These findings suggest several important directions for future research. First, these results need to be replicated, preferably in a longitudinal sample where true mediation can be assessed (as opposed to this cross-sectional sample). If these results are independently supported, then it would be important to establish the relative contributions of impulsive sensation seeking and parental history, perhaps via structural equation modeling. This would also allow for the inclusion of other relevant variables (e.g., externalizing psychopathology, level of response to alcohol) and would illustrate both the paths through which alcohol use and concomitant alcohol-tobacco use develop in adolescents and the relative strengths of associations. Second, the literature would benefit from further exploration of which neurocognitive or psychosocial traits associated with parental history of alcohol problems serve as the mechanisms of the risk imparted by parental history. While there is a clear relationship between parental history and greater likelihood of later alcohol use, later problem alcohol use and later concurrent alcohol-tobacco use, it is unclear what traits actually transmit the risk in this relationship.
