Abstract:
This study conducted Rasch analysis of the Depression, Anxiety and Stress Scale using responses of 400 participants to produce ordinal-to-interval conversion algorithms not available to date. Minor modifications such as removing the non-fitting item 5 and combining subscales items into super-items were necessary to achieve the best model fit ( χ2(15) = 20.28, p = 0.16). The results support internal validity of the modified 20-item Depression, Anxiety and Stress Scale and include ordinal-to-interval conversion tables for use with non-clinical populations that improve reliability and precision of the measure.
