Challenges in Obtaining Valid Causal Effect Estimates with Machine Learning Algorithms

TL;DR

This paper evaluates the performance of machine learning methods for causal effect estimation, highlighting their limitations and identifying approaches that improve bias and coverage in complex confounding scenarios.

Contribution

It demonstrates that ML-based double-robust estimators can outperform single-robust ones, especially when combined with sample splitting and rich modeling.

Findings

Double-robust estimators had less bias than single-robust ones.

ML algorithms with sample splitting and interactions achieved nominal coverage.

Single-robust ML estimators performed poorly in complex confounding scenarios.

Abstract

Unlike parametric regression, machine learning (ML) methods do not generally require precise knowledge of the true data generating mechanisms. As such, numerous authors have advocated for ML methods to estimate causal effects. Unfortunately, ML algorithms can perform worse than parametric regression. We demonstrate the performance of ML-based single- and double-robust estimators. We use 100 Monte Carlo samples with sample sizes of 200, 1200, and 5000 to investigate bias and confidence interval coverage under several scenarios. In a simple confounding scenario, confounders were related to the treatment and the outcome via parametric models. In a complex confounding scenario, the simple confounders were transformed to induce complicated nonlinear relationships. In the simple scenario, when ML algorithms were used, double-robust estimators were superior to single-robust estimators. In the complex scenario, single-robust estimators with ML algorithms were at least as biased as estimators using misspecified parametric models. Double-robust estimators were less biased, but coverage was well below nominal. The use of sample splitting, inclusion of confounder interactions, reliance on a richly specified ML algorithm, and use of doubly robust estimators was the only explored approach that yielded negligible bias and nominal coverage. Our results suggest that ML based singly robust methods should be avoided.