96 Social 45 (20091 961-964 hat th ponses,we d and gender as predictors.Analysis of the eval studies IAT.par less implicit stereotyping..(6)-.20.p.03. exposure to a pro -stereotypic exempla Study2 imin 1 led Black pl.We what l as in parti oth Am icipants' accessibility Blac that ack on nind w uld iudice re esulted from increase ciations bet een Black pe soposited change in hen the s alities hav for dy 2.we ex undan ed with lo activated when pr e Study Participants and procedure DT,and a :80% for 55 ms fo o.a mas skand hen arget rds or 1 nts completed tive and ing IATs ated (eg government president politician) rted 10r9 arget words eineut g-.ve ne to ded be criminali (e ds wer a common ster aire replaced with words that reflect"entaand" ed Ob eg. on a(strongly disagree)to(strongly agree)scale anticpatadcometoeterpeoplesinds hen they Results and discussion ont to th nary study and Studv there w of voting for a on a 1(hi e evalu score M-19 Results ower than previous IAT at the stiutio ealed a sizabledro IAT D-s e (M 02 alth gh the D- 50)or the ore (M 38)a0 stan ally lower th. o years prio ndstereope 2006:Devine et al.2002:Greenwald et al 1998) attitudes.latency to government words following theWhite We hypothesized that the salience of Obama created through the media blitz and strong support for his candidacy was responsi￾ble for the reduced levels of implicit stereotyping and prejudice. Hence during the Fall of 2008 when Obama was the Democratic presidential nominee and the publicity for the campaign was in￾tense, we conducted two studies to examine whether we repli￾cated the low level of implicit bias and whether the level of implicit bias was related to factors indicating that the bias reduc￾tion resulted from exposure to Obama. Indeed, the campaign pro￾vided a naturally occurring event during which the impact of exposure to a prominent counter-stereotypic exemplar could be examined. In Study 1, we examined whether the degree of implicit bias was related to the extent to which Obama as well as other po￾sitive exemplars were highly accessible when people thought of Black people. We also examined what participants anticipated comes to mind for others when they think of African Americans to explore social tuning effects. We hypothesized that having posi￾tive Black exemplars highly accessible in one’s mind would be re￾lated to lower levels of implicit prejudice and stereotyping. We also posited that if exposure to Obama produced a change in the pattern of associations activated when encountering Black peo￾ple, then the social category ‘‘Black” should activate qualities asso￾ciated with Obama; further, activation of these qualities should be associated with low implicit bias. Therefore, in Study 2, we exam￾ined whether qualities strongly associated with Obama as a polit￾ical figure (e.g., government, leader) were activated when primed with ‘‘Black” and whether activation of these qualities was related to the level of implicit prejudice. Study 1 Participants and procedure Two hundred and twenty-nine students (64% female; 80% White) at a Midwestern university participated for course extra￾credit. Participants completed evaluative and stereotyping IATs (Amodio & Devine, 2006) in counterbalanced order. On the evalu￾ative IAT, Black and White faces are sorted according to race and words (e.g., love, evil) were sorted according to valence (i.e., pleas￾ant/unpleasant). On the stereotyping IAT, the valence words were replaced with words that reflect ‘‘mental” and ‘‘physical” catego￾ries (e.g., brainy, run). Participants then listed the first five thoughts that came to mind when they thought of Black people and the first five thoughts they anticipated come to other people’smindswhen they think of Black peo￾ple. We recorded whether Obama and other positive exemplars were listed (e.g., Martin Luther King). Participants also rated their likelihood of voting for Obama on a 1 (highly unlikely) to 7 (highly likely) scale. Results D-scores (Greenwald, Nosek, & Banaji, 2003) were created for both IATs. Consistent with the preliminary study, we failed to ob￾serve anti-Black bias on either the evaluative IAT D-score (M = .03, SD = .50) or the stereotyping IAT D-score (M = .03, SD = .38) and neither was significantly different from zero, t(229) = .81, p = .42 and t(229) = 1.07, p = .28, respectively. These findings stand in stark contrast to prior work that has consistently found evaluative and stereotype IAT scores significantly greater than zero (see Table 1), indicating implicit bias against Blacks (e.g., Amodio & Devine, 2006; Devine et al., 2002; Greenwald et al., 1998).1 To examine whether thoughts of positive exemplars were asso￾ciated with participants’ implicit responses, we conducted regres￾sion analyses with the positive exemplar variable, likelihood of voting for Obama, and gender as predictors. Analysis of the evalu￾ative IAT revealed that participants who listed a positive exemplar on either list were less likely to respond with anti-Black bias, b = .17, t(226) = 2.52, p = .01. Similarly, for the stereotyping IAT, participants who listed a positive exemplar responded with less implicit stereotyping, b = .15, t(226) = 2.20, p = .03. Study 2 Consistent with our preliminary study, Study 1 revealed unusu￾ally low levels of implicit prejudice and stereotyping as well as ini￾tial evidence that the reduction of implicit bias is due to participants’ increased accessibility of positive Black exemplars. Study 2 was designed to examine whether decreases in implicit prejudice resulted from increased associations between Black peo￾ple and traits associated with Obama as a political figure. Thus, participants performed a lexical decision task (LDT) assessing the strength of association between Black primes and government-re￾lated words.2 In this study, participants completed only the evalua￾tive IAT because the stereotyping IAT would have been redundant with the LDT. We predicted that increased accessibility of govern￾ment-related words following a Black prime would be associated with decreased implicit prejudice. Methods Seventy-nine students (68% female, 85% White) at a Southern university participated for course-credit. Participants completed the evaluative IAT, the LDT, and a questionnaire. The trials of the LDT presented one of three primes (‘‘Black”, ‘‘White”, or ‘‘XXXXX”) for 55 ms followed by a mask and then one of 18 target words or 18 non-words (Dovidio, Evans, & Tyler, 1986). Target words included government-related (e.g., government, president, politician), crime-related (e.g., criminal), or neutral (e.g., vehicle) words. Crime-related target words were included because criminality is a common stereotype of Black Americans (Devine & Elliot, 1995). The questionnaire assessed participants’ likelihood of voting for Obama as in Study 1 and whether they believed Obama’s nomina￾tion had positively influenced their attitudes toward Black people on a 1 (strongly disagree) to 9 (strongly agree) scale. Results and discussion Contrary to the preliminary study and Study 1, there was some evidence of anti-Black bias on the evaluative IAT (D-score M = .19, SD = .20) t(76) = 8.27. p < .001. To determine if this level of bias was lower than previous IAT scores measured at the same institution, we merged these data with data collected during the Fall of 2006 (n = 41). The analysis revealed a sizable drop in D-scores (from .56 to .19), t(117) = 6.56. p < .001, d = 1.16, suggesting that although the D-scores differed significantly from 0, they were sub￾stantially lower than scores obtained just two years prior. Next we examined whether implicit prejudice was related to activation of government-related words when primed with ‘‘Black”. Participants’ D-scores were regressed on the latency to re￾spond to government words following the ‘‘Black” prime. The like￾lihood of voting for Obama, belief that Obama had changed their attitudes, latency to government words following the ‘‘White” 1 In both studies, we found that levels of explicit prejudice, measured via feeling thermometers and the Attitudes toward Blacks Scale (Brigham, 1993), were equiv￾alent to previous levels. 2 Fourteen undergraduates listed the first three words that came to mind when they thought of Barack Obama. Government-related words (e.g., president, leader) were tied with ‘‘Black” as the most frequently occurring. 962 E.A. Plant et al. / Journal of Experimental Social Psychology 45 (2009) 961–964