THE RORSCHACH AND CENTRAL NERVOUS SYSTEM PATHOLOGY: A CROSS-VALIDATION STUDY

Four Rorschach systems for determining the presence or absence of brain pathology are compared using a sample of 118 patients. Representativeness of the sample was secured by selecting only patients admitted to a neurology ward with complaints, signs, and/or symptoms referrable to the nervous system. The sample was divided, on the basis of the combined judgments of 2 neurologists, into cases presumably having brain pathology above the foramen magnum (84 cases) and those presumably having no such pathology (34 cases). The patients in both groups had undergone thorough and equivalent neurologic investigation. Results of the comparison of the 4 Rorschach systems with this criterion indicate that 3 of the 4 systems can distinguish between persons with and those without brain pathology with better than chance accuracy. The Piotrowski and Hughes systems are highly accurate (94% and 91%, respectively) when they identify a case as "organic" (only 6% and 9% false positives). However, when these systems identify a case as "nonorganic" the finding is inconclusive in that they fail to identify 62% of the organic cases (false negatives). The Ross and Ross system is slightly less accurate. The Dörken and Kral system yields results that can be explained in terms of chance and shows a systematic tendency to identify too many cases as "organics," resulting in 76% false positives. Only l of the 4 systems, Piotrowski's, maintains a validity level comparable to that obtained in previous studies.

The results are discussed and recommendations for derivation of more efficient Rorschach diagnostic systems advanced.