Relational Variables for Predicting Child Language Development from Language Transcripts
Larry Barnes

Abstract
Analyzing child language transcripts for the purpose of identifying developmental language progress is commonly used but met with challenges including the determination of what variables to analyze. A multivariate approach to analysis was used in this research design to determine if a set of variables is more predictive than others within language samples. Transcripts from 56 English proficient speaking fourth graders were used to examine correlation between nine variables. Results established statistical correlations exist between MLU, TTR, and
Mazes. Implication on transcript analysis is discussed.

Full Text: PDF