Multiple-cutoff regression discontinuity designs in educational evaluation

Abstract

The regression discontinuity design (RDD) can be used for evaluating cut-score based educational interventions. The design enables researchers and evaluators to estimate the local causal effect of an intervention provided to those who are most at-risk. In certain educational evaluations, an RDD with multiple cutoffs can also be applied, where different cutoffs across classrooms or schools are used to assign students to the intervention condition. The availability of multiple cutoffs allows estimation of a pooled average treatment effect across cutoffs and/or individual estimates at each cutoff location, allowing for the possibility of heterogeneous treatment effects. The purpose of this paper is to demonstrate the design and analysis of the multiple-cutoff regression discontinuity (MCRD) design for the evaluation of need-based educational programs. Using data obtained from a kindergarten mathematics intervention for low achieving students, we illustrate the utility of the MCRD design for examining the average and potential variability of the regression discontinuity (RD) treatment effect. Implications for evaluation research and practice in education are discussed.