Alcohol perceptions and behavior in a residential peer social network. Personalized normative feedback is a recommended component of alcohol interventions targeting college students. However, normative data are commonly collected through campus-based surveys, not through actual participant-referent relationship. In the present investigation, we examined how misperceptions of resident hall peers, both overall using a global question and those designated as important peers using person-specific questions, were related to students’ personal drinking behaviors. Participants were 108 students (88% freshman, 54% White, 51% female) residing in a single campus residence hall. Participants completed an online baseline survey in which they reported their own alcohol use and perception of peer alcohol use using both an individual peer network measure and a global peer perception measure of their residential peers. We employed network autocorrelation models, which account for the inherent correlation between observations, to test hypotheses. Overall, participants accurately perceived the drinking of nominated friends but overestimated the drinking of residential peers. Consistent with hypotheses, overestimation nominated friend and global residential peer drinking predictor. Interaction analyses showed that the relationship between global misperception and participant self-reported drinking was significant for heavy drinkers, but not non-heavy drinkers. The current findings explicate how student perceptions of peer drinking within an established of normative feedback interventions.