Unobservable stereotypes are more malleable than observable stereotypes in implicit person perception

The present research proposed a theoretical distinction among various stereotypes that we predicted would moderate their malleability in implicit person perception: the extent to which the stereotypes can be learned and validated with minimal or no indirect inference (i.e., their observability). We hypothesized that observable stereotypes would be less malleable than unobservable stereotypes in implicit person perception in the presence of counterstereotypic individuating information. This main hypothesis was tested in four studies, as were two alternative hypotheses that all implicit stereotype-relevant evaluations would show evidence of fast-learning, and that all such evaluations would provide support for slow-learning. Studies 1 and 2 tested these predictions in the domain of an observable (Study 1) and an unobservable (Study 2) racial stereotype, and Studies 3 and 4 in the domain of an observable (Study 3) and an unobservable (Study 4) gender stereotype. Considered as an aggregate, Studies 1 and 2 provided strong support for the hypothesis that the observability of an implicit stereotype would moderate its malleability. Studies 3 and 4 showed only limited support for this hypothesis. However, when all four studies were considered together, the hypothesis was supported. The finding that observable stereotypes in implicit person perception are less malleable than unobservable stereotypes is discussed in the context of current debates regarding the processes underlying implicit and explicit social cognition.