Falsification of Unconfoundedness by Testing Independence of Causal Mechanisms

TL;DR

This paper introduces a novel algorithm to falsify the assumption of no unmeasured confounding in observational studies by detecting dependencies in causal mechanisms across multiple environments, even under transportability violations.

Contribution

It presents a new two-stage procedure that identifies dependencies caused by unmeasured confounding without needing randomized data or strict transportability assumptions.

Findings

Effectively detects confounding in simulated data

Operates under violations of transportability assumptions

Controls false positives with high statistical power

Abstract

A major challenge in estimating treatment effects in observational studies is the reliance on untestable conditions such as the assumption of no unmeasured confounding. In this work, we propose an algorithm that can falsify the assumption of no unmeasured confounding in a setting with observational data from multiple heterogeneous sources, which we refer to as environments. Our proposed falsification strategy leverages a key observation that unmeasured confounding can cause observed causal mechanisms to appear dependent. Building on this observation, we develop a novel two-stage procedure that detects these dependencies with high statistical power while controlling false positives. The algorithm does not require access to randomized data and, in contrast to other falsification approaches, functions even under transportability violations when the environment has a direct effect on the outcome of interest. To showcase the practical relevance of our approach, we show that our method is able to efficiently detect confounding on both simulated and semi-synthetic data.