Nonparametric Estimation of Heterogeneous Treatment Effects: From Theory to Learning Algorithms

TL;DR

This paper provides a theoretical analysis of four meta-learning strategies for nonparametric estimation of heterogeneous treatment effects, guiding algorithm design and demonstrating practical performance with neural networks.

Contribution

It offers a theoretical comparison of meta-learners for treatment effect estimation and applies these insights to neural network-based algorithms.

Findings

Different meta-learners perform variably depending on data-generating processes.

Theoretical insights can guide the choice of meta-learning strategies.

Neural network architectures effectively implement the proposed meta-learners.

Abstract

The need to evaluate treatment effectiveness is ubiquitous in most of empirical science, and interest in flexibly investigating effect heterogeneity is growing rapidly. To do so, a multitude of model-agnostic, nonparametric meta-learners have been proposed in recent years. Such learners decompose the treatment effect estimation problem into separate sub-problems, each solvable using standard supervised learning methods. Choosing between different meta-learners in a data-driven manner is difficult, as it requires access to counterfactual information. Therefore, with the ultimate goal of building better understanding of the conditions under which some learners can be expected to perform better than others a priori, we theoretically analyze four broad meta-learning strategies which rely on plug-in estimation and pseudo-outcome regression. We highlight how this theoretical reasoning can be used to guide principled algorithm design and translate our analyses into practice by considering a variety of neural network architectures as base-learners for the discussed meta-learning strategies. In a simulation study, we showcase the relative strengths of the learners under different data-generating processes.