Developmental changes in interrelationships of visible language codes, oral language codes, and reading or spelling

Abstract

Structural equation modeling (EQS, Bentler 1985) was used to analyze covariance structures incorporating a visible language factor underlying memory for a word and memory for a letter in a word, an oral language factor underlying access to phonemic and semantic codes, and a reading or spelling factor underlying letterby-letter and whole word presentation on tasks dependent on both visible and oral language codes. Competing theoretical models, one with the visible-oral language covariance set to zero and the other with this covariance freely estimated, were evaluated for two reading tasks (lexical decision or naming) and one spelling task (written reproduction) at the beginning (2nd month) and end (8th month) of first grade ( N = 42). At the beginning of first grade, allowing a covariance between the visible and oral language factors improved the fit, but at the end of first grade it did not. When the covariance between the visible language and oral language factors was set to zero at the end of first grade, the direct effect of the visible language factor was significant for only the naming task. Results provide evidence for (a) developmental differentiation of the visible language and oral language systems as children gained in reading skill, and (b) the importance at the end of first grade of the visible language factor in prelexical but not postlexical processes.