Predicting binge drinking among undergraduate college students.

Abstract

The present investigation sought to provide a comprehensive examination of the binge drinking behavior of undergraduates at the University of Oklahoma so as to determine the best predictors of such behavior. Archival data consisting of a random sample of 503 undergraduates was analyzed utilizing multiple discriminant analysis procedures. A combination of demographic and social variables emerged as the strongest predictors of binge drinking with a modest overall classification accuracy (67.8%). Follow up robust tests of equality of means and subsequent Tamhane post hoc analyses were conducted. Consistent with previous investigations, variables such as Greek membership, ethnicity, use of tobacco products, and residency showed predictive power in predicting binge drinking. Moreover, frequent binge drinkers appeared to be a more psychologically maladjusted group using alcohol for self-medication purposes. Limitations and implications for future research are presented.