Background Early identification of emotional distress in university students requires brief yet psychometrically sound instruments that enable efficient screening without sacrificing interpretive precision. Although the Depression, Anxiety and Stress Scales-21 (DASS-21) is widely used, its length has prompted the development of several short forms. However, direct comparative evidence to guide the selection of these versions in applied settings, particularly in Mexican populations, remains limited. Methods This study compared five versions of the DASS (21, 14, 12, 9, and 8 items) within a single cohort of Mexican university students. Confirmatory factor analyses for ordinal data were used to test three-factor, hierarchical, and bifactor models. Additional evidence included hierarchical bifactor indices (ECV, ω H, PUC), reliability, score equivalence with the DASS-21, convergent validity with coping strategies, criterion-related validity based on recent functional impairment and self-reported prior psychological diagnosis, and measurement invariance by sex for the best-performing version. Results All versions showed adequate structural performance and external associations consistent with theoretical expectations. However, bifactor analyses indicated that the general distress factor became increasingly dominant as the number of items decreased, particularly in the DASS-9 and DASS-8, reducing discrimination among the specific dimensions. In contrast, the DASS-12 showed the most favorable balance between brevity, dimensional differentiation, and external validity, and also demonstrated scalar invariance by sex. Conclusion Although all evaluated versions were useful for screening general emotional distress, the DASS-12 emerged as the most balanced short form for Mexican university students, combining operational efficiency with adequate preservation of the three-factor structure and comparability across sex.
Castillo et al. (Fri,) studied this question.