Our findings provide a foundation for further investigation of the genetics of alcohol- and cannabis-related phenotypes, both gene-mapping and studies of the interaction of inherited vulnerabilities with environmental influences that lead to these outcomes. However, there are possible limitations that need to be kept in mind. One is the potential for retrospective reporting biases to influence results, for example, for greater lag time from the event to the time of reporting to translate into systematic underreporting of use and symptoms (Carney et al., 1998; Stockwell et al., 2004). Given the relatively narrow age range (and thus the relatively narrow range in lag times) in the sample and our adjustment for age at the time of interview, it is unlikely that such a bias greatly influenced results, but reports from prospective studies would be more accurate. To the extent to which the relative influences of heritable and environmental factors vary across cultures, the generalizability of our findings may also be somewhat limited. Conducting investigations similar to ours in other populations would reveal the nature of cross-cultural distinctions and commonalities in the
