The main effects model (Fig. 1) explained 20% and 21% of the variance in the INT and EXT dimensions, respectively. The full model that included main and interaction effects (9/11 exposure*childhood adversities) explained 21% and 22% of the variance in the INT and EXT dimensions, respectively. The full model provided an excellent fit to the data (CFI = 0.996, TLI = 0.994, RMSEA = 0.003). INT and EXT were correlated (r = 0.58). Participants had higher INT and EXT levels if they were exposed to 9/11 (EXT: β = 0.10, p < 0.001; INT: β = 0.11, p < 0.001). These results did not change significantly after adjusting for childhood maltreatment and parental psychopathology. Further, modification indices did not suggest that any associations between 9/11 exposure and individual disorders were unmediated by the association with the disorder with the latent dimensions (no difference in model fit exceeded a priori significance threshold of greater than 3.96), with the exception of PTSD. According to these criteria, there was significant model fit improvement when a direct effect of 9/11 exposure on the individual diagnosis of PTSD was included in the model (Modification = 5.66, p < 0.05).
