Measurement invariance of perfectionism measures in students with and without a history of non-suicidal self-injury

Abstract

Perfectionism is a transdiagnostic process which may be implicated in the onset and maintenance of non-suicidal self-injury. No study has evaluated whether reported differences in perfectionism between individuals with and without a history of self-injury represent genuine group differences or measurement artefacts. The present study reports an investigation of the measurement invariance of two common scales of perfectionism, the Frost Multidimensional Perfectionism Scale-Brief (FMPS-Brief) and the Clinical Perfectionism Questionnaire (CPQ), among university students (Mage = 20.48, SDage = 2.22, 75.3% female, 22.8% male) with and without a history of self-injury (total n = 711). Results revealed full residual error invariance for the two-factor model of FMPS-Brief, while the bifactor model of the FMPS-Brief and the two-factor model of the CPQ demonstrated partial metric invariance. Accounting for partial metric invariance, the bifactor model of the FMPS-Brief also demonstrated partial residual error invariance. The current findings suggest that observed differences using the FMPS-Brief reflect genuine differences in perfectionism between individuals with and without a history of self-injury. Further, while researchers using the bi-factor model can have confidence that the general factor can adequately assess group differences, differential item functioning should be considered if using the strivings and concerns factors. Finally, in the current data, the CPQ did not perform as expected in baseline model fit and future research should replicate assessments of measurement invariance in this measure.