Drinking Norm-Behavior Association Over Time Using Retrospective and Daily Measures

Objective: Social norm based interventions in college drinking are common but show mixed efficacy. Although such interventions assume a passive social-influence process, past research relied heavily on retrospective measures, leaving open the possibility that heuristic biases during recall may alternatively account for or inflate estimates of social influence from prospective norm drinking associations. The present study examined this possibility, using retrospective and daily aggregated measures of self and perceived peer drinking behavior. Method: For each of 3 years, students (N = 574; 288 men) reported on their drinking levels and perceptions of descriptive drinking norms, using., conventional retrospective reports over a month period and daily diary reports for 30 days. Using structural equation modeling, we tested cross-lag longitudinal models for evidence of social-influence/alternative processes and compared cross-lag effects across retrospective and daily aggregate models to determine the extent to which heuristic recall biases contribute to the norm-behavior association. Results: Perceptions of social norms had a small but reliable effect on changing drinking behavior across years, as indicated by model comparisons. Past drinking behavior also consistently shaped changing perceptions of drinking norms. These effects were not attributable to, nor inflated by, heuristic biases during retrospective reporting of personal and peer behavior. Conclusions: These results suggest that social influence and not heuristic biases contribute to the long-term norm drinking association but that alternative processes, whereby past drinking behavior shapes norm perceptions, contribute more to the norm drinking association. Implications for interventions designed to reduce college drinking are discussed. (J Stud. Alcohol Drugs, 71, 769-777, 2010)