The Almost Perfect Scale in Medical Students: Model Confirmation, Measurement Invariance, and Differential Item Functioning By Gender

Abstract

This study examined the factor structure of two common perfectionism scales – the Almost Perfect Scale – Revised (APS-R) and the Short Almost Perfect Scale (SAPS) - in medical students. It was found that both two-factor models hold for them, albeit marginally for the APS-R. Measurement invariance by gender showed that while configural invariance and metric invariance hold, scalar invariance does not, indicating that the means for men and women may not be meaningfully compared by using these scales. Additionally, several items exhibited differential item functioning, most of which are in the Discrepancy scale of the APS-R. Overall, the SAPS provides better fit with fewer biased items, and therefore is likely to be a better instrument for comparing perfectionism in men and women medical students, although direct comparison of group means should still be exercised with caution.