This article explores the relationship between the everyday use of statistical reasoning and the perception of erroneous group stereotypes. Recent research concerning a reasoning process analogous to a statistical analysis of covariance (ANCOVA) is reviewed. This research implies that the failure to engage in ANCOVA reasoning may play a role in the development and maintenance of stereotypes. This research also reveals that people can and do sometimes engage in sophisticated ANCOVA reasoning, and thus avoid forming erroneous group stereotypes. A number of factors influencing ANCOVA reasoning and group perception are identified, and the cognitive processes underlying their effects are considered. Finally, speculation is offered concerning the implications of statistical reasoning research for developing interventions that might help prevent the formation of erroneous group stereotypes.