People frequently expect those they interact with to behave in certain ways, to exhibit particular personality characteristics, or to demonstrate certain abilities or skills. Even when people's interpersonal expectancies are inaccurate, they nonetheless may form post-interaction impressions of others that are consistent with these expectancies. Such biased impressions may be especially damaging when they result from negative social stereotypes and prejudices (e.g., toward ethic groups, the mentally ill, the elderly, AIDS victims, etc.), as these impressions may lead to reduced opportunities (e.g., in employment, education, housing, etc.) for the stigmatized, and may serve to reinforce the initial negative expectancies. Although much work has investigated the processes that underlie expectancy confirmation, little research has attempted to identify those factors that determine when expectancy confirmation occurs and when it does not. Based on an analysis of the behavioral and cognitive mediators of the expectancy confirmation process, a model of expectancy- tinged dyadic social interaction is presented in which people's impression formation goals and self-presentational goals are posited to moderate expectancy confirmation. The model is based upon four primary assumptions. First, impression formation goals are proposed to regulate expectancy confirmation by moderating the extent to which people's expectancies influence the ways in which they gather information about others. Second, impression formation goals are proposed to also regulate expectancy confirmation by moderating the extent to which people's expectancies influence the ways in which they cognitively process the behavior of others. Third, self-presentational goals are proposed to regulate expectancy confirmation by moderating the extent to which people's expectancies influence their expressive behaviors toward other. And fourth, it is proposed that the allocation of behavioral and cognitive resources plays a vital role in determining both how, and to what extent, impression formation goals and self-presentational goals can effectively moderate the impact of expectancies. Preliminary research supporting several aspects of the model is described, and eight additional experiments, in which theoretically relevant perceiver goals and perceiver expectancies are manipulated within the context of relatively unconstrained social interactions, test the underlying assumptions of the presented model.
