Contextual variation of alcohol consumption and related problems among American college students

The severity of alcohol consumption and alcohol-related problems among American college students is heterogeneous depending on the contexts where they occur. Identifying contextual sources of variation and assessing their magnitude can serve as an important rationale for developing appropriate alcohol control policies. We examined the contextual variation of alcohol consumption and alcohol-related problems among a sample (N=54307) from the 1993, 1997, 1999 and 2001 Harvard School of Public Health College Alcohol Study of 119 colleges from 39 states in the US. Multilevel analyses were conducted on a three-level structure, where individuals were nested within colleges and colleges within states, and time was treated as a linear term. The state-level context (σ²=0.32, p<0.05) contributed a slightly greater proportion of the overall variation of binge drinking (binary outcome) than that of the college-level context (σ²=0.19, p<0.05). The combined contribution was 14%. With respect to whether students drank alcohol in the past year, the state-level variation was about two times greater than the variation from the college-level. Yet when the outcomes were related to heavy drinking (i.e., drunken in the last 30 days or frequent binge) as well as alcohol-related problems (i.e., hang over) among all students, the magnitude of the state-level variations were substantially lower, and college-level variations contributed a relatively larger contextual effect. Identifying and assessing contextual variations on a variety of drinking outcomes help guide research focus on college and state specific characteristics and related alcohol policies to reduce drinking-related problems.