ANDIn two previous studies (Bannister, 1960 and1962) thought disordered schizophrenics were discriminated from normals and other psychiatric groups (including non-thought disordered schizophrenics) by forms of the repertory grid test. The primary characteristics of thought disordered schizophrenics, in terms of such tests, appear to be their failure to manifest substantial intercorrelations between constructs and their inability to maintain in a second grid the specific pattern of intercorrelations found in the first. In Construct Theory terms (Kelly, 1955;Bannister, 1962~) schizophrenics are limited to an overly loose and inconsistent subsystem for construing people, in conventional terms their ideas about people are both poorly. related and unstable,The forms of repertory grid test used in the two previous studies are not suitable for clinical purposes in that they are cumbersome and lacking in normative data. The purpose of the present study was to produce a clinically economic and adequately standardised grid test for detecting the presence of schizophrenic thought disorder.TEST PROCEDURE Each subject was tested individually and was faced with the same array of eight passport-type photographs, four of men and four of women. S was told to study these photographs as he was to be asked questions about them.He was first asked which of the people whose photographs he had examined was the most likely to be kind. The photograph selected by S as most kind was turned face down and its number (written on the back) was entered by E in a prepared booklet as ranked first for kind. S was then asked to select the person most likely to be kind from the seven remaining photographs and this was turned face down and its number noted. In this way S ranked all eight photographs from the most kind to the least kind. The photographs were then turned face up, shuffled to change their desk positions and S was asked to select the person most likely to be stupid. Again his chosen photograph was turned face down, its number noted and S selected the next most stupid and so forth. In this way S rank-ordered the eight photographs on six constructs presented in the following order: kind, stupid, selfih, sincere, mean and honest.S was then told that he would now do the test again on the same photographs, ranking for the same qualities. He was told that this was not to test his memory but that he should undertake the test as if he were doing it for the first time. After S had finished the repeat of the grid and filled in the Mill Hill Vocabulary Scale, Form 2 (Senior) Part A, testing was completed.If subjects asked what meaning they were to attach to words like kind, stupid and so forth, they were told to use the words in their own personal sense. If subjects complained that the task was difficult, they were encouraged to do their best even if + This author gratefully acknowledges the support of the Medical Research Council whose direct grants financed his contribution to the studies.