The Role of Explicit Categorization in the Implicit Association Test

Abstract

The present study investigated how task-irrelevant attributes of a stimulus affected responses in a multi-attribute version of the Implicit Association Test (IAT). In Experiment 1, participants categorized images of Black and White male and female individuals on the basis of either race or gender. Both the race and gender of the individuals affected task performance regardless of which attribute was currently relevant to performing the task, yielding the IAT effects on both attributes. However, the influences of a task-irrelevant attribute depended on whether the task-relevant attribute was categorized compatibly or incompatibly with the underlying implicit biases. These results suggest that individuals are still categorized implicitly based on task-irrelevant social attributes and that the explicit categorization required in the standard IAT has a considerable impact on implicit social biases. Experiment 2 considered a third, non-social attribute (the color of the picture frame) and reproduced task-irrelevant IAT effects and their dependence on explicit categorization. However, Experiments 3 and 4 suggested that the task-irrelevant IAT effects based on social attributes are determined by whether the task-relevant attribute is a social or non-social attribute. The results raise fundamental questions about the basic assumptions underpinning the interpretations of the results from the IAT.