Prevalence of examinees screening positive on PID-5 validity scales and impact on scale scores, correlates, and reliability

Année :
Référence

*Leclerc P, Le Corff Y, Lapalme M, Bégin V, Forget K, Sellbom M. (2026). Prevalence of examinees screening positive on PID-5 validity scales and impact on scale scores, correlates, and reliability. Personality Disorders: Theory, Research, and Treatment.

Résumé

The Diagnostic and Statistical Manual of Mental Disorders’ (5th ed.; DSM-5) official self-report instrument to assess personality disorders is the Personality Inventory for DSM-5 (PID-5). To screen for response bias, three validity scales have been constructed from the PID-5 item pool to assess, respectively, inconsistency, overreporting, and underreporting. Nonetheless, the prevalence and impact of PID-5 response bias on diverse psychometric properties are largely underdocumented. Thus, this study aims at: (a) identifying the rates of positive screening of each validity scale, (b) identifying the degree of overlap between validity scales, and (c) documenting the impact of each bias on scale scores, criterion validity (correlations with self- and informant-assessed personality pathology), and internal consistency. Preregistered analyses were conducted among a large population-representative sample (n = 2,505, 51.2% females) and an at-risk sample of young adults (n = 321, 48.3% females). For each sample, respectively, plausible estimates were obtained for inconsistency (4.7% and 7.2%) and overreporting (2.9% and 5.6%), whereas the underreporting scale likely led to an overestimation of the prevalence (38.6% and 32.4%). Screening positive on two PID-5 validity scales was rare (<2% of cases). Each bias had distinctive impacts on scores, correlates, and reliability. Screening for multiple biases is paramount to replicating research findings and ensuring sound clinical conclusions. Recalibrating or developing a new PID-5 validity scale to detect underreporting among populations with low to medium levels of psychopathology, as well as a scale for acquiescent/counteracquiescent bias, would be warranted.

DOI