pith. sign in

arxiv: 2606.17600 · v1 · pith:RPPNZ2KNnew · submitted 2026-06-16 · 📊 stat.ME · math.ST· stat.ML· stat.TH

Proximal Mediation Analysis with Hidden Recanting Witnesses

Sihan Wu , Yang Bai , Yifan Cui This is my paper

Pith reviewed 2026-06-26 23:43 UTC · model grok-4.3

classification 📊 stat.ME math.STstat.MLstat.TH
keywords mediation analysisproximal causal inferencerecanting witnessespath-specific effectsmultiply robust estimatorsemiparametric efficiencyhidden confoundersdebiased machine learning
0
0 comments X

The pith

Proximal causal inference enables identification of path-specific effects when recanting witnesses are hidden and unknown.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Mediation analysis typically requires that treatment-induced confounders between mediator and outcome, known as recanting witnesses, are either absent or fully observed. This paper shows how proximal causal inference can identify direct and indirect effects even when those witnesses are unmeasured and unknown. It provides three identification strategies based on proximal variables and derives an efficient influence function for a multiply robust estimator. The estimator is consistent if at least one set of the required nuisance models is correct and achieves the semiparametric efficiency bound when all are correct at suitable rates. A debiased machine learning procedure allows point estimation and confidence interval construction.

Core claim

The paper establishes three novel identification strategies for path-specific effects using proximal variables in the presence of hidden recanting witnesses, leading to a proximal multiply robust estimator that is consistent under partial correct specification of nuisance models and attains the efficiency bound when all models are correctly specified.

What carries the argument

The proximal multiply robust estimator constructed from the efficient influence function, which uses bridge functions for proximal variables to handle the hidden confounders.

If this is right

  • The path-specific effects remain identifiable without direct observation of recanting witnesses.
  • The estimator stays consistent provided at least one set of nuisance models is correctly specified.
  • Under full correct specification and appropriate convergence rates, the estimator is asymptotically normal and efficient.
  • Valid confidence intervals can be constructed using the debiased machine learning procedure.
  • Methods apply to simulation studies and real data applications in mediation analysis.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This approach could apply to other causal decomposition problems involving unmeasured variables by adapting the proximal framework.
  • Researchers might test the method in settings where some recanting witnesses can be measured to compare with standard approaches.
  • Extensions to longitudinal or time-to-event data could follow similar identification logic.
  • The reliance on completeness conditions suggests exploring sensitivity analyses for violations of those assumptions.

Load-bearing premise

Suitable proximal variables exist along with bridge functions that satisfy the completeness conditions needed to identify the effects despite the hidden recanting witnesses.

What would settle it

In a dataset with known recanting witnesses, if the proximal estimates of path-specific effects differ from those obtained by standard mediation methods that use the witnesses, the identification strategies would be called into question.

Figures

Figures reproduced from arXiv: 2606.17600 by Sihan Wu, Yang Bai, Yifan Cui.

Figure 1.1
Figure 1.1. Figure 1.1: Causal diagram of a mediation model. In complex real-world systems, standard mediation analyses targeting NIE are often invalidated by the presence of so-called recanting witnesses (Avin et al., 2005; Petersen et al., 2006). To formally establish this structure, as illustrated in [PITH_FULL_IMAGE:figures/full_fig_p003_1_1.png] view at source ↗
Figure 1.2
Figure 1.2. Figure 1.2: Causal diagram of a mediation model with recanting witnesses. [PITH_FULL_IMAGE:figures/full_fig_p003_1_2.png] view at source ↗
Figure 2.1
Figure 2.1. Figure 2.1: Causal diagram of the sequential mediation model with unobserved confounding. [PITH_FULL_IMAGE:figures/full_fig_p007_2_1.png] view at source ↗
Figure 3.1
Figure 3.1. Figure 3.1: Causal diagram of the sequential mediation model with proxies. [PITH_FULL_IMAGE:figures/full_fig_p010_3_1.png] view at source ↗
Figure 5.1
Figure 5.1. Figure 5.1: Residual distributions of four estimators (POR, PIPW, PHE, PMR) across [PITH_FULL_IMAGE:figures/full_fig_p023_5_1.png] view at source ↗
read the original abstract

Mediation analysis is essential for decomposing the causal effect of a treatment into direct and indirect pathways. However, many practical settings rely on the stringent assumption that recanting witnesses, defined as treatment-induced mediator-outcome confounders, are either absent or fully known a priori. Such a requirement is often untenable, especially when these variables remain unobservable due to measurement difficulties or privacy constraints. In this paper, we leverage proximal causal inference to develop three novel identification strategies to address the challenge of identifying path-specific effects in the presence of unknown recanting witnesses. Building on this, we develop a semiparametric inference framework that derives the efficient influence function and proposes a proximal multiply robust estimator, which remains consistent if at least one set of nuisance models is correctly specified. When all nuisance models are correctly specified and converge at appropriate rates, the estimator is asymptotically normal and achieves the semiparametric efficiency bound. We provide a minimax optimization-based debiased machine learning procedure for point estimation and constructing valid confidence intervals. The performance of the proposed methods is demonstrated by simulation studies and a real data application.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper develops three identification strategies based on proximal causal inference to identify path-specific effects in mediation analysis when recanting witnesses (treatment-induced mediator-outcome confounders) are hidden/unobserved. It derives the efficient influence function (EIF) and constructs a proximal multiply-robust estimator that is consistent if at least one set of nuisance models is correctly specified; when all models are correct and converge at suitable rates, the estimator is asymptotically normal and attains the semiparametric efficiency bound. A minimax-optimization debiased machine-learning procedure is proposed for estimation and inference, with performance illustrated via simulations and a real-data example.

