| The purpose of the present study was to examine eight language measures commonly used by speech-language pathologists in Southern Ontario. The developmental sensitivity of each measure, the relationships between the measures, and the ability of the measures to discriminate between children with normal language development and children with impaired language development were investigated. A total of 85 children ranging in age from 36 to 71 months were included in the study. Of the subjects, 67 had normal language development and 19 were language impaired. Each subject was administered the eight language measures in random order. Correlations, principal components analysis, and discriminant function analysis were used to examine test raw scores and test z scores. Test raw scores were found to be correlated moderately to weakly with chronological age. The magnitude of the correlations decreased as age increased, indicating that the tests were most sensitive to developmental variation in younger ages. Further analysis of correlations between tests revealed strong to moderate relationships between the tests. These relationships were further investigated with principal components analysis, which revealed three clearly delineated components that were identified as word knowledge, grammar, and semantic content. Discriminant function analysis identified the tests that distinguished between subjects with normal language and subjects with impaired language. The identified tests for the entire sample represented all three of the components revealed by factor analysis. Tests that best discriminated among children with normal language and impaired language for each 12 month age band were also identified. The implications of the results and practical applications were discussed. |
