Wancata J, Alexandrowicz R, Benda N
Department of Psychiatry, University of Vienna, Austria.
Eur J Epidemiol. 2001;17(11):1001-4. doi: 10.1023/a:1020076311103.
An attempt was made to improve the validity of the 30-item General Health Questionnaire by excluding items. This exclusion was performed by using correlations of sum-scores with an external case criterion. This hierarchical approach based on subsets of items (stepwise hierarchical variable selection) resulted in a 9-item questionnaire whose discriminating performance was significantly better than that of the original version.