ABSTRACT. Objective: This study examined how social-infl uence processes operate during specifi c drinking contexts as well as the stability and change in these processes throughout the college years. Method: Using a measurement-burst design, a hybrid of longitudinal and daily diary methods, we assessed the relationship between event-specifi c descriptive drinking norms and personal drinking. College students (N = 523) completed a baseline survey followed by a 30-day daily diary each year for up to the 4 study years. The baseline survey assessed participant gender and social anxiety, and the daily survey assessed personal drinking and perceived peer drinking (i.e., event-specifi c descriptive norms) during social drinking events. Results: Multilevel modeling revealed that men's social drinking slightly increased over the 4 years, whereas women's drinking remained steady. Further, on social drinking days when event-specifi c descriptive norms were high, students drank more, but this relationship was stronger for men than women and did not change over time. However, men's drinking norm perceptions increased across years, whereas women's decreased. Social anxiety did not moderate the relationship between norms and drinking. Conclusions: We demonstrate that although gender differences exist in the stability and change of personal drinking, norms, and normative infl uence on drinking across the years of college, the acute social infl uence of the norm on personal drinking remains a stable and important predictor of drinking throughout college. Our fi ndings can assist with the identifi cation of how, when, and for whom to target social infl uence-based interventions aimed at reducing drinking. (J. Stud. Alcohol Drugs, 72, 633-641, 2011)