Success and failure in social norms interventions.

摘要

In their recent Journal of Studies on Alcohol article, Perkins and associates (2005) examined the statistical relationship among normative perceptions of drinking and self-reported drinking behavior among college students. In their introduction, Perkins et al. reviewed a study that we published in the Journal (Clapp et al., 2003). We wish to correct their representation of our findings.Our study (Clapp et al., 2003) utilized a quasi-experimental design to test the efficacy of a norms social marketing campaign. We employed an intensive norms social marketing campaign in an experimental residence hall and used students residing in a second hall as a comparison group. We collected pre- and postcampaign data at both halls to test the effectiveness of the campaign in reducing several drinking outcomes. Our analytical approach employed analysis of covariance (ANCOVA) models computed in the general linear model function in SPSS (SPSS Inc., Chicago, Illinois). Perkins et al. (2005) correctly reported our main finding: "Students exposed to the campaign subsequently perceived significantly less amounts [of alcohol] being consumed at parties, but failed to demonstrate a statistically significant reduction in the amounts consumed while increasing their days of drinking in the last 28 days, relative to the control group" (p.471). The authors, however, made several factual and conceptual errors in interpreting our other results.