Abstract
This paper reviews the current understanding of the statistical and psychometric issues associated with the severe discrepancy definition of learning disabilities used in PL 94-142. Statistical issues involve model specification, selection of variables, error rate control, and multivariate vs. univariate analysis. Psychometric issues involve appropriate use of reliability and validity information and inadequate conceptualization of the discrepancy construct with respect to concurrent, predictive, and construct validity. In conclusion it is proposed that the statistical and psychometric issues should not drive the psychological theory of learning disabilities, and psychologists should regain control of these issues.
Get full access to this article
View all access options for this article.
