The role of stereotyping in overconfident social prediction

A field study methodology was used to study the role that stereotyping and prototypicality assessment play in overconfident social prediction. Preliminary research verified the existence of distinct stereotypes about inhabitants of two campus residences, generated suitable behavioral prediction items relevant to those stereotypes, and demonstrated that those stereotypes afforded some predictive validity for residents in general and self-described prototypic residents in particular. Our principal study demonstrated that other students on campus, responding only to targets' photographs and residences, made predictions with greater subjective confidence than justified by their objective accuracy. Such overconfidence, however, resulted less from using stereotypes than from abandoning such stereotypes on the basis of dubiously valid (i.e., appearance-based) impressions about residents' prototypicality. The participants general tendency to express high confidence, even for counter-stereotypic predictions, and their inability to distinguish items offering high versus low stereotype validity also contributed significantly to the overconfidence phenomenon. Theoretical implications and limitations of the present methodology and results are discussed.