This paper argues that the practice of validating diagnostic tests solely against psychiatric diagnosis is pointless. The reason is, that if this method of validation is used, none of the consequences of the diagnostic label necessarily follow from the test score. Tests should be directly validated for the purpose to which they are being put. In clinical psychology, tests are usually used to provide a description of the patient, to provide a prognosis, to suggest a treatment, or to suggest some aetiology. Unless tests have been directly shown to have validity for these purposes, they should not be used.