Significance. If the proximal identification assumptions hold, the work meaningfully extends mediation analysis to settings with unmeasured recanting witnesses by importing proximal methods, while the multiply-robustness property and attainment of the efficiency bound (when all nuisances are correct) are practically valuable strengths. The combination of identification, EIF derivation, and debiased ML procedure provides a coherent semiparametric framework.

major comments (2)
  1. [Identification section / abstract] The three identification strategies (stated in the abstract and developed in the identification section) rest on the existence of proximal variables Z, W such that the relevant conditional-expectation operators satisfy completeness (invertibility) to recover the bridge functions that identify the path-specific effects. No sensitivity analysis, partial-identification bounds, or diagnostic for when completeness fails is provided; this is load-bearing because violation of completeness directly invalidates the identification equalities and therefore the consistency guarantee of the subsequent multiply-robust estimator.
  2. [Estimation / EIF derivation section] The abstract asserts that the proximal multiply-robust estimator remains consistent if at least one nuisance model is correct and attains the efficiency bound when all are correct and converge at appropriate rates, yet the manuscript summary does not exhibit the full EIF derivation, the explicit regularity conditions, or the proof of multiple robustness. Without these technical steps the central semiparametric claims cannot be verified.
minor comments (2)
  1. [Notation / preliminaries] Notation for the bridge functions and proximal variables could be introduced with a short table or diagram to aid readers new to proximal inference.
  2. [Simulation studies] The simulation section would benefit from explicit reporting of the chosen proximal variables and the degree to which the completeness condition is satisfied in the data-generating process.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. Below we respond point-by-point to the two major comments, indicating where we agree that revisions are warranted and where we believe the manuscript already addresses the concern.

read point-by-point responses

  1. Referee: [Identification section / abstract] The three identification strategies (stated in the abstract and developed in the identification section) rest on the existence of proximal variables Z, W such that the relevant conditional-expectation operators satisfy completeness (invertibility) to recover the bridge functions that identify the path-specific effects. No sensitivity analysis, partial-identification bounds, or diagnostic for when completeness fails is provided; this is load-bearing because violation of completeness directly invalidates the identification equalities and therefore the consistency guarantee of the subsequent multiply-robust estimator.

    Authors: We agree that the completeness condition is a key identifying assumption and that the manuscript would benefit from explicit discussion of its implications. While the paper follows the standard treatment of completeness in the proximal causal inference literature, we will add a dedicated paragraph in the Discussion section that (i) states the assumption clearly, (ii) notes that violation would invalidate the identification results, and (iii) outlines possible directions for future sensitivity or partial-identification analyses under weaker conditions. This addition will be made without altering the core identification theorems. revision: yes

  2. Referee: [Estimation / EIF derivation section] The abstract asserts that the proximal multiply-robust estimator remains consistent if at least one nuisance model is correct and attains the efficiency bound when all are correct and converge at appropriate rates, yet the manuscript summary does not exhibit the full EIF derivation, the explicit regularity conditions, or the proof of multiple robustness. Without these technical steps the central semiparametric claims cannot be verified.

    Authors: The full derivation of the efficient influence function, the statement of regularity conditions, and the proof of multiple robustness appear in Sections 3–4 and the supplementary appendix. To improve readability we will insert a short roadmap paragraph at the end of Section 2 that explicitly directs readers to these locations and will add a one-page outline of the multiple-robustness argument in the main text. The technical proofs themselves will remain in the appendix, consistent with journal conventions for semiparametric papers. revision: partial

Circularity Check

0 steps flagged

No significant circularity; derivation rests on external proximal assumptions

full rationale

The paper's identification strategies for path-specific effects are explicitly conditioned on the existence of proximal variables and bridge functions satisfying completeness/conditional independence conditions (as noted in the weakest assumption). The multiply robust estimator and its efficiency properties are then derived via standard semiparametric arguments (EIF derivation) from those assumptions. No equation reduces by construction to a fitted input, no load-bearing premise collapses to a self-citation, and no ansatz is smuggled via prior work by the same authors. The central claims therefore retain independent content from the proximal framework and are not equivalent to their inputs by definition.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on the existence of proximal variables and bridge functions satisfying completeness conditions; these are standard but non-trivial domain assumptions in proximal causal inference and are not independently verified in the abstract.

