Abstract
INTRODUCTION: This study investigates sex differences in three widely used mental health instruments: the Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS) and the Hopkins Symptom Checklist (HSCL)-5, and HSCL-10. The psychometric properties and differential item functioning (DIF) of the instruments were explored using measurement invariance tests and multiple indicators multiple causes (MIMIC) models to more fully inspect potential indicator-level differences that may lead to bias. METHODS: There were two adolescent samples in this study. One sample was in grade 8 (ages range: 13-14; 49% females), and the other was in grade 11 (age range: 16-17; 45% females). RESULTS: The results indicate (1) a dissimilar contribution of some indicators to their latent construct across sexes or a mean sex difference in indicators that are not captured by the construct, and (2) that several indicators of the instruments might be biased regarding sex. DISCUSSION: This study contributes to our knowledge of the complexity of sex differences in the measurement of adolescent mental health.