Cross-sectional psychometric validation, convergent validity, and measurement invariance of the DASS-21 in Mexican university students

Abstract

Objective: To evaluate the factor structure, reliability, and gender invariance of the DASS-21, and to examine convergent/criterion-related evidence with coping and functional impairment. Design: Cross-sectional, single-site psychometric validation in first-year undergraduates at a private Mexican university; multigroup CFA with ordinal/WLSMV; proctored digital administration; convergent/criterion analyses and known-groups tests. Methods: First-year students (N=1,251) completed the DASS-21 and an adapted Proactive Coping Inventory. Models estimated from polychoric correlations (WLSMV) compared a correlated three-factor solution, a second-order model, and a bifactor model; ESEM served as a robustness check. Gender invariance (women/men) followed a configural–metric–scalar (threshold) sequence. Associations with coping, prior diagnosis, and recent functional impairment were tested with multiplicity control. Results: The three-factor model showed acceptable fit (CFI=0.99; RMSEA=0.07). Hierarchical/bifactor solutions improved representation; bifactor indices (ECV=0.81; ωh=0.88) supported a predominant general factor. Reliability was good–excellent (ωsubscales≈0.85–0.88; ωtotal≈0.94). Configural, metric, and scalar (threshold) invariance held (ΔCFI≤0.01; ΔRMSEA≤0.015). Latent means indicated higher Stress and Anxiety in women (small effects), with no difference in Depression. DASS-21 scores correlated negatively with adaptive coping and positively with avoidant coping; higher scores were observed among students reporting prior diagnosis and recent functional impairment. Conclusion: Evidence supports the validity, reliability, and fairness of the DASS-21 for screening general distress and profiling subdomains in Mexican universities; priorities include norms and ROC-based cutoffs in verified clinical samples. Limitations: Single-site, non-probabilistic sample; cross-sectional design; clinical status by self-report; forced-completion digital setting may introduce minor response pressure.