My brand is +: when intersecting regularities transfer polarities instead of valence

Recent research suggests that valence can be transferred from one category to another through intersecting regularities (IR). For example, if two categories share a common response key, the valence of a source category (e.g., positive stimuli) can be transferred to a target category (e.g., a previously neutral brand). The present study investigates whether polarity differences of the source categories (i.e., asymmetric processing of these categories; Proctor & Cho, 2006) could be the underlying mechanism for differences in indirect and explicit measures. This hypothesis is motivated by the fact that the Implicit Association Test (IAT) is most often used as an indirect measure in IR tasks and is known to be susceptible to asymmetric processing. In other words, apparent valence differences between target categories could actually be artifacts arising from asymmetric processing of those categories, induced by asymmetry in the source categories. For explicit measures of valence, any observed differences would reflect genuine effects but might still be driven by the asymmetric processing of the source categories. Across four experiments, the polarities of the source categories were manipulated. The results show that polarity framing influenced IAT performance in Experiments 1 and 2. An alternative indirect measure—the Affect Misattribution Procedure—did not reveal valence differences (Experiment 3). The critical tests of the manipulation on explicit ratings were not significant in any individual experiment. However, a cross-experiment analysis revealed cautious evidence for a small effect.