whether the number of minor alleles at each of the SNPs was differentially associated with the probabilities of membership for any two trajectories. As such, we found evidence that CHRM2 genotypes were significantly associated with such differences when contrasting the decreasing low externalizing trajectory with both the decreasing moderate (across four of the nine SNPs; odds ratios ranging from 1.8 to 2.5) and stable high (across five of the nine SNPs; odds ratios ranging from 2.1 to 3.2) patterns of development in externalizing behavior. Even after an overly conservative Bonferroni correction for nine separate tests of the main effect, significance was maintained for three SNPs in relation to the first contrast and two SNPs in relation to the second contrast. Moreover, post-hoc analyses incorporating sex as a covariate revealed the same pattern of associations.
