A linear regression was performed (ESTIMATOR = MLR) in Mplus to test main and interaction effects of trauma type (nonsexual assaultive, nonassaultive, sexual assaultive) and fluid intelligence PGS on CPT performance (total errors, response time to cued Go) and frontal theta ERO measures. Each trauma type was included in the model as mutually exclusive categories and coded as follows: nonsexual assaultive trauma (no = 0, yes = 1), nonassaultive trauma (no = 0, yes = 1), and sexual assaultive trauma (no = 0, yes = 1). The fluid intelligence PGS was treated as a continuous predictor. Covariates included sex (male = 0, female = 1), age, number of CPT sessions, three ancestral principal components, and genotype array. Exploratory analyses were performed to investigate sex differences through a three‐way interaction (PGS × trauma × sex) for any significant interaction from the primary analysis. Prediction values for any significant interactions were calculated and plotted to better understand the nature of these interactions. All measures were included simultaneously in the pathway model to account for correlations observed among variables. Analyses were conducted in
