Obama v. Cognitive Bias

The New York Times recently reported on a study “showing that a performance gap between African-Americans and whites on a 20-question test administered before Mr. Obama’s nomination all but disappeared when the exam was administered after his acceptance speech and again after the presidential election.”

Earlier researchers had “assembled university students with identical SAT scores and administered tests to them, discovering that blacks performed significantly poorer when asked at the start to fill out a form identifying themselves by race. The researchers attributed those results to anxiety that caused them to tighten up during exams in which they risked confirming a racial stereotype.” Reviewers of the new study theorized that “Obama’s election could increase the sense of competence among African-Americans, and it could reduce the anxiety associated with taking difficult test questions.”

This reminded me of an earlier study that found that individuals’ implict biases shift significantly when they are immersed in situations that provide frequent exposure to admired members of traditionally stigmatized groups (e.g., famous African Americans or older people) and disliked members of traditionally valued groups (e.g.,infamous whites or younger people).

This involved use of the Implicit Association Test (IAT), which uses an individual test-taker’s response speed to assess the relative strength with which he or she associates certain targets with positive or negative attributes. For example, participants are typically asked to “classify stimuli representing two target concepts (e.g., flowers and insects, or Black and White groups) and evaluative attributes (e.g., good and bad words) using two designated keys. When the IAT is used to measure racial attitudes, people typically respond more quickly and easily if pleasant attributes share the same response key with White racial stimuli and unpleasant attributes share the same key with Black racial stimuli than vice versa.”

But the authors found that test-takers whose initial performance indicated a tendency to associate black targets with negative attributes (and a tendency to associate white targets with positive evaluations) displayed significant less implict bias of that sort after repeated exposure to pictures of admired African Americans (e.g., Dr, King) and disliked whites (e.g., Jeffrey Dahmer). Similarly, the authors found that test-takers’ automatic preference for younger over older people declined after encountering images of admired older and disliked younger individuals.

Together, these studies make me wonder about the effects we might expect now that pictures of President Obama are prominently displayed throughout the media as well as in government offices throughout the country . . .

You may also like...

2 Responses

  1. Gary Rosin says:

    Prof. Gelman (Columbia) has a discussion of this study on his blog “Statistical Modeling, Causal Inference and Social Science” blog:

  2. Gary Rosin says:

    Somehow the URL did not come thru: