Use regression discontinuity designs to replicate or check experimental impact estimates for education interventions¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The report demonstrates using a regression discontinuity design as a replication approach for experimental impact estimates of education interventions. It "attempts to replicate the estimated impacts from experimental studies of two different education interventions using a regression discontinuity design", comparing the RD-based estimates against the experimental benchmarks. This offers evaluators a non-experimental method for assessing whether experimental findings hold when random assignment is not available.

Design Implications¶

Context¶

Requirements¶

• A forcing variable and cutoff that permit regression discontinuity estimation for the intervention being evaluated

Constraints¶

• The report examines two education interventions, so the demonstrated replication scope is limited to those cases; differences between RD and experimental point estimates were sometimes nontrivial

Target Learners¶

• Students in education interventions evaluated by IES and Mathematica studies

Target Learning Goals¶

• Estimating the causal impact of education interventions on student outcomes

Related Strategies¶

• A replication approach for testing nonexperimental methods against experimental ITT impact estimates when experiments include substantial control crossover

Examples¶

-

Key Sources¶

• Philip M. Gleason, Alexandra M. Resch, Jillian A. Berk. (2012). Replicating Experimental Impact Estimates Using a Regression Discontinuity Approach. Washington, DC: National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education. https://ies.ed.gov/ncee/