Propensity Score Stratification Using Multilevel Models to Examine Charter School Achievement Effects

Abstract

The rapid expansion of charter schools reflects a growing interest in education reform and school turnaround efforts across the United States. However, national research on charter school performance remains inconclusive. Of particular debate is the impact of transferring from a traditional public school to a charter school on student achievement and growth. We employ propensity score stratification and multilevel models to balance key covariates between treatment and control groups of a cross-state sample of students, which provides a more complex picture of charter school achievement effects in a quasi-experimental context. We find that charter school students perform consistently better on standardized mathematic achievement measures, and some students grow faster than their traditional public school peers, although the differences are modest. Furthermore, our application of propensity score stratification using multilevel models reduces selection bias which arises from confounding variables and improves the accuracy of charter school effect estimations.

KEYWORDS:

• charter school achievement effect
• propensity score stratification
• multilevel model
• quasi-experimental design