Although there were no significant genotypic effects in our analysis of African ancestry subsample, interesting cross-phenotypic correlations between T1 and T2 emerged. In contrast to the link observed in the European ancestry subsample, T1 peer substance use was not correlated with T2 externalizing behavior in the African ancestry sample. T1 parental knowledge was, however, predictive of subsequent adolescent externalizing behavior at T2. This is consistent with the prior literature that shows that a high level of parental knowledge is a protective factor for adolescent externalizing behavior (Lac and Crano 2009; Lopez-Tamayo et al. 2016; Mason et al. 1994; Racz and McMahon 2011; Yap et al. 2017). Additionally, there was an inverse relationship between T1 adolescent externalizing behavior and T2 parental knowledge. This provides some evidence that youth behavioral problems could predict a change in parenting behavior (i.e., less parental knowledge). One explanation of this relationship could be because youth with behavioral problems are less likely to disclose details of their lives to parents. Additionally, parents may feel that they are unable to manage their high externalizing child and start to
