Aim. This study tested mediator, moderator and main effects models to determine the pattern of relationships between ethnicity, gender, temperament, and alcohol use . Design. A large-scale survey was employed. Respondents were interviewed in their homes. Participants were 523 Puerto Rican and 490 Irish American adults residing in the New York metropolitan area . Measures. Quantity and frequency of alcohol consumption were assessed with standard measures, and recent drinking problems with the substance abuse module of the Composite International Diagnostic Instrument . Findings. Rather than supporting a complex, nested model, the findings supported a main effects model and indicated that each of the three predictors was related to one or more drinking measure. While gender strongly influenced each drinking behavior, ethnicity only influenced frequency, and temperament only influenced recent drinking problems . Conclusions. The results support previous studies indicating consistent male-female drinking differences while highlighting a statistical strategy for comparing competing models of ethnic drinking behavior.