Five hundred and ninety-two Anglo, Mexican American, and American Indian students from nine selected 2-year colleges in the Southwest were given a self-administered questionnaire with items related to problem drinking (dependent variable) and various social psychological and demographic information (independent variables). Analyses included discriminant analysis, chi square, and eta. While ethnicity per se contributed little to the discrimination of problem from nonproblem drinkers, several other variables provided an overall correct classification (discrimination) rate of 79.73%. There variables were, in order of discriminating power: neighborhood people drink, veteran status, live life differently, close friend with drinking proble, drink because acquaintances drink, and age. It is hypothesized that the psychological principle of individual differences, and the significance of some other study variables, may haver overshadowed any ethnic effect. Measurement problems in this area are discussed as well as some potential research directions.