A large number of rating scales have been devised to assess the clinical construct of 'depression'. These scales have been universally used in research with little consideration of their content, or how they relate to accepted definitions of depressive disorder. The scales are often arbitrarily selected and used for the study on the assumption that all measure the same construct. The item analysis of a number of the most widely used depression scales reveals a variation in the areas of psychopathology they cover; some scales place greater emphasis upon the assessment of anxiety than upon depressed mood. Since disturbance in neurobiological systems is manifest by specific aspects of affective and behavioural malfunction, and since psychodynamic factors lead to particular cognitive sets, the advancement of research will depend upon the construction and validation of more refined measures than are provided by the present approach.