Alcohol use among college students: generating behavior prediction models for social and personal motivational contexts
Alcohol use and abuse are widespread and serious concerns for the college student population (Carnegie Foundation for the Advancement of Teaching, 1990). A multitude of studies have attempted to uncover the factors that influence college students' decisions about alcohol use. However, most preventive intervention programs that draw upon this research have failed to significantly influence college student drinking, suggesting that the complex network of factors involved in these decisions has not yet been uncovered(Meacci, 1990). In the present investigation, behavior prediction models for socially and personally motivated alcohol use among college students were generated through exploratory factor analysis and CALIS structural modeling procedures, instead of forcing the data to fit into a preconceived pattern. Adequate models were constructed for both the social and personal contexts across the model building sample of participants (n=167). Further, a cross-validation test (n=95) revealed that the models are likely to generalize to other populations of college students. Supplemental tests of model fit performed on men's (n=101)and women's (n=161) responses separately indicated that the social context model best predicts men's college alcohol use. However, women's responses resulted in a less effident fit to the social context model than did men's responses and a problematic fit to the personal context model. These results suggest that further investigations may be warranted in order to discover optimal models for personal context alcohol use and for women college students' alcohol use.