Difference-in-differences for mediation analysis using double machine learning

2026-02-27

AI Summary

This study addresses the identification of direct and indirect effects of a treatment on an outcome in settings involving multi-valued or continuous treatments and mediators. To this end, the authors propose a double machine learning framework that integrates difference-in-differences with mediation analysis, leveraging nonparametric machine learning methods to flexibly control for covariates under a conditional parallel trends assumption within a potential outcomes framework. This work is the first to embed difference-in-differences mediation analysis into a double machine learning structure, establishing formal identification conditions and asymptotic theory; the proposed estimator is shown to be asymptotically normal under standard regularity conditions. Simulation studies demonstrate favorable finite-sample performance, and the method is successfully applied to National Longitudinal Survey of Youth data to assess the direct and indirect effects of health insurance coverage on health outcomes.

Technology Category

• Machine Learning: Causal Learning
• Multiagent Systems: Mechanism Design
• Game Theory and Economic Paradigms: Mechanism Design

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Web Mining and Content Analysis: Machine learning and data science for the Web
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations

Abstract

We propose a difference-in-differences (DiD) framework with mediation for possibly multivalued discrete or continuous treatments and mediators, aimed at identifying the direct effect of the treatment on the outcome (net of effects operating through the mediator), the indirect effect via the mediator, and the joint effects of treatment and mediator, consistent with the framework of dynamic treatment effects. Identification relies on a conditional parallel trends assumption imposed on the mean potential outcome across treatment and mediator states, or (depending on the causal parameter) additionally on the mean potential outcomes and potential mediator distributions across treatment states. We propose ATET estimators for repeated cross sections and panel data within the double/debiased machine learning framework, which allows for data-driven control of covariates, and we establish their asymptotic normality under standard regularity conditions. We investigate the finite-sample performance of the proposed methods in a simulation study and illustrate our approach in an empirical application to the US National Longitudinal Survey of Youth, estimating the direct effect of health care coverage on general health as well as the indirect effect operating through routine checkups.

Problem

• mediation analysis
• difference-in-differences
• direct effect
• indirect effect
• dynamic treatment effects

Innovation

• difference-in-differences
• mediation analysis
• double machine learning
• direct and indirect effects
• conditional parallel trends