Evaluation of a typology of reading disability.

A typology of reading disability was derived from a cluster analysis of psychological and educational tests administered to 63 moderately to severely reading-disabled children and adolescents. The quality of the cluster solution was evaluated through the use of statistics designed to measure the degree to which the clusters were compact and well separated. The values of the evaluation statistics for the reading clusters were compared to the distribution of these statistics in a sample of 100 cluster solutions from comparable simulated data sets. The results of this comparison suggest that the derived reading-disability clusters represent true subtypes rather than a random partitioning of a homogeneous group of subjects. The commonalities among subjects in the same clusters appear to be such as to permit hypotheses about treatment based on cluster membership. The clusters included a language disorder with associated deficits in auditory memory and sound blending, an auditory processing disorder, a mixed auditory and visual processing disorder, and an auditory memory and visual processing disorder. This study did not find the "normal" and visual-perceptual subtypes reported in some other recent investigations.