Drinking onset was assessed with the question, “How old were you the first time you had a drink?” A drink was defined as a full drink (12 oz beer, 5 oz wine, or 1.5 oz spirits). To allow for comparison with Dawson and colleagues (2007), drinking onset was categorized into 3 groups: young (age ≤14), middle (age 15 to 17), and older onset (age ≥18). To allow a nonlinear effect of onset age, we used 2 dummy variables in the analyses: ONS17 indicated onset by age 17 (1 = onset age ≤17, 0 = onset age >17), ONS14 indicated onset by age 14 (1 = onset age ≤14, 0 = onset age >14). Using these codes allowed us to first estimate the effect of drinking by age 17 and then estimate whether there was a significant additional effect associated with drinking by age 14.
