Differentiating Between Learning Disability and Emotional Impairment with Adolescents (Diagnosis, Disturbed).

Abstract

The problem investigated in this study was that of making a differential diagnosis between learning disability and emotional impairment with adolescents in the school setting. Intellectual, achievement, self esteem, and behavioral variables were obtained for 145 adolescents with a mean age of 196 months, certified either learning disabled or emotionally impaired. Using a stepwise discriminant analysis, nine variables formed the set of optimal discriminating variables. These included four factors from the Devereux Adolescent Behavior Rating Scale; the Comprehension, Block Design, and Arithmetic subtests from the WISC-R; and math and spelling achievement test scores. A jackknifed classification demonstrated that 65.5 percent of the subjects were correctly classified using this set of optimal discriminator variables. Because a 65.5 percent classification rate is not high enough to warrant the usage of this optimal set of variables for the practical purpose of making a differential diagnosis, the conclusion was drawn that the two categories of learning disability and emotional impairment are too heterogeneous for making a differential diagnosis. Therefore, cluster analysis techniques were applied to the intellectual, achievement, self esteem, and behavioral scores of all 145 subjects. Four clusters of students resulted: those characterized by (1) learning and emotional adjustment problems, (2) low achievement, (3) significant disturbance and lowered potential, and (4) extreme underachievement and acting-out behavior. A series of stepwise discriminant analyses were performed resulting in the identification of the following set of optimal discriminating variables: the Arithmetic subtest from the WISC-R, a negative self esteem factor, and four factors from the Devereux Adolescent Behavior Rating Scale. A jackknifed classification demonstrated that 82.8 percent of the subjects were correctly classified using this set of optimal discriminator variables. The conclusions drawn included (1) that greater homogeneity is needed in the special education classification system when dealing with adolescents with learning and /or emotional problems, (2) that self esteem measures are important in the diagnostic process, (3) that there exist three to four subgroups of students within this population, and (4) that special education delivery systems must differentiate and be responsive to these four subgroups of students.