The data collection exercises which facilitated the development of the psychometric data is described in the Carr, Harris & James (2000) British Journal of Health Psychology paper.
Reliabilities of the DAS59
The internal consistency of the scale (Cronbach alpha) is 0.98 for the full scale, and from .98 to .74 for the factors. Test-retest reliability for the general population is 0.75, over a three month interval. Test-retest reliability for a clinical sample over a similar period is 0.86. Test- retest reliabilities for the factors ranged from. 79 (GSC) to .51 (FSC) (SSC = .70; SBSC = .73; NSC = .70)
Validities of the DAS59
Good content validity had been achieved by basing the original set of items upon patients’ descriptions of living with problems of appearance (Harris, 1982) and by ensuring that all aspects of patients’ experiences were retained throughout the development of the scale. The effectiveness of this strategy has been reflected in the consistently positive feedback we have received from patients and in the high level of internal consistency of the scale.
Using data from the earlier study to validate the experimental scale (DAS136), convergent construct validity was shown in a good positive correlation (0.63) between the DAS59 items total and the total SAD Scale score. Convergent construct validity was shown in moderate positive correlations with depression (0.52; BDI), anxiety (0.54; CCEI total), trait anxiety (0.48; STAI) and neuroticism (0.51; EPQ). Divergent construct validity is shown in the negative correlation with extraversion ( -0.21; EPQ) and the effective independence of hysteria (0.12; CCEI hysteria).