This article adresses the following issues: (1) can the GHQ be used as a severity measure for psychopathology. In other words, to what extent does it correctly follow changes in psychopathology, (2) does the alternative scoring of the GHQ, proposed by Goodchild and Duncan-Jones lead to superior results.

In a longitudinal design, 175 new psychiatric outpatients were administered the GHQ-28 and the Present State Examination on three occasions. The GHQ-28 is a self-report questionnaire that records psychopathology, the PSE is a structured psychiatric interview, that gives an accurate report of the patient's psychiatric state. The first administration took place shortly after the intake, the second and third 4 and 10 months later.

The analyses lead to the following conclusions: after a correction for retest-effect, the GHQ-28 seems a good state measure, that correctly follows changes in severity of psychiatric state (as established by the PSE-ID). However, the Goodchild and Duncan-Jones proposed alternative scoring of the GHQ did not lead to superior results.