Incorporating Latent Variable Outcomes in Value-Added Assessment

• In Event: Advances in Longitudinal Modeling

Abstract

This paper investigates the utility of incorporating latent variable modeling with value-added assessment to explore the degree to which outcome measurement assumptions influence teacher rankings. Two empirical cross-classified mixed effects models -- 1) cumulative, and 2) acute -- and four measurement model structures -- 1) summary scores, 2) factor scores, 3) direct latent variable modeling assuming longitudinal invariance, and 4) direct latent variable modeling with longitudinal noninvariance -- will be used to model multivariate student outcomes from the Early Childhood Longitudinal Study (Kindergarten Cohort). Empirical results will be used as population parameters in a subsequent Monte Carlo simulation to determine the degree to which analytic models differing on random effect and measurement structures recover known population parameters.

Authors

• Leslie R. Hawley, University of Nebraska - Lincoln

• James A. Bovaird, University of Nebraska - Lincoln