Second, to address the potential concern that the SSAGA-IV, which is a clinical interview, was not sensitive to adequately assessing environmental domains, we examined interactions between close friend substance use frequency measured from the IPA and PRS for heavy episodic drinking outcomes. Results of these regression models were somewhat consistent with the primary analyses (Supplemental Table 3), although PRS was not predictive of initial heavy episodic drinking status in these models. This may have been due to the limited power to detect effects (given the small sample size) as well as the possibility of cohort effects, as the smaller sample size was due to the fact that only certain COGA sites administered this instrument. Furthermore, the assessment of close friend substance use frequency was not completely analogous, as the IPA indexed the severity of close friend alcohol use whereas the SSAGA indexed the number of close friends who use substances.
