Regression Discontinuity Designs Within a Latent Variable Framework: Measurement Considerations for the Accuracy of the Treatment Effect

• In Event: AERA Graduate Student Research-in-Progress Session 3
  In Roundtable Session: Measurement and Modeling in Education Research (Table 5)

Abstract

Building on Soland et al. (2022), the study investigates the interplay of indicator reliability, number of indicators, and sample size on Type I error rates, power, and bias when the outcome of a regression discontinuity design is modeled within a latent variable framework. Preliminary findings reveal the nuanced effects of the examined factors on the dependent variables of interest. This research aims to enhance measurement practices, and therefore, provide valuable insights for improving inferential causal validity in RDDs when the variables are modeled within a latent variable framework.

Author

• Sarah Hammami, UNL