In the third 3-factor ESEM model (CFI = .97; TLI = .95, RMSEA = .01), ASPD was included as an indicator in the measurement part of the model, and the remaining PDs were included based on the alternative coding regarding social and/or occupational dysfunction proposed by Trull and colleagues (2010). Although these diagnostic rules yield lower prevalence rates of PDs, there are similarities with the estimates shown in Table 1. As shown in Table 3, the regression coefficients of avoidant and obsessive-compulsive PDs are significantly higher for factor 2 compared with factor 1. The regression coefficient of schizotypal PD is significant for factor 2 but not for factor 3, and the regression coefficient of borderline PD is significantly higher for factor 2 compared with factor 3. The alternative coding, however, yields some differences with the ESEM model in Table 1. The regression coefficients of paranoid and avoidant PDs are significantly related to each factor but no longer significant for the associations between schizoid PD and factors 1 and 3; schizotypal PD and factor 1; histrionic PD and factors 2 and 3; and narcissistic PD and factors 1 and 2.
