Measurement invariance of short-form measures of the Dark Triad across gender

Abstract

Given the prevalent use of short-form measures of the Dark Triad (DT) personality traits, much research attention has been devoted to examining the extent of psychometric equivalence of these scales across groups using measurement invariance (MI) analysis. However, an important limitation of MI analysis is that it provides limited information regarding the practical implications of non-invariance for group comparisons. The current study addressed this issue by supplementing MI analysis with additional assessments of practical significance of non-invariance. Specifically, in addition to examining gender invariance across two samples who completed the Dirty Dozen (DD) and the Short Dark Triad (SD3), we calculated effect size measures of measurement non-invariance that indicate the magnitude of non-invariance and the extent to which the observed mean gender differences are attributed to measurement non-invariance vs. true gender mean differences. The results provided support for the inference that observed mean gender differences on the DD and SD3 are largely construct-driven. Nonetheless, a few items on the Machiavellianism (SD3) and psychopathy (DD and SD3) scales showed gender non-invariance that reached the benchmark for medium effect. The results illustrate the advantage of supplementing MI analysis with standardized effect size indices that inform the practical implications of measurement non-invariance.