Used the 16 PF in a group screening battery for psychiatric inpatients. Validation was carried out using psychiatric diagnoses as the criterion. Four hundred and sixty-three patients were grouped into eight traditional diagnostic groups. Discriminant analysis was used to predict group membership and multivariate profile analysis was used to compared group means. No differences in profile elevation or shape were found, and four discriminant functions could classify correctly only 22% of the patients. Results indicated that the 16 PF is unsuitable for discriminating among psychiatric diagnostic groups.