Estimating treatment effect heterogeneity in randomized program evaluation

Abstract

When evaluating the efficacy of social programs and medical treatments using randomized experiments, the estimated overall average causal effect alone is often of limited value and the researchers must investigate when the treatments do and do not work. Indeed, the estimation of treatment effect heterogeneity plays an essential role in (1) selecting the most effective treatment from a large number of available treatments, (2) ascertaining subpopulations for which a treatment is effective or harmful, (3) designing individualized optimal treatment regimes, (4) testing for the existence or lack of heterogeneous treatment effects, and (5) generalizing causal effect estimates obtained from an experimental sample to a target population. In this paper, we formulate the estimation of heterogeneous treatment effects as a variable selection problem. We propose a method that adapts the Support Vector Machine classifier by placing separate sparsity constraints over the pre-treatment parameters and causal heterogeneity parameters of interest. The proposed method is motivated by and applied to two well-known randomized evaluation studies in the social sciences. Our method selects the most effective voter mobilization strategies from a large number of alternative strategies, and it also identifies the characteristics of workers who greatly benefit from (or are negatively affected by) a job training program. In our simulation studies, we find that the proposed method often outperforms some commonly used alternatives.

Published in at http://dx.doi.org/10.1214/12-AOAS593 the Annals of Applied Statistics (http://www.imstat.org/aoas/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (4)

• A weakly informative default prior distribution for logistic and other regression models
• On the "degrees of freedom" of the lasso
• Performance guarantees for individualized treatment rules
• Predicting accurate probabilities with a ranking loss

Cited by in corpus (58)

• Recursive Partitioning for Heterogeneous Causal Effects
• Discussion on "Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects" by Hahn, Murray and Carvalho
• A Survey of Learning Causality with Data: Problems and Methods
• Machine Learning Estimation of Heterogeneous Causal Effects: Empirical Monte Carlo Evidence
• High-dimensional regression adjustments in randomized experiments
• Heterogeneous Employment Effects of Job Search Programmes: A Machine Learning Approach
• Response Transformation and Profit Decomposition for Revenue Uplift Modeling
• Modern approaches for evaluating treatment effect heterogeneity from clinical trials and observational data
• Selective inference for effect modification via the lasso
• Experimental Evaluation of Individualized Treatment Rules
• Estimation of causal effects using instrumental variables with nonignorable missing covariates: Application to effect of type of delivery NICU on premature infants
• Targeted Estimation of Heterogeneous Treatment Effect in Observational Survival Analysis
• Heterogeneous Indirect Effects for Multiple Mediators using Interventional Effect Models
• False Discovery Rate Controlled Heterogeneous Treatment Effect Detection for Online Controlled Experiments
• Evaluating the Effectiveness of Personalized Medicine with Software
• Learning Triggers for Heterogeneous Treatment Effects
• Causal Falling Rule Lists
• Estimating Heterogeneous Treatment Effects with Item-Level Outcome Data: Insights from Item Response Theory
• On Inductive Biases for Heterogeneous Treatment Effect Estimation
• Learning Cost-Effective Treatment Regimes using Markov Decision Processes
• Propensity Score Weighting for Causal Subgroup Analysis
• Robust Recursive Partitioning for Heterogeneous Treatment Effects with Uncertainty Quantification
• Exploration of heterogeneous treatment effects via concave fusion
• A nonparametric Bayesian analysis of heterogeneous treatment effects in digital experimentation
• Minimax Optimal Nonparametric Estimation of Heterogeneous Treatment Effects
• Personalized Treatment Selection using Causal Heterogeneity
• Efficient Targeted Learning of Heterogeneous Treatment Effects for Multiple Subgroups
• Flexible machine learning estimation of conditional average treatment effects: a blessing and a curse
• A Groupwise Approach for Inferring Heterogeneous Treatment Effects in Causal Inference
• Optimal Statistical Inference for Individualized Treatment Effects in High-dimensional Models
• Practical Guidance on Modeling Choices for the Virtual Twins Method
• Sharp Inference on Selected Subgroups in Observational Studies
• Stable discovery of interpretable subgroups via calibration in causal studies
• A unified survey of treatment effect heterogeneity modeling and uplift modeling
• Asymptotic normality of a linear threshold estimator in fixed dimension with near-optimal rate
• Heterogeneous Treatment Effect Estimation through Deep Learning
• A Semiparametric Approach to Model Effect Modification
• Modelling Heterogeneity Using Bayesian Structured Sparsity
• Estimating Heterogeneous Causal Effects of High-Dimensional Treatments: Application to Conjoint Analysis
• Machine learning in the social and health sciences
• Assessment of Heterogeneous Treatment Effect Estimation Accuracy via Matching
• Regression Discontinuity Design with Multiple Groups for Heterogeneous Causal Effect Estimation
• A causal fused lasso for interpretable heterogeneous treatment effects estimation
• Interactive identification of individuals with positive treatment effect while controlling false discoveries
• Covariate adjustment in subgroup analyses of randomized clinical trials: A propensity score approach
• Enhanced inference for distributions and quantiles of individual treatment effects in various experiments
• The foundations of cost-sensitive causal classification
• Generalizing Off-Policy Learning under Sample Selection Bias
• Treatment Targeting by AUUC Maximization with Generalization Guarantees
• Uplift Modeling from Separate Labels
• A Nonparametric Method for Value Function Guided Subgroup Identification via Gradient Tree Boosting for Censored Survival Data
• Learning Optimal Distributionally Robust Individualized Treatment Rules
• Predictive Directions for Individualized Treatment Selection in Clinical Trials
• The Bias-Variance Tradeoff of Doubly Robust Estimator with Targeted $L_1$ regularized Neural Networks Predictions
• PRISM: Patient Response Identifiers for Stratified Medicine
• Structural modeling using overlapped group penalties for discovering predictive biomarkers for subgroup analysis
• Modelling hetegeneous treatment effects by quantitle local polynomial decision tree and forest
• A New Transformation Approach for Uplift Modeling with Binary Outcome