"Regression-Discontinuity Design" By Any Other Name Might Be Less Problematic
- 1 October 1991
- journal article
- research article
- Published by SAGE Publications in Evaluation Review
- Vol. 15 (5) , 605-624
- https://doi.org/10.1177/0193841x9101500505
Abstract
In "Random Measurement Error Does Not Bias the Treatment Estimate in the Regression- Discontinuity Design : II. When an Interaction Effect Is Present, " Trochim, Cappelleri, and Reichardt claim to "set the record straight regarding the RD design. " Instead, they defend the record of regression discontinuity (RD) and blur the statistical tests for treatment effect. Not withstanding their explicit assertions to the contrary, the Monte Carlo results they report show how the practical application of RD remains problematic and potentially biased. New testing strategies and restrictions for the application of RD are proposed.Keywords
This publication has 13 references indexed in Scilit:
- Random Measurement Error Does Not Bias the Treatment Effect Estimate in the Regression-Discontinuity DesignEvaluation Review, 1991
- Teaching to Talent: Evaluating an Enriched and Accelerated Mathematics ProgramJournal for the Education of the Gifted, 1989
- Estimating the superiority of a drug to a placebo when all and only those patients at risk are treated with the drug.Proceedings of the National Academy of Sciences, 1989
- Forecasting from fallible data: Correcting prediction bias with stein‐rule least squaresJournal of Forecasting, 1988
- Regression Discontinuity: Integrating Research and Program Design in Programs for the GiftedJournal for the Education of the Gifted, 1986
- Stein-rule least squares estimationEconomics Letters, 1986
- Advanced Econometric MethodsPublished by Springer Nature ,1984
- Specification Tests in EconometricsEconometrica, 1978
- Falsification and the Methodology of Scientific Research ProgrammesPublished by Cambridge University Press (CUP) ,1970
- Reforms as experiments.American Psychologist, 1969