The sample was approximately evenly divided among ascertainment sites, with 101 (16.7%) probands from Connecticut, 74 (12.2%) from Indiana, 88 (14.5%) from Iowa, 122 (20.1%) from New York, 81 (13.4%) from St. Louis, and 140 (23.1%) from California. Probands who were not interviewed at follow-up and thus excluded from this analysis, compared to probands who were interviewed at both time points, met fewer lifetime AUD criteria (M[SD]=9.45[1.89] versus 9.82[1.57], t(1245)=−3.72, p<.001) and conduct disorder criteria (M[SD]=2.01[2.14] versus 2.36[2.02], t(1245)= −2.92, p<.01), on average, and had a lower prevalence of major depressive disorder (14.96% versus 22.95%, χ2(1)=11.41, p<.001). Probands who were excluded because they did not have a first-degree relative with AUD, compared to those with an affected relative, met fewer conduct disorder criteria (M[SD]=2.04[1.84] versus 2.46[2.08], t(782)= −2.44, p=.02), and had a lower prevalence of females (20.79% versus 28.22%, χ2(1)=3.90, p=.05).
