Method: Ninth graders (N = 3,293, MAge = 14.08, 53% female) from 10 high schools participated from 2012 to 2013. Youth completed three surveys, 6 months apart over 12 months. SSC was measured with the 10-item social self-control scale (SSCS; Sussman et al., 2003) at Time 1. PE and NE of the effects of alcohol use were measured using the 12-item Modified Lyons Battery for Subjective Effects (MLBSE; Zeigler et al., 2010) at Time 2. Adolescents reported their quantity of alcohol use during the previous 30 days at Time 3.
Results: To account for nesting within schools and extreme non-normality of alcohol use, zero-inflated generalized linear mixed models for alcohol use were estimated. Lower SSC was associated with increases in alcohol consumption (RR = 1.07, 95% CI[1.04, 1.09]). Lower SSC was also associated with greater PE (β = .03, 95% CI[0.00, 0.06), and greater NE (β = .07, 95% CI[0.04, 0.10). With subjective evaluations and SSC estimated in the same model, only PE predicted increased number of drinks (RR = 1.16, 95% CI[0.06, 1.26]); SSC was not a significant predictor of alcohol use when controlling for PE (p = .17) and NE (p = .20). The indirect effects for PE were indicated to be greater than zero (p = .03), according to a Sobel test, but not for NE (p = .62).
Conclusions: The mediating effect of PE on the relationship between SSC and drinking may reflect these youth’s attempts to lubricate social interaction and cope with the repercussions of their negative actions toward peers. Although adolescents with low SSC may experience negative outcomes at a higher rate than their peers, potentially as a result of their irrational and aggressive behavior, NE did not predict alcohol use in the mediation model; additional mediators or moderators of the associations between SSC and alcohol use should be explored. Implications for adolescent substance use prevention include: 1) increasing teens’ abilities to engage in effective social interaction, and 2) altering PE of alcohol use.