Aims: To test the relative contribution of two key Social Learning Theory constructs, alcohol expectancies (AEs) and drinking refusal self-efficacy (DRSE), in predicting early adolescent drinking behavior and examine the possible mediational role of DRSE over AE.
Methods: High school students (N = 192, mean age 14) were administered measures assessing AE (Drinking Expectancy Questionnaire--Adolescent version; DEQ-A), DRSE (Drinking Refusal Self-Efficacy Questionnaire--Revised Adolescent version; DRSEQ-RA) and indices of alcohol consumption and problem drinking. Age, gender, peer drinking, tobacco use and positive and negative behavioral characteristics were included in the statistical models as known predictors of alcohol misuse. Subjects were followed up at 12 months, with 88.5% retention.
Results: Initial confirmatory factor analyses verified factor structures of the DEQ-A and DRSEQ-RA. Prospective structural models controlling for Time 1 drinking behavior, age, gender, peer alcohol use, tobacco use and behavior problems identified that DRSE but not AE was associated with problem drinking 12-month post-initial assessment. DRSE mediated AE in predicting problem drinking.
Conclusion: Results suggest that DRSE is a more salient cognitive construct than AE in early adolescence alcohol use. In this age group, prevention and treatment strategies that build refusal self-efficacy may be more effective than strategies that challenge AEs.