and frequency of alcohol use as well as the number of times each of 27 alcohol-related problems occurred (see Supplemental Material for examples). From these assessments, the following alcohol outcomes were calculated by dividing the respective cumulative totals by the number of years of drinking and performing a square root variance-stabilizing transformation to address the fact that they could be considered count data: mean number of drinks consumed per year, days on which binge drinking occurred (defined as ≥ 5 drinks) per year, and times alcohol-related problems were experienced per year (mean age at assessment was 24.2 [SD 2.7, range 17.0–31.0]). In addition, in light of recent changes to the definition of alcohol and other substance use disorders in DSM-5 (APA, 2013), we combined DSM-IV abuse and dependence into one lifetime AUD outcome (absent/present); mean age of most recent AUD assessment was 23.4 (SD 3.2, range 15.7–30.8). The mean number of drinking years was 11.5 (SD 3.5, range 3.3–23.0). Mean age at first drink (i.e., more than a sip) was 13.7 (SD 2.4, range 5.0–19.3). Despite using SRE scores from only the first five drinking occasions to reduce the potential confounding effect of acquired tolerance, we included drinks before SRE-5
