A comment on van der Vorst et al. (2010).

article on adolescent drinking, van der Vorst and colleagues (2010) made recommendations for policy and parenting that are unsupported by their own data and are, I would argue, potentially harmful. The article was presented as a refutation of the idea that parents should model moderate/mealtime drinking for their adolescent offspring (as by permitting the adolescent a few sips of an alcoholic beverage with dinner). Van der Vorst et al. (2010) concluded: [O]ur fi ndings yield substantial evidence that adolescents' alcohol use leads to more drinking over time and problem drinking, regardless of the context [emphasis added], with whom adolescents are drinking, or their age. Based on this, we think that alcohol prevention programs should recommend that parents prohibit their children from drinking at home and outside the home and make parents more aware of their possible preventive role in the initiation phase of drinking. (p. 113) The authors' claim that their conclusions apply " regardless of the context " is unwarranted because their assessment of " drinking with parents " did not distinguish among contexts at all. Moderate/mealtime drinking (assuming that some of it occurred) was lumped together with drinking under unspecifi ed circumstances and in unlimited amounts—as long as a parent was present. The unsurprising fi nding was that " the more the older adolescents drank at home, the more they drank outside the home " (p. 109). Of course, unlimited provision of alcohol to one's adolescent offspring has never been seriously recommended as a way to inculcate social learning. In addition, to my knowledge, no one has suggested that the amount of drinking an adolescent does at home will be monotonically and inversely related to the amount of drinking the adolescent does elsewhere—yet that appears to be what is addressed in the analysis. The question of interest to both science and public policy is whether appropriately supervised adolescents drink differently from other adolescents. This brings us to the authors' conclusion that parental " supervision " of adolescents' drinking did not deter an overall increase in drinking. The operational defi nition of " supervision " here was whether parents cross-examined the adolescents before and after a night out with friends—not whether the parents provided alcohol for moderate, directly monitored consumption at family meals or gatherings. Thus, the authors' broad conclusions about " supervision " do not mean what they might appear to mean. What is especially frustrating is that …