RESUMO
BACKGROUND: Risk factors of alcohol abuse generally have been examined for their additive or direct effects on the development of alcohol-related problems. This study was designed to assess the interaction between two important predictors: parental history of alcohol problems and positive expectancies regarding alcohol consumption. METHODS: The positive expectancies, reported parental alcohol problems, and alcohol use and problems of 169 first-year college students were assessed at their entrance to college, and their alcohol use and problems were assessed again approximately 3 months later. The main effects of positive family history and expectancies as well as their interaction in predicting changes in alcohol use and problems were examined in hierarchical regression analyses. RESULTS: A parental history of alcohol-related problems was related to greater alcohol problems at the two assessment times, and positive expectancies for the effects of alcohol were related to both alcohol problems and alcohol use. The parental history x positive expectancies term significantly added to the prediction of changes in alcohol problems (8% incremental increase in explained variance) even after baseline problems were entered in a prior step in the equation. Follow-up exploration revealed that the interaction was accounted for by high levels of alcohol problems at time 2 being reported by those high in positive expectancies and reporting high parental alcohol problems. The parental history x positive expectancies interaction term also added slightly to the prediction of changes in alcohol use amounts but accounted for a modest 1% incremental variance. CONCLUSIONS: Reported parental history and positive alcohol expectancies interacted to predict increases in alcohol problems over the course of the first semester of college. These results suggest that risk factors for alcohol problems may conjointly interact to confer heightened risk. Such interactive models may further assist in identifying at risk young adults. Limitations such as the reliance on self-report measures and the predominantly female sample are discussed.
