The graph of the aggression data suggests that MAOA-L was associated with increased aggression scores for children with moderate trauma histories, but that MAOA genotype had little effect for children with extreme trauma experiences (Figure 1). To test this hypothesis, an exploratory GEE analyses was conducted examining the effect of trauma exposure and genotype for children with low to moderate trauma histories (e.g., scores 0–2) and proposing no effect of genotype for children with more severe trauma exposure (e.g., scores ≥ 31). In the moderate trauma range, 25% of the children had a history of verified maltreatment and removal from home. The analysis used the whole sample to estimate the model. For children with trauma scores below the threshold (TIES < 3), the main effects and interaction term were estimated. When an individual child’s trauma score exceeded the threshold, the model converted the effect of MAOA genotype and the G×E interaction term to zero.
