First, we estimated frequencies of parental separation, substance use, and sexual debut, and average age at initiation of each substance and at sexual debut, in the sample as a whole and as a function of parental AUD status. Second, we used survival analysis to estimate the strength of association of parental separation and parental AUDs with each of the 4 outcomes: time in years since birth to initiation of alcohol, cigarette, and cannabis use, and sexual debut. Within this step, we (i) used Cox regression to test the associations of mother-only and father-only AUD with each outcome, then performed post-hoc Wald chi-square tests to assess the equality of the coefficients. Since the coefficients for mother-only and father-only AUD did not differ for any outcome (p-values ranging from .17 to .94), we combined them into a single variable representing alcohol problems in either parent; (ii) graphed cumulative failure rates using the Kaplan-Meier survival function. Third, we estimated Cox proportional hazards regression models predicting age at first use of alcohol, cigarettes, cannabis, and age at sexual initiation from parental AUD (modeled with
