Limitations on the use of scales in psychiatric research.

This paper discusses some issues involved in the construction and interpretation of scales in psychiatric research. The issues covered include scale construction, measures of scale homogeneity, the logic of factor analysis and the interpretation of effects manifesting in scale-data, especially interaction effects. The aim of the paper is to "de-mystify" some of the psychometric techniques commonly available, and to emphasize that their appropriate use depends on both a firm understanding of the content area to which they are applied and a rudimentary knowledge of what the methods purport to do.