This experiment was undertaken to test two contrasting explanations of the effects of eye gaze on social perceptions and outcomes. A social meaning model holds that differing levels of gaze have such clear meaning that gaze alone accounts for the reactions to it. A nonverbal expectancy violations model holds that normative behaviors are expected in social interactions with strangers and that violating these expectations produces different results depending on whether the violator is deemed highly “rewarding” or “nonrewarding.” This experiment, the third in a series, proposed to extend the violations model by incorporating the concept of positive and negative types ofviolations. Subjects (N = 145) interviewed one of four confederate interviewees who manipulated one of three levels of eye gaze (nearly constant, normal, andnearly constant aversion) and who were assigned one of two levels of reward (highly qualified for the job or highly unqualified). Differential gaze behavior resulted in varied impressions of attraction, credibility, and relational communication, with gaze aversion producing consistently negative effects. Interpretations and communication consequences were mediated by reward, gender, and confederate differences.