Simple and honest confidence intervals in nonparametric regression

DOI10.3982/QE1199zbMATH Open1454.62124arXiv1606.01200WikidataQ109039003 ScholiaQ109039003MaRDI QIDQ5116129FDOQ5116129

Authors: Timothy B. Armstrong, Michal Kolesár 

Publication date: 24 August 2020

Published in: Quantitative Economics (Search for Journal in Brave)

Abstract: We consider the problem of constructing honest confidence intervals (CIs) for a scalar parameter of interest, such as the regression discontinuity parameter, in nonparametric regression based on kernel or local polynomial estimators. To ensure that our CIs are honest, we use critical values that take into account the possible bias of the estimator upon which the CIs are based. We show that this approach leads to CIs that are more efficient than conventional CIs that achieve coverage by undersmoothing or subtracting an estimate of the bias. We give sharp efficiency bounds of using different kernels, and derive the optimal bandwidth for constructing honest CIs. We show that using the bandwidth that minimizes the maximum mean-squared error results in CIs that are nearly efficient and that in this case, the critical value depends only on the rate of convergence. For the common case in which the rate of convergence is $n^{-2/5}$, the appropriate critical value for 95% CIs is 2.18, rather than the usual 1.96 critical value. We illustrate our results in a Monte Carlo analysis and an empirical application.

Full work available at URL: https://arxiv.org/abs/1606.01200

Recommendations

• Robust nonparametric confidence intervals for regression-discontinuity designs
• scientific article; zbMATH DE number 646821
• Automatic bandwidth choice and confidence intervals in nonparametric regression
• Pointwise Confidence Intervals in Nonparametric Regression with Heteroscedastic Error Structure
• Bootstrap confidence intervals in nonparametric regression with built-in bias correction

zbMATH Keywords

nonparametric regressionconfidence intervalsregression discontinuity

Mathematics Subject Classification ID

Nonparametric regression and quantile regression (62G08) Nonparametric tolerance and confidence regions (62G15)

Cited In (19)

• Estimation and inference of seller's expected revenue in first-price auctions
• Does health behavior change after diagnosis? Evidence from fuzzy regression discontinuity
• Honest confidence regions for nonparametric regression
• Testing continuity of a density via \(g\)-order statistics in the regression discontinuity design
• Confidence intervals for nonparametric regression functions under negatively associated errors
• Nonlinear and nonseparable structural functions in regression discontinuity designs with a continuous treatment
• Confidence intervals for intentionally biased estimators
• Minimizing sensitivity to model misspecification
• A Correction for Regression Discontinuity Designs With Group-Specific Mismeasurement of the Running Variable
• External Validity in Fuzzy Regression Discontinuity Designs
• Optimal HAR inference
• A more credible approach to parallel trends
• Title not available (Why is that?)
• Adaptation bounds for confidence bands under self-similarity
• Accurate confidence intervals in regression analyses of non-normal data
• On uniform confidence intervals for the tail index and the extreme quantile
• Intercept estimation in nonlinear selection models
• Simple confidence intervals for MCMC without CLTs
• Minimax linear estimation at a boundary point

This page was built for publication: Simple and honest confidence intervals in nonparametric regression

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5116129)