Comparison of Predicted-difference, Simple-difference, and Premorbid-estimation methodologies for evaluating IQ and memory score discrepancies.

Discrepancies between WAIS-III and WMS-III scores for a group of 39 males and 48 females with a history of TBI were examined using three methodologies: Predicted-difference, Simple-difference, and Premorbid-estimation methods. Overall, the Predicted-difference method tended to classify the fewest individuals as impaired based on statistical rarity of discrepancies (11-16% classified as impaired), while the regression-based Premorbid method tended to classify the fewest individuals as impaired based on clinical rarity of discrepancies (4-8% classified as impaired). Degree of agreement is reported and was substantial. The only comparison between methods to reach statistical significance was the Predicted-difference method classifying subjects as impaired at a higher rate than other methods for Auditory Delayed memory index (Cochran's Q = 7.00, P < .05). Findings suggest a combination of estimates of premorbid functioning and regression-based predicted scores is optimal for interpreting IQ/memory score discrepancies. Clinical implications are discussed.