Identifying Sociodemographic Community Variables and High School Characteristics Predictive of Subsequent College...

Collegiate alcohol consumption is multi-determined by both precollege (e.g., institutional and community background, academic characteristics, precollege drinking) and college variables. However, precollege variables are rarely investigated. Self-reports of collegiate alcohol use were predicted by novel sources of data from university, census, and public administrative records.

3,720 first-time, college freshmen at a large, public, Midwestern university (54% female) participated in a longitudinal study of self-reported drinking prior to matriculation and the first four semesters. Data were merged with registrar records of ACT scores, high-school core GPA, first-generation college student status, high school administrative data and census tract information associated with senior year community of residence.

Factor analyses of census and high school characteristics data yielded two census factors of Socioeconomic Advantage and Black/Urban community, and four school factors of Graduation Rates, Predominantly White student composition, Human Capital and School Attendance. Academic Challenge, a reverse-scored composite of entering academic aptitude and achievement, was most strongly associated with increased alcohol use across all five assessments (ORs between 1.33-1.64 and 1.38-1.76 for men and women, respectively). Effects for the census variables were smaller and confined to men's precollege or freshman Fall assessments (OR's1.21-1.56) with students from higher SES and more Black/Urban communities reporting higher levels of alcohol involvement. With minor exceptions, no effects were found for school factors.

Academic Challenge appeared most associated with collegiate substance use across all assessments. Census and high school administrative records constitute additional sources of data which may improve prediction of substance use risk.