Gender-based differential item functioning in the Cannabis-Associated Problems Questionnaire: A replication and extension

Abstract

BackgroundGender bias in measures of cannabis problems may differentially affect how men and women endorse items. This gender invariance might mask, exaggerate, or otherwise obscure true distinctions in experiences of cannabis consequences. MethodsThe Cannabis-Associated Problems Questionnaire (CAPQ), a measure of interpersonal deficits, occupational impairment, psychological issues, and physical side effects related to cannabis use, contained items with gender-based differential item functioning (DIF) in previous work—a finding we aim to replicate and extend (Lavender, Looby, & Earleywine, 2008). ResultsIn a sample of 4053 cannabis users, gender differences were apparent in global scores on the CAPQ. A DIF analysis revealed two gender-biased items, including one identified previously. Removal of these items did not significantly alter the scale’s relation to cannabis use. Gender differences on the CAPQ persisted after removal of the two problematic items, indicating true gender differences still exist in men and women’s experiences of cannabis-related consequences. Gender appeared to significantly contribute to scores on the full CAPQ and the short-form of the CAPQ with biased item removed, even after controlling for indices of cannabis use. ConclusionsThese findings suggest that the CAPQ evidences less gender bias than previously thought, perhaps due to diminishing gender-based stereotypes. Future work might opt to use the short form of the CAPQ to minimize gender-based DIF. In addition, potential biases in measures of substance use problems deserve more attention.