Abstract
The relationship between religiosity and alcohol use among a sample of middle-adolescent boys and girls was examined using two-wave (with a 1-year interval) prospective longitudinal data. A primary aim was to determine the extent to which religion differentially predicts three outcomes: (a) the decision to use alcohol, (b) the choice among users of how much and how often to drink, and (c) the experience of alcohol-related problems. Regression analyses revealed that religious salience was associated negatively with later decisions to use alcohol (net of the effects for gender and age), but this association became nonsignificant when controlling for peer, family, and school influences. Likewise, although religious attendance had a moderate negative relationship with subsequent alcohol problems, this relationship disappeared when more fully specified models were estimated. By contrast, religious attendance predicted decreases in the quantity and frequency of alcohol use even in the presence of peer, family, and school variables.