A technique for identifying the subgroup membership of certain misarticulating children

Cluster analysis was used to identify two homogeneous clusters of 8-9 1/2-year-old children who misarticulated /s/, /r/, or both. The analysis was based on the children's scores on 40 measures of language, reading, auditory processing, and other variables. Discriminant function analysis was then used to identify a subset of five measures and a means of computing classification scores. These measures and the classification scores can be used to identify the cluster membership of new subjects. The use of classification scores for identifying cluster membership was cross-validated against cluster analysis of a second group of children. The two clusters are described in terms of their performance on language and reading measures.