Evaluation of a level of response to alcohol-based structural equation model in adolescents.

OBJECTIVE A low level of response (LR) to alcohol relates to a family history of alcoholism and predicts future heavier drinking and alcohol-related problems. The current analyses evaluate how LR functions within the context of a Social Information Processing Model, using 238 subjects aged 13 to 19 years from the Collaborative Study on the Genetics of Alcoholism (COGA). METHOD A structural equation model (SEM) was used to evaluate the relationship among (1) a family history (FH) of alcoholism, (2) the LR to alcohol, (3) expectations as measured by the Alcohol Expectancy Questionnaire and (4) the recent pattern of drinking among parents in the home as predictors of the maximum recent quantity of drinking, the maximum number of drinks ever consumed in 24 hours and the number of alcohol problems in the teenage subjects. RESULTS When tested in the SEM, LR functioned as a mediator of the relationship between an FH of alcoholism and alcoholic outcome, and expectancy functioned as a partial mediator of the relationship between LR and outcome. Invariance testing revealed that the SEM performed similarly in light and heavy drinkers, the two sexes, and older versus young subjects. Both the measurement model and SEM had good characteristics of fit, and direct paths within the model explained 49% of the variance of outcome. CONCLUSIONS Consistent with results from the San Diego Prospective Study, when tested in the more heterogeneous COGA population, LR functioned as a mediator of the relationship between the FH and alcoholic outcome. In these adolescents, LR not only had a direct relationship to alcoholic outcome but, in contrast to prior results in adults, also appeared to operate through expectations of the effects of alcohol in predicting heavier drinking and a higher rate of alcohol-related problems among teenagers.