CHILDREN VULNERABLE TO LEARNING FAILURE: IDENTIFICATION AND DESCRIPTION OF CONSISTENTLY IDENTIFIABLE SUBTYPES

The purpose of this study was to identify and describe homogeneous subgroups within a sample of kindergarten children identified as vulnerable to learning failure on the basis of poor performance on SEARCH, a scanning instrument for the identification of potential learning disabilities. The sample of 141 children (85 males, 56 females) was selected from a much larger sample of approximately 950 children screened over a period of three years in the kindergartens of five schools in a large urban school district. Profiles of perceptual and cognitive functioning, which consisted of twenty standardized scores on eleven WPPSI and nine SEARCH subtests, were analyzed using hierarchical, agglomerative, average-linkage cluster analysis, with Euclidean distance as the similarity measure. The cluster analytic solution was checked by a k-means iterative partitioning procedure. Multidimensional Scaling (MDS) procedures were used to confirm the adequacy of the cluster solution, and to establish the number and consistency of dimensions necessary to conceptualize intersubject differences. HGROUP cluster analysis classified 84% of the children into one of seven homogeneous subgroups. Three distinct language disability subgroups were identified, along with a Developmental Gerstmann syndrome subgroup, a developmental immaturities subgroup, and a subgroup manifesting mixed deficits in visual-motor integration and associative learning. The seventh subgroup, which was free of perceptual or cognitive deficits, contained a significantly higher frequency of children rated as demonstrating severe personality disturbance. Significant differences among subgroups were found with respect to sex ratio, abnormal pencil grip, and presence of neurological impairment. Subgroup differences detected on WRAT reading, Woodcock Word Identification, and Woodcock Word Attack scores did not hold up when the effects of intelligence differences among subgroups were removed. Involvement of all children in an intervention program, however, confounded the attempt to measure subgroup differences in reading achievement. Application of MDS procedures to two separate samples identified four consistent dimensions and three characteristics that accounted for intersubject proximity. The first two of these dimensions, general intelligence, and area of deficit (i.e. audio-linguistic versus visuo-spatial) were accounted for by the cluster solution. The third, perceptual intactness and body image awareness, was also related to neurological impairment, and served to differentiate subjects within homogeneous subgroups. . . . (Author\u27s abstract exceeds stipulated maximum length. Discontinued here with permission of author.) UM