Examining differential item functioning and group differences across student and community samples on the Personality Inventory for the DSM-V-Short Form

Abstract

This study forms one part of a larger project of evaluating the Personality Inventory for the Diagnostic and Statistical Manual-V (DSM-V) Short Form (PID-5-SF) in the South African context. Among others, its objectives include investigating its psychometric properties and the development of preliminary norms. For the project, non-clinical adult data were collected from a community sample (n = 729) and from students (n = 629). To determine if the data can be combined into a single adult sample for further research, this initial study seeks to investigate if there are any differences across the student and community samples. No meaningful group differences would suggest that the data can be merged, while differences would suggest that the student and community data should be further analysed separately. To that end, the present study investigated: (1) Differential Item Functioning (DIF) and (2) group differences on the PID-5-SF across the student and community data. Analyses using the ‘lordif’ package in R found no meaningful DIF based on McFadden’s pseudo-R-square thresholds. In addition, no substantive group differences were observed on any of PID-5-SF’s 25 facets.
Contribution: This study shows that student data collected on the PID-5-SF can, in the present case, be considered representative of the broader community. In turn, this facilitates further ongoing work on its psychometric properties and the development of preliminary norms. In this way, it will contribute to the international literature on the PID-5-SF’s psychometric functioning and enable further applied research on personality disorders among practitioners and researchers in South Africa.