Evidence Library

Statistical Power for Regression Discontinuity Designs in Education: Empirical Estimates of Design Effects Relative to Randomized Controlled Trials
Evidence | Publications

Statistical Power for Regression Discontinuity Designs in Education: Empirical Estimates of Design Effects Relative to Randomized Controlled Trials

Working Paper 8
Published: Jun 30, 2012
Publisher: Princeton, NJ: Mathematica Policy Research
Human Services Education Strengthening and Disseminating Research Labor: Strengthening and Disseminating Research

Download

Full Publication

Authors

John Deke

Lisa Dragoset

Using data from four previously published education studies, this working paper finds that a study using a regression discontinuity design needs between 9 and 17 times as many schools or students as a randomized controlled trial to produce an impact with the same level of statistical precision. The need for a large sample is driven primarily by bandwidth selection, not adjusting for random misspecification error.

Efficiency Meets Impact.
That's Progress Together.

To solve their most pressing challenges, organizations turn to Mathematica for deeply integrated expertise. We bring together subject matter and policy experts, data scientists, methodologists, and technologists who work across topics and sectors to help our partners design, improve, and scale evidence-based solutions.

Work With Us