Comparing the temporal stability of self-report and interview assessed personality disorder

Findings from several large-scale, longitudinal studies over the last decade have challenged the long-held assumption that personality disorders (PDs) are stable and enduring. However, the findings, including those from the Collaborative Longitudinal Personality Disorders Study (CLPS; Gunderson et al., 2000), rely primarily on results from semistructured interviews. As a result, less is known about the stability of PD scores from self-report questionnaires, which differ from interviews in important ways (e.g., source of the ratings, item development, and instrument length) that might increase temporal stability. The current study directly compared the stability of the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV) PD constructs assessed via the Schedule for Nonadaptive and Adaptive Personality (SNAP-2; Clark, Simms, Wu, & Casillas, in press) with those from the Diagnostic Interview for DSM-IV Personality Disorders (Zanarini, Frankenburg, Sickel, & Yong, 1996) over 2 years in a sample of 529 CLPS participants. Specifically, we compared dimensional and categorical representations from both measures in terms of rank-order and mean-level stability. Results indicated that the dimensional scores from the self-report questionnaire had significantly greater rank-order (mean r=.69 vs. .59) and mean-level (mean d=0.21 vs. 0.30) stability. In contrast, categorical diagnoses from the two measures evinced comparable rank-order (mean κ=.38 vs. .37) and mean-level stability (median prevalence rate decrease of 3.5% vs. 5.6%). These findings suggest the stability of PD constructs depends at least partially on the method of assessment and are discussed in the context of previous research and future conceptualizations of personality pathology.