decision. After participants selected their candidate, they ranked the same nine criteria in order of their importance in their decisions. Participants were again confronted with a surprise second decision: Unlike in Study 1, in which both candidates were White, participants again saw one White and one Black candidate. We varied which candidate was superior on which dimension: In one version, the Black candidate had the higher GPA (3.58 vs. 3.32) while the White candidate had more AP classes (5 vs. 2); in the other version, the Black candidate had more AP classes and the White candidate had the higher GPA. Participants selected their candidate, then turned the page, completed a series of demographic questions, then rated the extent to which they felt it was appropriate to consider race when making decisions between two candidates (1: never appropriate to 5: always appropriate).
RESULTS AND DISCUSSION First Selection Replicating the previous study, the majority of participants (72%) selected the Black candidate, χ2 (1) = 5.83, p < .05. Justification Participants again justified this choice using nonracial criteria. The majority (83%) ranked the qualification that favored the Black candidate—GPA—more highly than AP classes, χ2 (1) = 12.45, p < .001. Use of Race Once again, participants ranked race last in importance to their decision (M = 8.17, SD = 1.44), significantly lower than personal statement, the next lowest ranked criterion (M = 6.66, SD = 1.70), paired t(28) = 3.14, p < .01.
COLORBLINDNESS AND DIVERSITY
Second Selection The second selection, between two new candidates who varied in race, was our measure of interest. In Study 1 we found that White participants continued to select based on the same criteria they had used in the first decision when choosing between two White candidates in a second decision. We might therefore expect to find a preference for the candidate with the higher GPA at Time 2, but instead we observed a significant preference for the candidate with more AP classes (69%), χ2 (1) = 4.17, p < .05. Thus, one strategy from Study 1—choosing based on the same criteria—was abandoned. Another possible strategy—making sure to select two Black candidates—was also not in evidence, as participants actually showed a significant reversal in preference, selecting the White candidate 59% of the time, χ2 (1) = 5.70, p < .02. If participants were not being consistent on race or on criterion, what strategy were they using? Participants’ choices—while violating being consistent on either criteria or race—were perfectly in line with results from our pretest: The majority of participants (66%) chose one White and one Black candidate, while far fewer participants chose two Black (24%) or two White (10%) candidates, χ2 (2) = 14.35, p < .01. But was it specifically concern about appearing colorblind that led participants to select one candidate of each ethnicity? We examined responses to the question which assessed participants’ beliefs about the appropriateness of using race to further explore their motivations. Interestingly, participants who chose either two White (M = 2.67, SD = 2.08) or two Black (M = 1.57, SD = .54) candidates reported thinking use of race was more appropriate than those who chose one White and one Black candidate (M = 1.21, SD = .42), F(2, 26) = 5.41, p < .01. Ironically, those participants who thought using race was most inappropriate were those who were most careful to select candidates of both races.2
GENERAL DISCUSSION In two studies, we showed that Whites attempt to cultivate an appearance of colorblindness when making decisions based on race, though the studies revealed quite different techniques for doing so. In Study 1, both White and non–White participants showed an overwhelming preference for Black candidates over equally qualified White candidates for admission to college, and both groups demonstrated post–decisional rationalization of their choices by inflating the value of criteria that favored their chosen candidates. Only White participants, however—more uncomfortable with their use of race—continued to use these same criteria in a subsequent decision between two additional White candidates. While White participants were a model of consistency in Study 1, they seemed to be a model of inconsistency in Study 2, making second selections that were neither consistent on the basis of race nor qualifications. Instead, participants were careful to use the strategy that a pilot study suggested was most effective for appearing unbiased: Selecting one candi2. Another explanation for participants switching from picking Black to picking White candidates is that demonstrating their lack of bias on the first selection licensed them to express bias in favor of Whites on the second decision (Monin & Miller, 2001). The results from the preliminary study, however, indicate that decision–makers who choose candidates of different races are seen as less biased, suggesting that participants may switch from a Black candidate to a White one not because they are licensed to express increased bias but because ironically they are trying to appear unbiased.
NORTON ET AL.
date of each race. These results demonstrate the importance that decision–makers assign to the goal of appearing colorblind, as they are willing to forgo cognitive consistency (Festinger, 1957) in order to meet what is seemingly a more important goal: Appearing unbiased. Finally, while our participants are college students and not actual college admissions officers, their efforts to balance out racial composition by favoring a White over a Black candidate after having favored a Black over a White candidate offer further evidence that desires to increase minority representation of college campuses—far from determining every decision made by admission committees— likely impact only a small percentage of college admissions decisions (Fryer & Loury, 2005). Still, our results do suggest that desires to honor diversity can lead individuals to favor minority applicants for admission to college in some cases, but important qualifiers are in order. Our participant population, college students, may be particularly concerned about appearing nonprejudiced due to heightened norms for political correctness across most college campuses, though such concerns are present in varying degrees across cultures and social groups (e.g., Galanter, 1984; see Sowell, 2004, for a review). von Hippel and Gonsalkorale (2005), for example, showed that non-Asian participants were more likely to eat unappetizing Asian cuisine when asked to do so by a member of that group. Clearly, however, such behavior remains far from the norm in most situations. While countless investigations have documented racism towards minorities, in domains ranging from housing (Turner et al., 2002) to employment (Bertrand & Mullainathan, 2004) to health care (Laveist, Arthur, Morgan, Plantholt, & Rubinstein, 2003), empirical studies that demonstrate favoritism toward minorities are few and far between (see Aberson & Ettlin, 2004). Indeed, research shows that the same processes demonstrated here underlie discriminatory decisions against females in employment decisions (Norton et al., 2004; Uhlmann & Cohen, 2005) and against minorities and females in jury selection (Norton, Sommers, & Brauner, 2007; Sommers & Norton, 2007). While biases toward and biases against minorities clearly are far from balancing out, the present investigation suggests that despite such different motivations (favoring majority or minority groups), individuals may utilize similar strategies to mask their decision–making.
