Modeling Carryover Classroom Effects: A Comparison of Two Multilevel Structural Equation Modeling Approaches

• In Event: Fancy Structural Equation Modeling and Hierarchical Linear Modeling

Abstract

Educational research data usually have complex structure, where students are cross-classified with classrooms over multiple grade-levels. Accounting for such dependency among student-level observations across multiple academic years requires use of specialized multilevel models. In addition to random within-year classroom effects, carryover effects (from previous years’ classrooms on later years’ outcomes) should be accounted for as well. This study compared two multilevel SEM-based approaches for incorporation of the so-called carryover effects. The empirical dataset was comprised of N=1,506 students’ kindergarten, first-, and second-grade math national percentile scores. Analysis results provided methodological and practical implications for educational researchers, who wish to employ such models for analyzing complex cross-classified data.

Authors

• Yusuf Kara, Southern Methodist University

• Akihito Kamata, Southern Methodist University

• Dylan Elise Farmer, Southern Methodist University

• Annie Wright, Southern Methodist University

• Scott K. Baker, Center on Teaching and Learning