Use of Multiple Discriminant Function Analysis in Evaluation of a State-Wide System for Identification of Educationally Handicapped Children

Publisher Website
Google Scholar
Abstract
The purpose of this study was to evaluate a state-wide system for the identification of educationally handicapped children through the use of multiple discriminant function analysis. The assessment information upon which diagnostic specialists based their classification decisions was used concurrently to predict the classification decisions which the diagnosticians made for 477 children representing school districts from throughout the State of Utah. The rationale for this procedure was (1) to manifest the variables along with their relative weights upon which the diagnostic classifications were based, (2) to gain evidence of consistency with which diagnostic standards were followed, and (3) to examine how accurately the diagnostic classifications made in the State of Utah could be duplicated statistically. The results indicated that the “mentally retarded,” “culturally disadvantaged,” “slow learner,” and “no-significant-problem” groups could be efficiently identified statistically, but the distinctions between those groups were based almost entirely on the variables of Full Scale IQ and race. The other two classifications of “learning disabled” and “emotionally handicapped” could not be efficiently identified statistically, and the consistency of standards employed for those classifications was questioned. Implications for current practice are discussed and directions for future research are suggested.

This publication has 4 references indexed in Scilit: