Does the Brief Implicit Association Test Measure Semantic or Affective Valence Representations?

Valence, the representation of a stimulus as positive or negative, is fundamental to conceptualizing attitudes and their empirical research. Valence has two potential representations: semantic and affective. The current line of studies investigates the degree to which the congruency effect of the Brief Implicit Association Test (BIAT), often used as an indirect evaluation measure, reflects affective or semantic aspects of valence. In three preregistered experiments (N = 1,056, with 352 participants each), we examined how the congruency effect of the BIAT reflects these aspects. In all three experiments, we used a repeated exposure manipulation, which typically causes a habituation effect on affective but not on semantic aspects of valence, to differentiate between the two types. In the first experiment, repeated exposure occurred before the BIAT, while in the second and third experiments, it was performed in the context of the BIAT task. We utilized three dependent variables: feelings-focused self-reports (measuring participants’ reports about their feelings), knowledge-focused self-reports (measuring semantic evaluations), and the BIAT congruence effect. Supported by Bayesian analysis, we found consistent evidence that the repeated exposure manipulation influenced feelings-focused self-reports but did not affect knowledge-focused self-reports or the BIAT. The results suggest that the BIAT effect is sensitive to semantic (and not affective) representations of valence. Implications for attitude theory and measurement are discussed.