A confirmatory factor analysis of the Hospital Anxiety and Depression scale: Comparing empirically and theoretically derived structures
Article first published online: 24 DEC 2010
2000 The British Psychological Society
British Journal of Clinical Psychology
Volume 39, Issue 1, pages 79–94, March 2000
How to Cite
Dunbar, M., Ford, G., Hunt, K. and Der, G. (2000), A confirmatory factor analysis of the Hospital Anxiety and Depression scale: Comparing empirically and theoretically derived structures. British Journal of Clinical Psychology, 39: 79–94. doi: 10.1348/014466500163121
- Issue published online: 24 DEC 2010
- Article first published online: 24 DEC 2010
- Received 23 February 1998; revised version received 13 July 1999
- Cited By
Objectives. To compare the fit of various factor solutions for the Hospital Anxiety and Depression scale (HAD; Zigmond & Snaith, 1983).
Design. A cross-sectional postal survey was used to collect the data from community-based participants in the West of Scotland Twenty-07 study.
Methods. The HAD scale, a 14-item self-administered measure of anxiety and depression, was completed by 2547 participants from three age cohorts (aged approximately 18, 39 and 58 years). Using confirmatory factor analyses four models suggested by prior exploratory factor analyses were compared to a model derived from Clark and Watson's (1991) tripartite theory of anxiety and depression.
Results. The model derived from the tripartite theory of anxiety and depression (with factors labelled negative affectivity, anhedonic depression and autonomic anxiety) produced the closest fit to the data. This model produced a good fit in all three cohorts although group comparisons suggested that there were variations in the strength of some factor loadings across the three age groups. A model that had a hierarchical arrangement of the three factors in the tripartite model was also produced. This model fit the data equally as well as did the ‘flat’ tripartite model.
Conclusions. Three factors appear to underlie the HAD scale. Research is needed that examines whether or not using sub-scales based on these factors increases the ability of the HAD scale to detect cases of anxiety and depression.