REFERENCES change’s sake. Journal of Consumer ReAberson, C.L., & Ettlin, T.E. (2004). The aversive search, 29, 293–305. racism paradigm and favoring African Americans: Meta–evidence of two types Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University of favoritism. Social Justice Research, 17, Press. 25–46. Bertrand, M., & Mullainathan, S. (2004). Are Fryer, R.G., & Loury, G.C. (2005). Affirmative action and its mythology. Journal of EcoEmily and Greg more employable than nomic Perspectives, 19, 147–162. Lakisha and Jamal? A field experiment on labor market discrimination. American Gaertner, S.L., & Dovidio, J.F. (1986). The aversive form of racism. In J. Dovidio & S. Economic Review, 94, 991–1013. Gaertner (Ed.), Prejudice, discrimination, Brehm, J.W. (1956). Postdecision changes in the and racism (pp. 61–89). Orlando, FL: Acadesirability of alternatives. Journal of Abdemic Press. normal and Social Psychology, 52, 384–389. Drolet, A. (2002). Inherent rule variability in consumer choice: Changing rules for
COLORBLINDNESS AND DIVERSITY Galanter, M. (1984). Competing equalities: Law and the backward classes in India. Berkeley, CA: University of California Press. Gratz v. Bollinger. 539 U.S. 234 (2003). Grutter v. Bollinger. 539 U.S. 306 (2003). Hodson, G., Dovidio, J.F., & Gaertner, S.L. (2002). Processes in racial discrimination: Differential weighting of conflicting information. Personality and Social Psychology Bulletin, 28, 460–471. Hopwood v. Texas, 78 F.3d 932 (5th Cir. 1996). Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480–498. Laveist, T. A., Arthur, M., Morgan, A., Plantholt, S., & Rubinstein, M. (2003). Explaining racial differences in receipt of coronary angiography: The role of physician referral and physician specialty. Medical Care Research and Review, 60, 453–467. Mather, M., Shafir, E., & Johnson, M.K. (2000). Misremembrance of options past: Source monitoring and choice. Psychological Science, 11, 132–138. Monin, B., & Miller, D.T. (2001). Moral credentials and the expression of prejudice. Journal of Personality and Social Psychology, 81, 33–43. Munnell, A., Tootell, G. M. B., Brown, L. E., & McEneaney, J. (1996). Mortgage lending in Boston: Interpreting HMDA data. American Economic Review 86, 25–53. Nisbett, R.E., & Wilson, T.D. (1977). Telling more than we can know. Psychological Review, 84, 231–259. Norton, M.I., Sommers, S.R., Apfelbaum, E.P., P u r a , N. , & Ar i e l y , D . ( 2 0 0 6 ) . Colorblindness and interracial interaction: Playing the “Political Correctness Game.” Psychological Science, 17, 949–953. Norton, M.I., Sommers, S.R., & Brauner, S. (2007). Bias in jury selection: Justifying prohibited peremptory challenges. Journal of Behavioral Decision Making, 20, 467–479. Norton, M.I., Sommers, S.R., Vandello, J.A., & Darley, J.M. (2006). The psychology of mixed motives: The impact of race on decision–making. Psychology, Public Policy, and Law, 12, 36–55. Norton, M.I., Vandello, J.A., & Darley, J.M. (2004). Casuistry and social category bias. Journal of Personality and Social Psychology, 87, 817–831.
111 Pyszczynski, T., & Greenberg, J. (1987). Toward an integration of cognitive and motivational perspectives on social inference: A biased hypothesis–testing model. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 20, pp. 297–340). San Diego, CA: Academic Press. Russo, E.J., Medvec, V.H., & Meloy, M.G. (1996). The distortion of information during decisions. Organizational Behavior and Human Decision Processes, 66, 102–110. Scott, M.B., & Lyman, S.M. (1968). Accounts. American Sociological Review, 33, 46–62. Shafir, E., Simonson, I., & Tversky, A. (1993). Reason–based choice. Cognition, 49, 11–36. Snyder, M.L., Kleck, R.E., Strenta, A., & Mentzer, S.J. (1979). Avoidance of the handicapped: An attributional ambiguity analysis. Journal of Personality and Social Psychology, 37, 2297–2306. Sommers, S.R., & Norton, M.I. (2006). Lay theories about White racists: What constitutes racism (and what doesn’t). Group Processes and Intergroup Relations, 9, 117–138. Sommers, S.R., & Norton, M.I. (2007). Race–based judgments, race–neutral justifications: Experimental examination of peremptory use and the Batson challenge procedure. Law and Human Behavior, 31, 261–273. Sowell, T. (2004). Affirmative action around the world. New Haven: Yale University Press. Steinberg, J. (2002, December 8). Using synonyms for race, college strives for diversity. The New York Times. Turner, M.A., Freiberg, F., Godfrey, E., Herbig, C., Levy, D.K., & Smith, R.R. (2002). All other things being equal: A paired testing study of mortgage lending institutions. Washington, D.C.: The Urban Institute. Uhlmann, E.L., & Cohen, G.L. (2005). Constructed criteria: Redefining merit to justify discrimination. Psychological Science, 16, 474–480. von Hippel, W., & Gonsalkorale, K. (2005). “That is bloody revolting!”: Inhibitory control of thoughts better left unsaid. Psychological Science, 16, 497–500. Wolsko, C., Park, B., Judd, C.M., & Wittenbrink, B. (2000). Framing interethnic ideology: Effects of multicultural and colorblind perspectives on judgments of groups and individuals. Journal of Personality and Social Psychology, 78, 635–654.