Can adolescents differentiate between depression, anxiety and stress? Testing competing models of the Depression Anxiety Stress Scales (DASS-21)

Abstract

The present study used the Exploratory Structural Equation Modeling (ESEM) procedure and a bifactor-ESEM framework to evaluate the construct validity of the Depression Anxiety Stress Scales (DASS-21) among adolescents. We also evaluated measurement invariance and latent mean differences across gender and age, and investigated the convergent validity of DASS-21 by examining relationships with positive and negative affect. A total of 1906 adolescents (60.7% females, Mage = 16.54) completed the DASS-21, whereas 998 adolescents (62.6% females; Mage = 16.98) completed both the DASS-21 and a measure of positive and negative affect. The results showed that the bifactor-ESEM model was the best representation of the data and this model proved to be invariant across gender and age. The findings indicated a strong general factor of emotional distress underlying responses to all DASS-21 items, but also suggested that Depression and Anxiety subscales possess substantial amount of specificity over and above the general factor. The Stress subscale showed little specificity after partialling out the general factor. Depression subscale explained variance of both positive and negative affect over and above the variance already explained by the general factor, whereas the Anxiety and Stress subscales did not. Our findings show that a bifactor-ESEM framework is a valuable tool for examining structural validity of DASS-21.