Dimensions of adolescent problem drinking

Issues involved in defining and measuring adolescent problem drinking have occupied the attention of alcohol researchers since the mid-1970s. However, appropriate definitions and measures of problem drinking remain important and unanswered conceptual and research questions. The research reported in this study was designed to address these issues by identifying multiple dimensions of adolescent problem drinking and examining correlates of these dimensions using multivariate regression models. A longitudinal sample of 2,771 students who had tried alcohol was used in these analyses. The sample was followed over three waves of data collection, between 1985 and 1990, in middle and high schools located in a southeastern U.S. county. Data from the third wave were used to characterize problem-drinking behavior. Results of exploratory factor analyses supported the conclusion made by others that adolescent problem drinking is a multidimensional phenomenon. Three dimensions measuring (1) level or frequency of alcohol use, (2) problems related to drinking and (3) symptoms of dependency were identified and were only moderately intercorrelated. The independent measures used to model the three problem-drinking measures were the same for each model, and the significant predictors and proportions of variance explained by the predictors differed widely across models. Results of the regression models confirmed the uniqueness of the three measures and suggested that the potency of specific risk factors varies for different types of problem drinking. Prevention implications and recommendations for this area of research are discussed.