High-risk behaviors, such as driving after drinking and getting into a physical fight, are common on college campuses. There are clear and strong associations between alcohol consumption and increased likelihood of risky behaviors amongst college students. A dose-response model developed by Gruenewald and colleagues suggests there may be substantial differences in the level of risk associated with a specific level of alcohl consumption between typically light and heavy drinkers. Somewhat surprisingly, while a strong association between alcohol consumption and high-risk behaviors has been established, little is known about the degree to which drinking contexts may affect those risks. Furthermore, no one has quantified the distinct risks associated with (1) selection of a drinking context, (2) use of alcohol in that context, and (3) level of use in that context. The current application uses a quantitative theoretical framework to assess heterogeneity of dose-response and context-specific risks related to drinking using Safer California Colleges, a large survey data set available from a recently concluded evaluation of the impacts of environmental intervention programs on college drinking. From 2003 to 2008, 42,171 college student surveys were collected, with 79.5% of respondents reporting that they drank in the past year. We use data from 14,252 drinkers in 2003 and 2004 (pre-intervention years). Students provided data on drinking patterns, use of drinking contexts (e.g., Greek parties, parties in residence halls), characteristics of those contexts (e.g., other intoxicated persons), and high-risk behaviors (e.g., rode with a driver who was drunk). These replicated cross-sectional data provide assessments of college drinkers'use of drinking environments across 14 campuses in California.
The aims of the proposed project are to (1) Determine whether college students'probability of engaging in high-risk behaviors varies by alcohol dose and (2) Quantify the extent to which high-risk behaviors vary in relation to background non-drinking risks, risks related to drinking, and level o use in different contexts. We will also examine whether context-specific risks are moderated by time spent in each context, number of persons in that context, and others'drinking levels (subjectively assessed). This information can be used to target preventive interventions to reduce access to high-risk contexts, eliminate all drinking in high-risk contexts, and/or reduce heavy drinking in high-risk contexts. Understanding the conditions under which alcohol use is most likely to lead to risky behaviors will benefit both prevention practitioners and consumers of this information, and will facilitate campus-wide efforts to intervene in specific high-risk drinkig settings (e.g., cooperative community programs to restrict off-campus access and use).
We propose research to determine whether college students'probability of engaging in high-risk behaviors (e.g., riding with a driver who was drunk) varies by alcohol dose and to quantify three distinct risks due to selection of a drinking context, use of alcohol in that context, and level of use in that context. This information can be used to target preventive interventions to reduce access to high-risk contexts, eliminate all drinking in high-ris contexts, and/or reduce heavy drinking in high-risk contexts. Understanding the conditions under which alcohol use is most likely to lead to risky behaviors will benefit both prevention practitioners and consumers of this information and will aid in the development of recommendations with a much more practical edge, such as risks are elevated when attending off-campus parties regardless of drinking level.