Abuse of alcohol is a major problem on many college campuses. To reduce problem drinking, interventions must be directed toward those factors associated with problem drinking. In this study we examined how perceptions of the role of alcohol related to problem drinking. We surveyed 318 college students who were taking health related classes. Students voluntarily completed questionnaires in their regular classroom setting. The questionnaire included nineteen items concerned with drinking behavior or perceptions regarding the effects of alcohol. Data were analyzed using SAS programs to conduct frequency counts, factor analysis, and multiple regression. Results indicated that among female students, 77.16% reported having ever been drunk; 22.54% reported having been drunk more than 20 times in the last year; 49% reported having been drunk at least one time in the last month. Among male students 85.71% reported having ever been drunk; 46.03% reported having been drunk more than 20 times in the last year; 66.34% reported having been drunk at least once in the last month. Factor analysis allowed us to identify a problem drinking factor and two alcohol perception factors (items were included in a factor if they loaded at .60 or higher). The problem drinking factor consisted of seven items. An example item concerned the number of times the student had been drunk in the last 12 months. The alcohol perception factors were called “social bonding” (five items) and sexiness (three items). Example items are “alcohol facilitates a connection with peers” (social bonding) and “alcohol makes women sexier” (sexiness). Multiple regression was conducted using problem drinking as the dependent variable and social bonding and sexiness as the predictor variables. Separate analyses were conducted for males and females. Results indicated that for both males and females the two predictor variables accounted for a significant (p<.0001) amount of the variation in problem drinking (Rsquare = .312 for females and .204 for males) with social bonding the more important of the two predictor variables (Rsquare for social bonding = .283 for females and .202 for males). Results suggest that programs to prevent problem drinking may wish to address the issue of alcohol and social bonding by: (1) identifying positive ways to promote social bonding that avoid problem drinking (or avoid alcohol altogether), (2) demonstrating that problem drinking can often be detrimental to social bonding. Future research should examine the impact of interventions that address the relationship of problem drinking and social bonding.