and lower emotional stability that was not associated with sociability in boys. Both temperamental styles showed mediation through reduced emotional stability in girls. This supports previous literature suggesting that girls in particular may be more prone to adopt problematic alcohol use as a means of dealing with emotional distress and is consistent with the higher rates of comorbid depression and alcohol problems observed in women in adulthood (Grant et al., 2004). It is also consistent with methods of subtyping alcohol problems in which alcohol use associated with depression and anxiety is more prevalent in women (Cloninger, 1987). The association with peer characteristics in boys only was unexpected. Some of the pivotal work on the role of peer deviance in substance use has been done in males (Dishion et al., 1995), but in other work, there has been evidence for stronger genetic influences on peer substance use, and stronger associations with the adolescent's own substance use, in females (Dick et al., 2007). The latter study examined peer influences later in adolescence, so it is possible that the relative influence of peer characteristics may vary for the sexes at different developmental stages. This warrants further, more systematic, exploration in future research.
