Twin and sibling pair correlations were lower in AA compared to EA women (Table 2). Details of the model-fitting procedure are presented in Supplementary Table 11. The model-fitting indicated that there were significant familial influences on both cigarette and cannabis use in AAs; however, we could not determine the extent to which this familiality was attributable to genetic, shared environmental, and/or twin specific environmental factors (either all A parameters could be dropped or all C+T parameters could be dropped, but a model dropping A, C, and T parameters simultaneously was rejected). For EAs, there were genetic influences on both cigarette and cannabis use and either source of familial environment, C or T, could be excluded from the model, but not both.
