Modeling Relations Between Reading and Writing

Abstract

The three chapters take different approaches to model relations between reading and writing: Synthesizing existing models of reading and writing and key empirical findings (Kim); using structural equation modeling (SEM) to test alternative models of relations between reading and writing (Jimenez et al.); and using meta-analytic structural equation modeling (Ahmed &Wagner). Each of the models can be used to generate testable predictions. Examples of how this can be done using latent change score modeling of longitudinal data are described. Latent change score modeling is a new approach to modeling longitudinal data that combines the strengths of SEM causal modeling and latent growth curve modeling.