A Multivariate Perspective for Aptitude-Achievement Discrepancy in Learning Disability Assessment

A multivariate model is proposed for assessing a severe discrepancy between intellectual ability and actual achievement in learning disabled (LD) children. The model follows work conducted by Cronbach, Gleser, Nanda, and Rajaratnam (1972). First, a universe score is obtained for aptitude and achievement using multiple regression. The difference between these universe scores is examined for a severe discrepancy. Formulas for computing universe difference score variance, the standard error of the difference score, and its generalizability coefficient are given.