Latent Change Score Modeling of Developmental Relations Between Vocabulary and Reading Comprehension

The present study utilized a sample of 316 first grade students followed through fourth grade to evaluate the dynamic relations between vocabulary and reading comprehension as measured by two expressive vocabulary measures and two passage comprehension measures to determine the leading and /or lagging relations between the two variables. Using latent change score (LCS) modeling, three sets of models were fit to the repeated measurements of vocabulary knowledge and reading comprehension. In the first set of models, four competing univariate models were fit separately to the two constructs. For vocabulary, the model in which the changes-to-changes parameters were estimated was determined to be the best fitting model (Voc; χ2 (35) 50.30, p < .045; RMSEA = 0.037), and the dual change model was determined to be the best fitting model for reading comprehension (χ2 (36) = 254.56, p < .045; RMSEA = 0.037). In the second set of models, four competing bivariate models were fit jointly to the two constructs. It was determined the model with the coupling parameter (relating previous level of one variable to future changes in the other variable) estimated vocabulary to changes in reading comprehension, was the best fitting model (χ2= 362.35 df = 25, p < .001, RMSEA = 0.077). Lastly, four extensions of the bivariate DCS model were fitted to the data, and the model in which only the vocabulary changes to reading comprehension changes parameter was estimated was the best fitting model (χ2= 350.80, df = 123, p < .001, RMSEA = 0.077). These results describe a dynamic system in which yearly changes in comprehension were significantly determined by prior level of comprehension (β = -0.573) and by prior yearly changes in vocabulary (ξ = 0.434). Furthermore, there is a dynamic system such that yearly changes in vocabulary knowledge were significantly determined by previous level of vocabulary knowledge (β = -0.327). These findings suggest vocabulary instruction to be vital for future reading comprehension abilities, but further investigation of this longitudinal relationship is warranted to determine if reciprocal relations exist outside of this sample.