Reinforcing earlier findings from other data, college senior fraternity/sorority members are more likely to consume alcohol frequently. Large reductions in estimates upon controlling for time spent partying, and to a lesser extent cigarette use and intramural sports involvement, suggest considerable unobserved heterogeneity in the relationship. Yet, effects remain substantive and are invariant to conditioning on numerous further measures of socializing, sports participation, academic performance and mental health. The conclusion holds when non-member comparison groups are restricted to drinkers who smoke, party and/or play intramurals, or matched to members based on drinking propensities, suggesting that
fraternity/sorority membership raises alcohol use frequency.