Lasso adjustments of treatment effect estimates in randomized experiments

Abstract

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 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 OLS for estimation performs particularly well, and it chooses a smoothing parameter based on combined performance of Lasso and OLS.

References in corpus (3)

• Agnostic notes on regression adjustments to experimental data: Reexamining Freedman's critique
• On regression adjustments in experiments with several treatments
• Randomization Does Not Justify Logistic Regression

Cited by in corpus (20)

• High-dimensional regression adjustments in randomized experiments
• Regression adjustment in completely randomized experiments with a diverging number of covariates
• Randomization Tests for Weak Null Hypotheses in Randomized Experiments
• A general theory of regression adjustment for covariate-adaptive randomization: OLS, Lasso, and beyond
• Shrinkage Estimators in Online Experiments
• Design-Based Ratio Estimators and Central Limit Theorems for Clustered, Blocked RCTs
• Covariate-adjusted Fisher randomization tests for the average treatment effect
• Regression-adjusted average treatment effect estimates in stratified randomized experiments
• Model-assisted analyses of cluster-randomized experiments
• Regression adjustments for estimating the global treatment effect in experiments with interference
• Design-based theory for Lasso adjustment in randomized block experiments and rerandomized experiments
• Heterogeneous Treatment Effect Estimation through Deep Learning
• Principled estimation of regression discontinuity designs
• Sharp bounds for variance of treatment effect estimators in the finite population in the presence of covariates
• Model-assisted complier average treatment effect estimates in randomized experiments with non-compliance and a binary outcome
• Rerandomization and Regression Adjustment
• On High Dimensional Covariate Adjustment for Estimating Causal Effects in Randomized Trials with Survival Outcomes
• Estimation of Discrete Choice Models: A Machine Learning Approach
• Randomization-based joint central limit theorem and efficient covariate adjustment in stratified $2^K$ factorial experiments
• Rerandomization with Diminishing Covariate Imbalance and Diverging Number of Covariates