There are also several limitations to acknowledge in the present study. First, the size of the sample precluded testing several potentially important hypotheses. With respect to sex, for example, it has been suggested that the drinking behaviors of adolescent girls are more susceptible to the influences of peers than are those of boys (Dick et al., 2007b). Moreover, males and females have been shown to differ markedly in their prevalences of individual externalizing disorders (Kessler et al., 1994; Newman et al., 1996; Moffitt et al., 2001) and general externalizing problems (Hicks et al., 2007). Although genotype was not differentially associated with sex in this study, girls were clearly overrepresented in the decreasing low externalizing trajectory. Replication within a larger sample would allow us to incorporate sex as a moderator or to examine relevant influences in relation to sex-specific patterns of development. In addition, a larger sample may provide the power necessary to model development separately within the distinct domains of aggression and delinquency.
