期刊
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
卷 113, 期 27, 页码 7383-7390出版社
NATL ACAD SCIENCES
DOI: 10.1073/pnas.1510506113
关键词
randomized experiment; Neyman-Rubin model; average treatment effect; high-dimensional statistics; Lasso
资金
- NSF [DMS-11-06753, DMS-12-09014, DMS-1107000, DMS-1129626, DMS-1209014]
- Computational and Data-Enabled Science and Engineering in Mathematical and Statistical Sciences (Focused Research Group) [1228246, DMS-1160319]
- AFOSR [FA9550-14-1-0016]
- NSA [H98230-15-1-0040]
- Center for Science of Information, a US NSF Science and Technology Center [CCF-0939370]
- Department of Defense for Office of Naval Research [N00014-15-1-2367]
- National Defense Science and Engineering Graduate Fellowship Program
- Direct For Mathematical & Physical Scien [1228246] Funding Source: National Science Foundation
- Direct For Mathematical & Physical Scien
- Division Of Mathematical Sciences [1513378] Funding Source: National Science Foundation
- Direct For Mathematical & Physical Scien
- Division Of Mathematical Sciences [1209014] Funding Source: National Science Foundation
- Division Of Mathematical Sciences [1228246] Funding Source: National Science Foundation
We provide a principled way for investigators to analyze randomized experiments when the number of covariates is large. Investigators often use linear multivariate regression to analyze randomized experiments instead of simply reporting the difference of means between treatment and control groups. Their aim is to reduce the variance of the estimated treatment effect by adjusting for covariates. If there are a large number of covariates relative to the number of observations, regression may perform poorly because of overfitting. In such cases, the least absolute shrinkage and selection operator (Lasso) may be helpful. We study the resulting Lasso-based treatment effect estimator under the Neyman-Rubin model of randomized experiments. We present theoretical conditions that guarantee that the estimator is more efficient than the simple difference-of-means estimator, and we provide a conservative estimator of the asymptotic variance, which can yield tighter confidence intervals than the difference-of-means estimator. Simulation and data examples show that Lasso-based adjustment can be advantageous even when the number of covariates is less than the number of observations. Specifically, a variant using Lasso for selection and ordinary least squares (OLS) for estimation performs particularly well, and it chooses a smoothing parameter based on combined performance of Lasso and OLS.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据