Revisiting the structure of Diagnostic and Statistical Manual of Mental Disorders, fifth edition, Section II personality disorder criteria using individual participant data meta-analysis.

Abstract

The factor structure of personality disorder (PD) criteria has long been debated, but due to previous heterogeneous findings, a common structure to represent covariation among the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV)/DSM-5 Section II PD criteria remains an open question. This study integrated individual participant data from 25 samples (N = 30,545) to conduct factor analyses of PD criteria. Measurement invariance tests across gender, clinical status, and assessment method indicated substantial structural differences between interview-based and self-report measures. In interviews, a confirmatory 10-factor model with factors representing specific DSM-5 PDs showed a major misfit, with results from exploratory factor analyses suggesting that this was due to a relatively small number of substantial secondary loadings. In self-reports, a confirmatory 10-factor model showed greater misfit than in interviews, and exploratory solutions were more complex. When five factors were extracted, the factors showed some similarity to maladaptive trait domains such as Negative Affectivity and Disinhibition, but there were substantial differences in factor content between interviews and self-reports. In bifactor models, a general factor explained more common variance in self-reports, whereas the content of general factors was similar in both assessment methods. Our findings suggest that interview and self-report measures of PD criteria are not structurally equivalent. To advance research on the structure of PD, it may be useful to consequently focus on the shared variance of multiple methods. For this purpose, future multimethod studies should combine interviews and self-reports with other assessment methods such as informant reports. (PsycInfo Database Record (c) 2025 APA, all rights reserved).