Few psychiatric disorders have yet been adequately validated and it is still an open issue whether there are genuine boundaries between the clinical syndromes recognized in contemporary classifications, or between these syndromes and normality. In the long run validation depends on the elucidation of aetiological processes. There are, however, a number of strategies which clinicians could use, but at present rarely do, in order to improve and validate existing classifications. Most of these involve studying populations which have been deliberately chosen to represent a broader grouping than a single diagnostic category, or even a group of related categories.