Multiply Robust Causal Mediation Analysis with Continuous Treatments

TL;DR

This paper introduces a new multiply robust estimator for causal mediation analysis with continuous treatments, leveraging kernel smoothing and cross-fitting to relax assumptions and improve inference robustness.

Contribution

It extends influence function-based estimators to continuous treatments using kernel smoothing and cross-fitting, enabling robust and flexible causal mediation analysis.

Findings

Estimator is multiply robust and asymptotically normal.

Allows for slower convergence rates of nuisance parameters.

Applicable without strong parametric assumptions.

Abstract

In many applications, researchers are interested in the direct and indirect causal effects of a treatment or exposure on an outcome of interest. Mediation analysis offers a rigorous framework for identifying and estimating these causal effects. For binary treatments, efficient estimators for the direct and indirect effects are presented by Tchetgen Tchetgen and Shpitser (2012) based on the influence function of the parameter of interest. These estimators possess desirable properties such as multiple-robustness and asymptotic normality while allowing for slower than root-n rates of convergence for the nuisance parameters. However, in settings involving continuous treatments, these influence function-based estimators are not readily applicable without making strong parametric assumptions. In this work, utilizing a kernel-smoothing approach, we propose an estimator suitable for settings with continuous treatments inspired by the influence function-based estimator of Tchetgen Tchetgen and Shpitser (2012). Our proposed approach employs cross-fitting, relaxing the smoothness requirements on the nuisance functions and allowing them to be estimated at slower rates than the target parameter. Additionally, similar to influence function-based estimators, our proposed estimator is multiply robust and asymptotically normal, allowing for inference in settings where parametric assumptions may not be justified.