Toward a meaningful metric of implicit prejudice.

Abstract

[Correction Notice: An Erratum for this article was reported in Vol 100(5) of Journal of Applied Psychology (see record 2015-40760-001). there are errors in some of the values listed in Table 6 that do not alter any of the conclusions or substantive statements in the original article. The corrected portion of Table 6 is in the correction. The positive intercepts in this table represent the estimated IAT score when the criterion has a value of zero (suggesting attitudinal neutrality), except in the equation examining voter preference in Greenwald et al. (2009), where the intercept estimated the IAT score of Obama voters.] The modal distribution of the Implicit Association Test (IAT) is commonly interpreted as showing high levels of implicit prejudice among Americans. These interpretations have fueled calls for changes in organizational and legal practices, but such applications are problematic because the IAT is scored on an arbitrary psychological metric. The present research was designed to make the IAT metric less arbitrary by determining the scores on IAT measures that are associated with observable racial or ethnic bias. By reexamining data from published studies, we found evidence that the IAT metric is "right biased," such that individuals who are behaviorally neutral tend to have positive IAT scores. Current scoring conventions fail to take into account these dynamics and can lead to faulty inferences about the prevalence of implicit prejudice.

A closer look at the discrimination outcomes in the IAT literature

“…Blanton et al . () found that participants with neutral (non‐discriminatory) scores actually had positive (biased) IAT‐scores and concluded that the IAT must be right‐biased (e.g., showing too much bias) since it is not properly aligned with the discrimination outcome. They regard these discrimination outcomes as the gold standard to compare the IAT against (although they do admit that they have imperfections).…”

“…The discussion has so far almost exclusively focused on either the conceptual link (e.g., does implicit bias predict discrimination) or on methodological issues with the IAT (e.g., if the zero point is meaningful; Blanton et al ., ). In the present research, we will switch the focus to the discrimination outcomes themselves.…”

“…They conclude that the link between IAT and discrimination is, at best, very weak. Furthermore, in a recent related study, Blanton, Jaccard, Strauts, Mitchell and Tetlock () turn things around and predict the IAT‐scores based on the behavioral outcome. They find a peculiar pattern in that those who have neutral outcomes (i.e., treat people equally) actually show bias against the ethnic or racial minority on the IAT.…”

Using the IAT to predict ethnic and racial discrimination: Small effect sizes of unknown societal significance.

“…Relying on zero-order correlations between the IAT and all criteria in general is too crude and begs the fundamental question of the behavioral implications of different IAT scores (Blanton, Jaccard, & Burrows, 2015;Blanton, Jaccard, Strauts, Mitchell, & Tetlock, 2015). Greenwald et al (2015) offer two examples that they contend illustrate how the average effect size estimates for race IATcriterion correlations might have real-world consequences.…”

Not So Fast: Ten Challenges to Importing Implicit Attitude Measures to Media Psychology

“…Blanton et al (2006) found evidence against the assumption of equal b 2 and b 4 weights for the IAT and, indeed, Greenwald, Nosek, and Banaji (2003) offered a scoring algorithm, in part, as an explicit acknowledgment of the problem. Unfortunately, the new scoring algorithm adjusts for GPS in ways that introduce a host of new problems (see Blanton, Jaccard, Strauts, Mitchell, & Tetlock, 2015).…”