Assessment of Brain Damage: Discriminant Validity of a Neuropsychological Key Approach with the McCarron-Dial System
Description: The present study investigates the predictive accuracy of a key approach to interpretation of the verbal-spatialcognitive (VSC) and sensorimotor (SM) factors of the McCarron-Dial System (MDS). The subjects include 99 brain damaged and 30 normal adults. The following research questions are addressed: (a) Does the neuropsychological key classify brain damaged and non-brain damaged subjects at a level significantly above chance? (b) Among the brain damaged subjects, does the neuropsychological key identify right brain damage, left brain damage and diffuse brain damage at an accuracy level significantly above chance? (c) Is the neuropsychological key approach superior to the empirical model derived from discriminant function analysis in predictive accuracy? The neuropsychological key correctly classifies 90% of the cases as brain damaged and 90% of the cases as non-brain damaged, for a total of 89.9% predictive accuracy. The obtained Kappa coefficient of .74 is statistically significant. The key accurately classifies 71.4% of the brain damaged group as right damage, 70% as left damage, and 93.8% as diffuse damage, for a total predictive accuracy of 7 9.5%. The Kappa coefficient of .68 is statistically significant. Chi square analysis of the difference between the key approach and multiple discriminant function analysis reveals that no significant difference is present between the accuracy of the two approaches in differentiating between brain damaged and non-brain damaged, or in differentiating among left, right and diffuse brain damage. The results support the validity of a neuropsychological key approach to interpretation of the McCarron-Dial System, although cross-validation is indicated to confirm the stability of these results. Differences in sex, educational level and racial composition of the comparison groups may have affected the results obtained. Refinement of the key in future research and the addition of test instruments assessing memory, auditory processing, attention and emotional/behavioral variables are recommended.
Date: December 1985
Creator: Norton, Carole Lynn