Abstract
Some relationships between language measures based on verbal output (Mean Length of Response and Number of Different Words) and those based on structural analysis (Length-Complexity Index and Developmental Sentence Types) were investigated by longitudinally analyzing the early language development of 13 children. Seven recordings were made of eight boys and five girls at two- to four-month intervals beginning at an average age of 21 months. A method for quantifying the DST was devised, and all the recordings were analyzed to compare scores for the measures. It was found that (1) variability in the rate of increase in language development scores made age a poor basis for comparing children, using any language measure; (2) there was considerable agreement among measures in reflecting the rate of language development; (3) all the language measure scores seemed to be equally related to language growth, so each provided a reasonable index of the complexity of the underlying language structure; and (4) the reliability of trained scorers was as good for measures based on structural analysis as it was for those based on verbal output.