An examination of early reading skill covariance structure invariance across Spanish-speaking English language learners in grades 3 and 4 and their native English-speaking peers.

摘要

A long history of research on the construct of early reading exists for students who have grown up speaking English. Much less research, however, exists on the ways in which this construct might differ for students from different language backgrounds, the focus of this study. This dissertation examines the construct validity of early reading assessments across samples of Spanish-speaking third- and fourth-grade English Language Learner (ELL) students with different levels of English proficiency and their native English-speaking peers using confirmatory factor analysis (CFA) and a structural equation modeling (SEM) approach. Invariance was tested across four parameters: factor loadings of measured variables on the latent variables of alphabetic principle, fluency, and comprehension; variances of latent variables; covariances between latent variables; and variances of error terms associated with each measured variable. Students were grouped into three levels of English proficiency: Low ELL ( n =110), High ELL (n = 110) and Native English Speaker (n = 214).; Results indicate the measurement model was not invariant across the three groups. The complete model was unable to be estimated for the High ELL group due to difficulties associated with the latent variable Alphabetic Principle. The complete model ran for both the Low ELL and Native Speaker groups. The complete model was not invariant across these two groups in any of the parameters.; A truncated model examining only the parameters associated with Fluency and Comprehension ran for all three groups. Results from the truncated model suggest invariance of these two constructs across High ELL and Native Speaker groups on all but one parameter (the error term variance associated with Literal Comprehension) and across all three groups across 8 parameters (the variances of the latent variables fluency and comprehension, the factor loadings for the three measures of comprehension, the covariance between fluency and comprehension, and the error terms associated with Passage Reading Fluency and Evaluative Comprehension). Appendices include example measures used to assess students' skill in alphabetic principle, fluency, and comprehension.