The validity and predictive utility of the Classroom Assessment Scoring System Pre-K in Head Start classrooms

Abstract

This dissertation critically examines the measurement structure and predictive validity of the Classroom Assessment Scoring System Pre-K (CLASS Pre-K) within Head Start settings, with implications for research, practice, and policy. Drawing on a comprehensive review of empirical studies, the analysis reveals that both the traditional three-factor model and the bifactor model of CLASS Pre-K exhibit poor model fit, high inter-factor correlations, and inconsistent item loadings, raising concerns about their structural validity in diverse classrooms. Additionally, CLASS Pre-K scores show inconsistent and modest associations with child outcomes, particularly in socio-emotional development. These limitations reflect deeper conceptual shortcomings in the framework, including its narrow scope and limited cultural responsiveness. In high-need environments like Head Start, quality teaching involves trauma-informed care, culturally relevant pedagogy, and family engagement—dimensions largely absent from CLASS metrics. In this dissertation, I argue that reliance on CLASS Pre-K in high-stakes accountability systems such as the Designation Renewal System may inadvertently penalize programs serving marginalized communities. To improve equity and accuracy in classroom quality assessment, this work recommends revising the CLASS framework to include culturally grounded and contextually relevant indicators. It calls for participatory, mixed-methods research and subgroup-specific validation studies to develop more responsive tools that align with the strengths and realities of diverse early learning settings.