This important report by Slomkowski et al. (2005) underscores the importance of genetically informed research designs for gaining new understanding of environmental influences on the use of tobacco, alcohol or other addictive substances. The paper accomplishes this goal in at least two ways. First, it demonstrates that the smoking behavior of one sibling is related significantly to the smoking behavior of a second sibling, even after controlling for the degree of genetic relatedness between the two siblings. Indeed, the findings showed that the correspondence in sibling smoking was equally high for unrelated as for full siblings or dizygotic twins in at least some situations. Results such as these make it clear that the shared environments of siblings, in addition to non-shared influences and genetic similarities, are likely to play a powerful role in generating risk for tobacco and, quite probably, other types of substance use. This demonstration of environmental effects net of common genes simply cannot be achieved in the absence of a genetically informed research design.

A second important contribution of the study was to show that a potentially important interpersonal mechanism, social connectedness, provided a plausible means through which siblings might influence one another's smoking behaviors. That is, the findings indicated that the shared environmental effect, but not the genetic effect, was moderated by social connectedness. When siblings were highly involved with one another in diverse activities, then one sibling's smoking was especially likely to be associated with the second sibling's smoking. Again, this influence operated net of any genetic effects. It seems reasonable that siblings who spend time together, and who share common friends and activities, would be more likely to engage in substance-using behaviors to a similar degree, if for no other reason than that their high level of social involvement would provide numerous opportunities for one sibling to be present when the other sibling is smoking, drinking or using illicit drugs. When common and perhaps substance-using friends are added to this equation, it seems it would be difficult for an initially non-smoking sibling not to at least experiment with smoking.

As important as these two major contributions are, however, the study also sets the stage for gaining more complete understanding of the multiple pathways through which family and peers probably affect risk for smoking and other forms of substance use. Simply put, genetically informed studies of this type increase confidence in the reality of environmental influences on variations in individual behaviors, and that confidence enhances the value of more traditional observational studies that do not include a behavioral genetics component. For example, Slomkowski and colleagues found that peers, parents and siblings had independent effects on the smoking behavior of a focal sibling. To be sure, additional research needs to be conducted to gain a more thorough understanding of sibling effects alone; however, there is also a need to understand the larger social environment in which sibling relationships operate.

To illustrate this point, over a decade ago my colleagues and I (Melby et al. 1993) examined the empirical credibility of a theoretical model that postulated specific connections among parental, peer, sibling and a focal adolescent's (n = 204) tobacco use. The results of that study showed that parent and sibling smoking were related to the acquisition of peers who also smoked by the focal adolescent. The findings indicated that parent and sibling smoking in a sense ‘granted permission’ to the focal adolescent to associate with tobacco-using peers, who had a direct influence on the focal adolescent's smoking. Interestingly, however, the child-rearing style of parents also affected association with peers who smoked, suggesting that the substance-using behaviors of parents were only part of the story regarding family influence. Other reports in this program of research have shown that parenting style moderates sibling influences on alcohol use and problems with alcohol (Conger et al. 1994) and that sibling effects may be indirect through peer associations in affecting change in drinking behaviors (Conger & Rueter 1996). These findings suggest that future research should look not only at the internal dynamics of sibling relationships to understand their influence, as proposed by Slomkowski and her collaborators, but also at the broader social context of sibling relationships and how it shapes and is shaped by the sibling bond. Such research will take full advantage of the important contributions of the current study.