Although research has examined the content of alcohol-outcome expectancies and also the role of alcohol use in aggressive and sexual behaviors, few studies have linked the two lines of inquiry. To examine the efficacy of outcome expectancies for predicting actual behavior, 64 male social drinkers, aged 21 to 25 years, completed questionnaires and, three to five days later, participated in an experimental session. The questionnaires assessed personal effects of alcohol as well as hostility and sex guilt. Subjects were divided into two groups: "expect-alcohol" and "expect-tonic." The groups were divided again into those who received tonic only and those who actually received the alcohol (one part vodka to five parts tonic). During the sequenced experimental session, subjects first interacted with a confederate, then consumed enough alcohol (or placebo) to become moderately intoxicated, and then rated four types of stimulus slides at their own pace. The slides depicted violence, mild erotica, violent erotica, and neutral scenery. Unobstrusive recordings were made of the time spent viewing and rating each slide. An analysis of the results showed that pre-existing expectations regarding the psychosocial effects of alcohol consumption had a clear influence on overt behavior. Specifically, expectations that alcohol enhanced aggressiveness, disinhibition, and euphoria led to increased viewing of violent and deviant slide materials, for subjects who believed that their drinks contained alcohol. In contrast, these alcohol-outcome expectancies were not predictive of viewing times for subjects who did not expect alcoholic drinks. These findings are consistent with a cognitive social learning perspective, which would support a direct link between expectancies and post-consumption behavior. (BL)