Distributional Causal Mediation via Conditional Generative Modeling

TL;DR

This paper introduces Distributional Causal Mediation Analysis (DCMA), a generative modeling framework that estimates entire outcome distributions and their causal effects, capturing complex nonlinear mediations beyond mean effects.

Contribution

DCMA is a novel framework that learns conditional generative models to recover and compare outcome distributions under different treatments using observational data.

Findings

DCMA accurately estimates distributional causal effects in simulations.

It captures rich distributional contrasts like Wasserstein and energy distances.

Empirical results demonstrate effectiveness on real-world datasets.

Abstract

Mediation analysis has traditionally focused on outcome-level summary contrasts, such as mean effects, which may obscure substantial distributional changes induced by complex and nonlinear causal mechanisms. We propose Distributional Causal Mediation Analysis (DCMA), a generative learning framework for identifying and estimating treatment effects on entire outcome distributions transmitted through multiple mediators. DCMA learns conditional generative models for the mediators and the outcome, recovering the relevant conditional distributions from observational data. Leveraging the identification formulas, it reconstructs interventional outcome distributions via Monte Carlo forward simulation by noise resampling, enabling the capture of both classical summary effects and rich distributional contrasts such as energy distance and the Wasserstein distance. Analytical error bounds are derived to decompose how estimation errors in the learned conditional models propagate to the reconstructed interventional outcome distributions. The empirical effectiveness of DCMA is demonstrated through numerical experiments and real-world data applications.