Not all Stereotype-Incongruent Information is Created Equal: The Impact of Sample Variability on Stereotype Change

In two experiments participants saw stereotype-incongruent information that was either high or low in variability. Following exposure participants generated distributions of group members from which central tendency and variability information was calculated. The comprehensive retrieval model (Garcia-Marques & Mackie, 1999) predicted greater change in perceived dispersion following high versus low variability incongruent information, but greater central tendency change following low versus high variability exposure. In Experiment 1 participants generated more variable distributions after exposure to high versus low variability inconsistent information and showed greater central tendency revision in the low compared to high variability condition, supporting predictions. These effects were replicated in Experiment 2, which demonstrated similar effects for congruent information.