ERIC Number: ED565869
Record Type: Non-Journal
Publication Date: 2011-Sep
Abstractor: As Provided
Reference Count: 21
Heaping-Induced Bias in Regression-Discontinuity Designs. NBER Working Paper No. 17408
Barreca, Alan I.; Lindo, Jason M.; Waddell, Glen R.
National Bureau of Economic Research
This study uses Monte Carlo simulations to demonstrate that regression-discontinuity designs arrive at biased estimates when attributes related to outcomes predict heaping in the running variable. After showing that our usual diagnostics are poorly suited to identifying this type of problem, we provide alternatives. We also demonstrate how the magnitude and direction of the bias varies with bandwidth choice and the location of the data heaps relative to the treatment threshold. Finally, we discuss approaches to correcting for this type of problem before considering these issues in several non-simulated environments. The following figures are appended: (1) Rejection Rates at 5% Level (Assuming iid Errors) Using Alternative Data-Generating Processes; (2) Rejection Rate at 5% Level (Clustering SEs) Using Alternative Data-Generating Processes; and (3) Fraction of Births Recorded in 100s of Grams and Ounces Over Time.
Descriptors: Statistical Bias, Regression (Statistics), Research Design, Monte Carlo Methods, Prediction, Error of Measurement, Data Collection, Evaluation, Error Correction, Statistical Data, Birth, Body Weight, Statistical Distributions
National Bureau of Economic Research. 1050 Massachusetts Avenue, Cambridge, MA 02138-5398. Tel: 617-588-0343; Web site: http://www.nber.org
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: National Bureau of Economic Research
Identifiers - Location: California
IES Cited: ED560820