Doubly robust treatment effect estimation with missing attributes

TL;DR

This paper investigates methods for estimating treatment effects when some attributes are missing, demonstrating that doubly robust approaches generally outperform standard methods in simulations and real-world data.

Contribution

It introduces and evaluates doubly robust modifications for treatment effect estimation with missing data, showing their superior performance over traditional methods.

Findings

Doubly robust methods outperform non-robust baselines in simulations.

Doubly robust estimators produce confidence intervals aligned with randomized trials.

No single method is universally best across all scenarios.

Abstract

Missing attributes are ubiquitous in causal inference, as they are in most applied statistical work. In this paper, we consider various sets of assumptions under which causal inference is possible despite missing attributes and discuss corresponding approaches to average treatment effect estimation, including generalized propensity score methods and multiple imputation. Across an extensive simulation study, we show that no single method systematically out-performs others. We find, however, that doubly robust modifications of standard methods for average treatment effect estimation with missing data repeatedly perform better than their non-doubly robust baselines; for example, doubly robust generalized propensity score methods beat inverse-weighting with the generalized propensity score. This finding is reinforced in an analysis of an observations study on the effect on mortality of tranexamic acid administration among patients with traumatic brain injury in the context of critical care management. Here, doubly robust estimators recover confidence intervals that are consistent with evidence from randomized trials, whereas non-doubly robust estimators do not.