[Psychiatric classification by means of a discriminatory application of Q factor analysis (author's transl)].

An attempt is made to separate different pairs of psychiatric patient groups by means of a modified form of Q factor analysis comparable to discriminant analysis. The psychopathological state of 454 patients had been rated using two psychiatric rating scales, the IMPS (Inpatient Multidimensional Psychiatric Scale) by Lorr et al. and the psychopathological and somatic scales of the AMP documentation system. Out of these patients the four most frequently occurring groups (schizophrenia, paranoid form, n = 45; schizophrenia, unspecified form, n = 47; depressive psychosis, n = 44; depressive neurosis, n = 53) were selected. Each patient group was divided randomly into two samples, an analysis sample and a validation sample. Only those items were selected which discriminated best between any two analysis samples. Using G indices a Q factor analysis was calculated, and the results improved by a criterion-related additional rotation. The resulting weights were transferred to the validation samples in order to have a cross validation. The mean percentage of correct placements within the validation samples was 83%.