Some hypothesized predictors of drinking onset did not perform as expected in this sample. Regarding prior reports of the relation of race to an earlier AFD (Dawson et al., 2008), the SDPS is largely Caucasian and this demographic breakdown may have hindered finding any racial/ethnic differences. As for lower parent education as a risk factor, the current sample was also fairly homogenous, with most families having parents with higher education reflective of the original selection criteria. Finally, these subjects reported low rates of depressive symptoms, tobacco use, and other illicit drug use which may have reflected their relatively young age. Significant effects of these characteristics may emerge with increased sample size and the availability of an older age range in future data collections. This issue may be especially relevant to the performance of tobacco use as a predictor of drinking onset, as <10% of this sample had previous smoking experience. The relatively low rate of smoking might also reflect the relatively high education levels of the families in the sample.
