Premium
Why Does the Repertory Grid Test Indicate Thought Disorder?
Author(s) -
FRITH C. D.,
LILLIE F. J.
Publication year - 1972
Publication title -
british journal of social and clinical psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.479
H-Index - 92
eISSN - 2044-8260
pISSN - 0007-1293
DOI - 10.1111/j.2044-8260.1972.tb00781.x
Subject(s) - consistency (knowledge bases) , psychology , test (biology) , construct (python library) , reliability (semiconductor) , repertory grid , grid , construct validity , psychometrics , clinical psychology , cognitive psychology , social psychology , artificial intelligence , computer science , mathematics , paleontology , power (physics) , physics , geometry , quantum mechanics , biology , programming language
Low scores on the Bannister & Fransella (1967) Grid Test of Schizophrenic Thought Disorder need not necessarily be the result of a loose construct system. For example, a low discriminability of the elements would cause subjects to make ‘errors’ in assigning ranks to elements giving low Intensity and Consistency scores. A direct estimate of the extent of these ‘errors’ is given by the Element Consistency score which is equivalent to test‐retest reliability. A retrospective study of the performance of 95 psychiatric patients on the Grid Test was carried out. There was a significant relation between test performance and psychiatric assessment of thought disorder. However, Element Consistency scores also successfully predicted psychiatric assessments of thought disorder and Intensity and Consistency scores. When Intensity and Consistency scores uncontaminated by ‘errors’ in performance were estimated by partialling out Element Consistency they showed no relationship to psychiatric assessments of thought disorder. It is therefore proposed that the poor performance of thought‐disordered patients on the Grid Test need not be due to a loose construct system, but could be due to a difficulty in extracting relevant information from complex visual stimuli which results in a low discriminability of the elements presented in the Grid Test.