Development and Validation of Kindergarten Dynamic Assessments of Early Reading and Language.

Abstract

PURPOSE The purpose of this study was to develop and evaluate the initial reliability and validity evidence of the dynamic assessment (DA) of early reading and language as a second-stage screener in kindergarten, the first year of formal schooling. The DA comprises three subtests that capture students' ability to learn letter sounds and blending and infer word meaning from context by providing a series of graduated prompts, from the least to most helpful sequence, depending on students' responsiveness and documenting students' learning process. METHOD Concurrent data on DA and static screeners at the beginning of kindergarten (N = 164, Mage = 5.60, 4-6 years old) were analyzed using item response theory, confirmatory factor analyses, and receiver operating characteristics curve models. RESULTS Item response theory analyses identified the optimal number of prompts for each DA subtest, and they were rescaled accordingly. Confirmatory factor analyses of the rescaled DA indicated three distinct subscales comprising letter sound, blending, and vocabulary learning with high reliability coefficients and construct validity evidence demonstrating theoretically expected correlations with the static screener. Finally, when DA was added to the static screener for identifying students with disabilities, DA improved specificity by reducing the number of false positives. CONCLUSIONS Our preliminary investigation demonstrated the DA's appropriate reliability and validity as a supplemental screener. Future directions include evaluating the DA's predictive validity evidence and classification accuracy with a large student sample and norm-referenced standardized measures, addressing the limitation of the current study using school-identified disability classification as a criterion measure. SUPPLEMENTAL MATERIAL https://doi.org/10.23641/asha.29425259.