| The Implicit Association Test (IAT; Greenwald, McGhee, Schwartz, 1998) has been presented to the research community and to the public as a measure of attitudes. More than this, it has been said to have the ability to measure prejudices, stereotypes, and self-conceptions that people do not realize that they have. The task operates on the premise that, when two concepts are conceptually close (i.e., a “match” of concepts), it is easier to pair them together in a single judgment than when the two concepts are conceptually distinct (i.e., a “mismatch” of concepts). Performance on this categorization task is operationalized as the difference in reaction time for match judgments and mismatch judgments. These difference scores are then interpreted as measures of various social constructs, particularly attitudes, stereotypes, and self-concepts. In this dissertation, I provide a review showing that difference scores influence statistical measures of associations in ways that are not consistent with the interpretation of past IAT results. More damaging, I will show that an accurate interpretation of these past associations fails to support past conclusions made using the measure. Following this review, I report two experiments showing that these statistical properties obscure fairly basic relationships between IAT measures and external criterion scores. They also created the appearance of relationships that were mistakenly interpreted as evidence that the IAT is a valid measure of attitudes, stereotypes and self-concepts. In response to the problems of difference scores, I applied a more defensible analytic method to the IAT, based on linear regression. This was done to determine if responses on match judgments interacted with responses on mismatch judgments in theoretically meaningful ways to predict criterion scores. No evidence of such interactions was found. This review and these results shed serious doubt on the validity of the IAT as a measure of attitudes, stereotypes, and identities. |
