Diagnostic and classification accuracy for mildly mentally handicapped children

David E. Carter
The purpose of this study was to examine the diagnostic and classification accuracy of placement decisions for Mildly/Educably Mentally Handicapped (M/EMH) children in British Columbia. Evidence from the United States suggests that classification decisions are often made on the basis of idiosyncratic student behaviours and the subjective opinions of educators, not on the basis of empirical evidence. Although Canadian special education practice is often based on that of the U.S., no major study
more » ... f the accuracy of diagnosis and classification has been undertaken in this country. Based on a review of the literature, internationally accepted criteria for the diagnosis and classification of M/EMH students were formulated. In addition, variables that might influence the use of these criteria were identified. Elementary age students from two metropolitan Vancouver school districts who had been suspected of being M/EMH during a two-year period served as subjects (n=106). Of these 57 were classififed as M/EMH and 49 were classified as regular education. An evaluation of IQ, adaptive behaviour, reading and arithmetic achievement, maladaption, and visual and hearing acuity was performed for each subject. Preliminary data analyses permitted the formation of an achievement composite score and the pooling of subjects from the two districts. Using an internationally accepted two-factor diagnostic model, analyses were performed to investigate the classification accuracy for the sample. Cut-off criteria used with the two-factor model were adjusted to those of both the American Association for Mental Deficiency and the draft B.C. Special Education Guidelines. Where subjects could not be confirmed by the application of these models, sources of classification error were identified. Next, a series of discriminant function analyses, each representing a historical step in the development of diagnostic and classification models, were performed and the classification accuracy of each examined. Finally, a full model of all measured va [...]
doi:10.14288/1.0053701 fatcat:kfk4scr6dfbypl647wkzshes7q