Abstract
Background: The Depression Anxiety Stress Scale‐21 (DASS‐21) has been widely studied in different populations as a brief screening tool to assess features of psychological distress. Because of its multidimensionality, this scale is better addressed within a complex system.Objective: Therefore, this research is aimed at investigating the psychometric properties of the scale using network analysis and other statistical analysis to evidence validity of the scale based on its internal structure, since only two studies using classical methods have previously tested it on Brazilian samples.Methods: In this cross‐sectional study, an online survey was used to recruit a sample of 4017 (Mage = 31.16; SD = 10.31 years), of whom 70.3% were college students, 69.5% were assigned female at birth, 60.8% declared themselves to be White, and 22.6% reported they lived in the southeastern region of the country. The DASS‐21 was analyzed using uni‐, bi‐, and multivariate analyses, including exploratory graph analysis (EGA) and bootstrap EGA.Results: EGA and BootEGA generated and confirm a model with four factors, which expressed a better fit than other replications (χ2 (183) = 899.09, Comparative Fit Index [CFI] = 1.00, Tucker–Lewis Index [TLI] = 1.00, root mean square error of approximation [RMSEA] = 0.03, and standardized root mean square residual [SRMR] = 0.03). The new four‐factor model has excellent composite reliability (rho_C range = 0.84–0.90), which indicates a high level of accuracy of the measured factor. It is also invariant across genders and type of activity performed.Conclusion: The classic structure of the DASS‐21 (three factors) was replicated, but a four‐factor model proved to be the one with the best statistical indices for nonclinical sample, providing new insight into the internal structure of this scale. In addition, there was factorial invariance between the groups compared. Implications for the use of the new structure are discussed.
Funder
Conselho Nacional de Desenvolvimento Científico e Tecnológico