Influence of Environmental and Social Characteristics of Drinking Contexts On College Students' High Risk Drinking: A Multi-Level Analysis of a Randomized Controlled Trial

Introduction:

High-risk alcohol use has been identified by U.S. college presidents as the number one campus life problem (NIAAA, 2005) due to its association with multiple adverse outcomes (Hingson et al., 2009). While numerous studies have investigated individual characteristics that are predictive of heavy alcohol use (Baer, 2002; Kuntsche et al., 2005), research focused on the social contexts in which students drink is limited (Paschall & Saltz, 2007).  The purpose of this study was to examine the environmental and social characteristics of drinking contexts that serve as risk and protective factors of high-risk drinking.

Methods:

We analyzed fall 2010 and 2011 self-report data from California public university students (N= 6903 students in 14 schools).  We examined how background (age, gender, race/ethnicity, and past drinking behavior), environmental (keg availability, monetary charge for admission, and enforcement of the minimum drinking age) and social (party size and percent intoxicated) characteristics influenced students’ drinking to intoxication at six settings. For each setting, Stata v12 was used to estimate four multilevel logistic regression models (students nested within schools): 1) null model, 2) background only, 3) background and environmental, and 4) the full model with background, environmental, and social parameters.

Results:

Over 50% of respondents drank to intoxication the last time they attended a party set at a Greek House, Residence Hall, On-Campus event, or Off-Campus house party. There was variation by context in environmental and social characteristics. Multilevel logistic regression analyses showed the presence of a keg was positively associated with drinking to intoxication at a Greek party (OR = 1.41, p<0.01); this association was present in model 3 for campus, off campus, and outdoor parties but attenuated in the full model.  Charging a monetary cost for admission was associated with a higher odd of drinking to intoxication at a bar in model 3, and this association was attenuated in the full model.   In the fully estimated models, enforcement of the minimum drinking age was associated with a lower odds of drinking to intoxication at a campus event; this association was also observed for the bar setting in model 3, and attenuated in the fully estimated model. For every setting, there was a positive association between the percent of attendees at a party perceived to be drunk and drinking to intoxication.

Conclusions:

These results highlight that, depending on the setting, limiting the social availability of alcohol and enforcing the minimum drinking age can serve to reduce the odds of drinking to intoxication. Future research should examine mechanisms by which these factors influence behavior.