The relationship between attitudes and the Implicit Association Test.

Abstract

The goal of these studies was to closely examine the relationship between the Implicit Association Test (IAT) and explicit attitudes. In Studies 1 and 2, I examined the correlations between the IAT and explicitly measured attitudes in a fine-grained analysis. Correlations between the IAT and explicit attitude measures failed to emerge despite examining explicit attitudes in two independent samples, at multiple levels of specificity, and in a domain in a domain where participants were willing and able to report their explicit attitudes. In Study 3, I investigated the relationship between the IAT, explicitly measured attitudes, and behavior and found that while explicit attitudes predicted a choice behavior, the IAT did not. Taken together, these results support a dual process view of the IAT and explicit attitudes in which the IAT and explicit attitudes tap separate, independent underlying processes. In Study 4, I found that the IAT was affected by exposing participants to new associations between attitude objects, while the explicit attitudes remained unchanged. These results support an environmental association interpretation of the IAT in which IAT scores are interpreted to reflect the extent to which various attitudinal objects are associated in a person's environment rather than the extent to which the person endorses the attitude objects. Finally, in Studies 5 and 6, I investigated the possibility that importance moderates the relationship between the IAT and explicit attitudes. Although the results of Study 5 were inconclusive, Study 6 provided evidence for a strong positive relationship between IAT scores and their explicit attitudes when the attitude domain was rated to be of high importance, but a much weaker correlation when the attitude domain was rated to be of low importance. These studies provide evidence that the IAT is not simply a better or more direct measure of attitudes than explicit attitude measures, but these two types of measures tap different underlying constructs. The IAT measures exposure to associations, whereas explicit attitude measures tap endorsement of associations.