than unlabeled. Those in the labeled group had lower family incomes, less education, lower levels of perceived social support and social involvement, and were more likely to have a first-degree relative or a live-in partner with alcohol problems. Age, gender, racial/ethnic group status, and levels of PAS were not significantly different across groups. Supplementary Material S41 provides a partial correlation matrix for the endogenous variables of the SEM model. We note that compared to our analytic sample, those who were not in the analytic sample had significantly lower rates of past-year AUD (72.3% vs. 2.6%, respectively) and past-year internalizing disorders (33% vs. 19%), but similar levels of ISS (M = 42.6 vs. M = 42.5) and PAS (M = 37.4 vs. M = 37.9, respectively) (data not shown).
