112 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / JANUARY 2008 The Contextual Role of Alcohol Outlet Density in College Drinking* RICHARD SCRIBNER, M.D., M.P.H., † KAREN MASON, M.S., KATHERINE THEALL, PH.D., NEAL SIMONSEN, PH.D., SHARI KESSEL SCHNEIDER, M.S.PH., † LAURA GOMBERG TOWVIM, M.S.PH., † AND WILLIAM DEJONG, PH.D. † Louisiana State University School of Public Health, Epidemiology Program, 2120 Lakeshore Drive, New Orleans, Louisana 70122 Received: May 11, 2007. Revision: August 24, 2007. *This research was supported by National Institute on Alcohol Abuse and Alcoholism grant R01 AA014679. † Correspondence may be sent to Richard Scribner at the above address or via email at: rscrib@lsuhsc.edu. Shari Kessel Schneider and Laura Gomberg Towvim are with the Education Development Center, Inc., New- ton, MA. William DeJong is with the Department of Social and Behavioral Sciences and Youth Alcohol Prevention Center, Boston University School of Public Health, Boston, MA. 112 ABSTRACT. Objective: The objective of the study is to examine the relationship between the physical availability of off-campus alcohol and drinking outcomes among college students. Method: A multilevel analy- sis of students (N = 17,051) nested within college campuses (N = 32) was conducted. Four problem-drinking-related outcomes (i.e., average number of drinks when partying, frequency of drunkenness in past 2 weeks, 30-day frequency of drinking, and greatest number of drinks in one sitting) along with individual level covariates of drinking were in- troduced at the student level. The physical availability of alcohol was assessed as the number of on-premise and off-premise alcohol outlets within 3 miles of campus per 1,000 students enrolled. Results: Higher densities of on-premise alcohol outlets were strongly related to drink- ing outcomes even after controlling for individual predictors of college drinking. The association indicated that the campus means for the aver- age number of drinks when partying and the number of drinking occa- sions in the past 30 days were, respectively, 1.13 drinks and 1.32 occasions greater when the outlet density was 2 SDs higher. Conclu- sions: Off-campus, on-premise outlet density is strongly associated with college-drinking outcomes. Given the limited number of modifiable fac- tors that affect college drinking, on-premise outlet density represents a potential modifiable means of addressing the problem. (J. Stud. Alco- hol Drugs 69: 112-120, 2008) A LCOHOL-RELATED PROBLEMS AMONG college students represent a serious threat to American higher education. In 2002, a task force of the National Institute on Alcohol Abuse and Alcoholism (NIAAA) issued recom- mendations for addressing this problem. The task force’s report, “A Call to Action,” argued that reducing the magni- tude of the problem will require a change in campus cul- ture and that this change can occur only by addressing both individual and environmental factors that maintain that cul- ture (NIAAA, 2002). Within the college drinking literature, a large number of individual factors impacting college drinking have been identified, and a number of interventions have been devel- oped to address these factors with varying degrees of suc- cess (NIAAA, 2002). At the environmental level, the situation is much different. Although the literature frequently highlights the importance of environmental factors, only a few multicampus studies have been designed to character- ize the campus-level factors associated with college drink- ing (Toomey et al., 2007; Wechsler et al., 2000a,b). Instead, most research on environmental risk factors involves stud- ies of the environment at a single university (Clapp and Shillington, 2001) or uses compositional measures of the environmental risk based on aggregated students’ responses (Wechsler et al., 2000b). Although much can be learned from these studies, most of the constructs of interest are campus level or contextual, and examining them requires a study design in which predictors at both the student and the campus level are assessed directly. The College Alcohol Study (CAS), a series of four stan- dardized surveys conducted between 1993 and 2001 at 119 colleges and universities, has provided a number of oppor- tunities to study campus-level factors (Weschler et al., 2002). Historically the CAS’s primary focus has been to monitor changes in student-level risk over time and not to charac- terize campus-level differences between campuses. This is reflected in the use of the generalized estimating equation approach (Wechsler et al., 2002), which accounts for the between-campus variance in the grouping of drinkers but essentially treats this variance as a nuisance rather than a point of analysis for understanding the sources of between- campus variance. In addition, conducting standardized, cam- pus-level assessments has been beyond the scope of the CAS, presumably because of the expense entailed in di- rectly observing conditions at 119 campuses. Despite these limitations, the CAS has provided some of the most impor- tant campus-level research to date. For example, Weitzman et al. (2003) assessed the physical availability of alcohol (i.e., density of alcohol outlets within 2 miles of campus) at eight CAS sites and observed bivariate correlations with drinking levels. In addition, Kuo (2003) surveyed a sample