Confounding Feature Acquisition for Causal Effect Estimation

TL;DR

This paper addresses the challenge of efficiently acquiring missing confounding features in observational studies to improve causal effect estimation, proposing two strategies and demonstrating their effectiveness across multiple methods.

Contribution

It introduces two novel feature acquisition strategies, covariate balancing and outcome error reduction, tailored for confounder missingness in causal inference.

Findings

Outcome error reduction (OE) outperforms covariate balancing (CB) in sample efficiency.

Proposed methods improve causal effect estimation across five different models.

Visual analysis highlights differences between acquisition strategies.

Abstract

Reliable treatment effect estimation from observational data depends on the availability of all confounding information. While much work has targeted treatment effect estimation from observational data, there is relatively little work in the setting of confounding variable missingness, where collecting more information on confounders is often costly or time-consuming. In this work, we frame this challenge as a problem of feature acquisition of confounding features for causal inference. Our goal is to prioritize acquiring values for a fixed and known subset of missing confounders in samples that lead to efficient average treatment effect estimation. We propose two acquisition strategies based on i) covariate balancing (CB), and ii) reducing statistical estimation error on observed factual outcome error (OE). We compare CB and OE on five common causal effect estimation methods, and demonstrate improved sample efficiency of OE over baseline methods under various settings. We also provide visualizations for further analysis on the difference between our proposed methods.