Abstract
Objectives
The study investigates the psychometric properties of the Depression, Anxiety, and Stress Scales (DASS‐21).
Method
The DASS‐21 was administered to a community sample twice over the course of 7 months ( = 235 and = 171; 51% female, M
age = 22 years). A partial credit Rasch model was used to determine the scales’ longitudinal invariance and reliability.
Results
The DASS‐21 mostly demonstrated longitudinal invariance across administrations. The Depression and Anxiety subscales were sufficiently reliable to distinguish among the suggested severity categories, but the Stress scale was not.
Conclusion
Several revisions of the DASS are recommended. It is also suggested that the Stress scale should be used with caution, given its low reliability.