Among families of European ancestry, effects of parental separation and parental alcoholism on timing of substance involvement were substantial, even in Cox models adjusting for family background, offspring psychopathology, and physical and sexual abuse during childhood. Compared to twins from nonalcoholic intact families, twins from families where a parent was alcoholic or parents separated were more likely to initiate substance involvement during very early adolescence. Risks to twins from alcoholic separated families were further increased, suggestive of a dose-response relationship. Results from propensity score analyses largely confirm survival-analytic findings for European ancestry twins, increasing our confidence in the specificity of risks associated with parental separation. Effects of parental separation on offspring substance involvement were observed across much of the risk spectrum and, despite elevated rates of parental alcoholism, twins from intact families at high predicted probability of separation demonstrated reduced risk, particularly for alcohol and cannabis involvement.
