Proximal Path-Specific Inference
Not on the map. Why?
Abstract (arXiv)
Causal mediation analysis has been extended to estimate path-specific effects with multiple intermediate variables, isolating treatment effects through a mediator of interest while excluding pathways through its ancestors. Such analyses address bias from recanting witnesses, i.e., treatment-induced mediator-outcome confounders. However, existing methods typically rely on stringent assumptions precluding general unmeasured confounding, which are often violated in practice. In this paper, we relax these restrictions by leveraging observed covariates as proxy variables to accommodate unmeasured confounding among the treatment, recanting witness, mediator, and outcome. Using proximal confounding bridge functions, we develop four nonparametric identification strategies for the path-specific effect. We further derive the efficient influence function and propose a quadruply robust, locally efficient estimator. To handle high-dimensional nuisance parameters, we propose a proximal debiased machine learning approach. We theoretically guarantee that our estimator achieves √n-consistency and asymptotic normality even when machine learning estimators for nuisance functions converge at slower rates. Our approaches are validated via semiparametric and nonparametric simulations and an application to the CDC WONDER Natality study, estimating the path-specific effect of prenatal care on preterm birth through preeclampsia, independent of maternal smoking during pregnancy.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 25 of its 55 references that are in the map. First posted on arXiv 10 May 2026.
Builds on
25 of 55 references.
References · 25
- 2013Direct and Indirect Effectscited by 276
- 2016Identifying Causal Effects With Proxy Variables of an Unmeasured Confoundercited by 138
- 2022Semiparametric doubly robust targeted double machine learning: a reviewcited by 129
- 2020An Introduction to Proximal Causal Learningcited by 91
- 2020Minimax Estimation of Conditional Moment Modelscited by 83
- 2020Semiparametric proximal causal inferencecited by 58
- 2021Causal Inference Under Unmeasured Confounding With Negative Controls: A Minimax Learning Approachcited by 43
- 2021Proximal Learning for Individualized Treatment Regimes Under Unmeasured Confoundingcited by 21
- 2021Minimax Kernel Machine Learning for a Class of Doubly Robust Functionals with Application to Proximal Causal Inferencecited by 20
- 2021Proximal Causal Inference for Complex Longitudinal Studiescited by 16
- 2021Proximal mediation analysiscited by 10
- 2021Theory for identification and Inference with Synthetic Controls: A Proximal Causal Inference Frameworkcited by 9
13 more
- 2019Non-parametric efficient causal mediation with intermediate confounderscited by 8
- 2017On semiparametric estimation of a path-specific effect in the presence of mediator-outcome confoundingcited by 6
- 2021Causal Inference with Hidden Mediatorscited by 6
- 2022Proximal Causal Inference for Marginal Counterfactual Survival Curvescited by 5
- 2022Proximal Survival Analysis to Handle Dependent Right Censoringcited by 5
- 2022Optimal Treatment Regimes for Proximal Causal Learningcited by 5
- 2023Proximal Causal Inference for Synthetic Control with Surrogatescited by 4
- 2023Doubly Robust Proximal Causal Learning for Continuous Treatmentscited by 4
- 2024On Identification of Optimal Dynamic Treatment Regimes with Proxies of Hidden Confounderscited by 2
- 2024Identification and multiply robust estimation of causal effects via instrumental variables from an auxiliary populationcited by 2
- 2025On Multiple Robustness of Proximal Dynamic Treatment Regimescited by 2
- 2024General targeted machine learning for modern causal mediation analysiscited by 1
- 2024Nonparametric Estimation of Path-specific Effects in Presence of Nonignorable Missing Covariatescited by 1
Built on by
Citing papers per year: 2026: 1.
1 paper.
- 2026Proximal Mediation Analysis with Hidden Recanting Witnesses16 Jun 2026