Salience asymmetries in the Implicit Association Test

Abstract

The Implicit Association Test (IAT) is the most popular indirect measure of attitudes in social psychology. It has been suggested that salience asymmetries are a non-associative contaminant of the IAT that threatens the accurate assessment of attitudes. Salience asymmetries in the IAT are claimed to correspond with visual search asymmetries, and differences in target familiarity. In this thesis, I propose that processing fluency is the common mechanism underlying both visual search asymmetries and familiarity. Several experiments were conducted to determine whether visual search asymmetries, familiarity, or processing fluency most reliably corresponds with salience asymmetry effects in the IAT.

The first series of experiments revealed that processing fluency is a better predictor of salience asymmetry effects in the IAT than is visual search asymmetry (Chapter 2). In Chapter 3, a novel method was developed to distinguish between the effects of valence and salience in the IAT. Using this method, I demonstrated that the effects of salience in the IAT are consistent with a fluency account of salience asymmetries. Familiarity was also shown to produce salience asymmetry effects in the IAT (Chapter 4), which is also consistent with the fluency account. When fluency and familiarity were set against each other in Chapter 5, it was processing fluency, rather than familiarity, that predicted salience asymmetry effects in the IAT. Although processing fluency is a good predictor of salience asymmetries, the results of Chapter 6 reveal that the fluency account cannot explain all examples of salience asymmetries in the IAT.

The data presented here are consistent with the view that the more fluently processed target category is compatible with the pleasant attributes on the grounds of salience asymmetries. The current experiments suggest that when there are valence differences between the target categories, salience asymmetries can potentially distort IAT effects. When the positive target category is more salient, salience asymmetries appear to increase IAT effects. In contrast, when the negative target category is more salient, salience asymmetries appear to decrease IAT effects. However, further evidence is required to determine how the effects of salience and valence combine in the IAT.