A Multilevel Latent Variable Modeling Approach to the Regression Discontinuity Design

• In Event: AERA Poster Session 6
  In Poster Session: Division D Section 1 Poster Session 2

Abstract

Regression discontinuity designs allow estimating local average treatment effects when assignment of an individual to treatment is determined by their location on an assignment variable. The design is especially useful in education research, where ethical concerns can forestall the use of randomization. The current study proposes a multilevel latent regression discontinuity model, which reflects two characteristics typical in application of the design in education research: The prevalent use of latent constructs and the hierarchical structure of data due to cluster sampling. A simulation study shows the proposed model recover model parameters well and provides less biased treatment estimates than the conventional modeling approach. This study provides a more rigorous modeling framework than the conventional approach in many education studies

Authors

• Monica Morell, University of Maryland - College Park

• Ji Seung Yang, University of Maryland