The current study has some limitations that should be kept in mind when interpreting results. First, substance use history was reported retrospectively and was therefore subject to potential recall biases. However, by using the earliest reports of first use (across multiple waves of data collection) and adjusting for lag times between age at first use and age at time of report, we were able to minimize these biases while gathering data from women who had passed through the age of risk for substance use initiation. Second, although our final model of cross-substance use initiation provides clear support for the primary role of heritable factors, the wide range in confidence intervals for A and C in the univariate models suggests that evidence should be interpreted as preliminary and that further studies are needed to confirm that familial influences on substance use onset are attributable to genetic rather than shared environmental sources. Third, due to power limitations, we were unable to test the equal environment assumption (EEA), the assumption that correlations in exposure to environmental events relevant to the trait under study
