Thus, the current results must be viewed in light of the methods used. First, the sample was relatively small and it is possible some of the results regarding cross correlations and predictors of T2 drinking may have been significant if more subjects had been available. Second, data are offered only on males as they were the largest group available for these prospective analyses, and the outcome may be different for women. Third, subjects were all from Southern California, were predominantly White, and were recruited from a narrow age range. In addition, these men had alcoholic relatives, and it is difficult to know whether the same results would be seen for a family history negative group. These issues combine to raise questions about both the robustness and generalizability of the findings, and additional studies are needed. A fourth caveat reflects our choice of using oral (as opposed to intravenous alcohol challenges [Ramchandani et al., 2002]), because the oral route adds variability to the BAC. While that point is worth noting, our procedure of giving alcohol based on both weight and percent
