an additional 11% of this variance could be explained when the moderating effect of teen susceptibility to influence was taken into account. Given that simple stability of substance use accounted for almost half of the variance in age 16 use, this means that the composite marker of apparent susceptibility to teen influence was accounting for almost 20% of the residual or change variance in substance use over time. Given likely noise and error variance in this substance use measure, this suggests a potentially quite strong process being tapped by this factor. Notably, prior research on potential peer influences in adolescence has generally found only modest effects, and rightly noted that much apparent peer influence is simply due to selection of similarly behaving peers (Dishion & Owen, 2002; Ennett & Bauman, 1994). The present results suggests that we may be in danger of substantially underestimating the importance of peer influence processes if we do not account for the likelihood that such processes are far more applicable to some teens than to others.