axioms (1)
  • domain assumption Existence of proximal variables and bridge functions satisfying the completeness or conditional independence conditions required for identification of path-specific effects.
    Invoked to justify the three identification strategies when recanting witnesses are hidden.

pith-pipeline@v0.9.1-grok · 5722 in / 1282 out tokens · 45040 ms · 2026-06-26T23:43:06.032418+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

50 extracted references · 10 canonical work pages · 3 internal anchors

  1. [1]

    Identification of causal effects using instrumental variables

    Joshua D Angrist, Guido W Imbens, and Donald B Rubin. Identification of causal effects using instrumental variables. Journal of the American Statistical Association, 91 0 (434): 0 444--455, 1996

    1996

  2. [2]

    Identifiability of path-specific effects

    Chen Avin, Ilya Shpitser, and Judea Pearl. Identifiability of path-specific effects. In IJCAI International Joint Conference on Artificial Intelligence, pages 357--363. International Joint Conferences on Artificial Intelligence, 2005

    2005

  3. [3]

    Proximal Path-Specific Inference

    Yang Bai, Sihan Wu, Baoluo Sun, and Yifan Cui. Proximal path-specific inference. arXiv preprint arXiv:2605.09462, 2026

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  4. [4]

    Double fairness policy learning: Integrating action fairness and outcome fairness in decision-making

    Zeyu Bian, Lan Wang, Chengchun Shi, and Zhengling Qi. Double fairness policy learning: Integrating action fairness and outcome fairness in decision-making. arXiv preprint arXiv:2601.19186, 2026

    work page arXiv 2026

  5. [5]

    Efficient and adaptive estimation for semiparametric models, volume 4

    Peter J Bickel, Chris AJ Klaassen, Ya’acov Ritov, and Jon A Wellner. Efficient and adaptive estimation for semiparametric models, volume 4. Springer, 1993

    1993

  6. [6]

    Department of Labor

    Bureau of Labor Statistics, U.S. Department of Labor . National longitudinal survey of youth 1997 cohort, 1997-2023 (rounds 1-21). Produced and distributed by the Center for Human Resource Research (CHRR), The Ohio State University, 2026

    1997

  7. [7]

    Double/debiased machine learning for treatment and structural parameters

    Victor Chernozhukov, Denis Chetverikov, Mert Demirer, Esther Duflo, Christian Hansen, Whitney Newey, and James Robins. Double/debiased machine learning for treatment and structural parameters. The Econometrics Journal, 21 0 (1): 0 C1--C68, 2018

    2018

  8. [8]

    Semiparametric proximal causal inference

    Yifan Cui, Hongming Pu, Xu Shi, Wang Miao, and Eric Tchetgen Tchetgen. Semiparametric proximal causal inference. Journal of the American Statistical Association, 119 0 (546): 0 1348--1359, 2024

    2024

  9. [9]

    Minimax estimation of conditional moment models

    Nishanth Dikkala, Greg Lewis, Lester Mackey, and Vasilis Syrgkanis. Minimax estimation of conditional moment models. Advances in Neural Information Processing Systems, 33: 0 12248--12262, 2020

    2020

  10. [10]

    Proximal mediation analysis

    Oliver Dukes, Ilya Shpitser, and Eric J Tchetgen Tchetgen. Proximal mediation analysis. Biometrika, 110 0 (4): 0 973--987, 2023

    2023

  11. [11]

    Robust inference on population indirect causal effects: the generalized front door criterion

    Isabel R Fulcher, Ilya Shpitser, Stella Marealle, and Eric J Tchetgen Tchetgen. Robust inference on population indirect causal effects: the generalized front door criterion. Journal of the Royal Statistical Society Series B: Statistical Methodology, 82 0 (1): 0 199--214, 2020

    2020

  12. [12]

    On multiple robustness of proximal dynamic treatment regimes

    Yuanshan Gao, Yang Bai, and Yifan Cui. On multiple robustness of proximal dynamic treatment regimes. arXiv preprint arXiv:2510.20451, 2025

    work page arXiv 2025

  13. [13]

    Minimax kernel machine learning for a class of doubly robust functionals with application to proximal causal inference

    AmirEmad Ghassami, Andrew Ying, Ilya Shpitser, and Eric Tchetgen Tchetgen. Minimax kernel machine learning for a class of doubly robust functionals with application to proximal causal inference. In International conference on artificial intelligence and statistics, pages 7210--7239. PMLR, 2022

    2022

  14. [14]

    Causal inference with hidden mediators

    Amiremad Ghassami, Alan Yang, Ilya Shpitser, and Eric Tchetgen Tchetgen. Causal inference with hidden mediators. Biometrika, 112 0 (1): 0 asae037, 2025

    2025

  15. [15]

    Proximal Causal Inference for Hidden Outcomes

    Helen Guo, Ilya Shpitser, and Elizabeth L Ogburn. Proximal causal inference for hidden outcomes. arXiv preprint arXiv:2605.09849, 2026

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  16. [16]

    Population intervention models in causal inference

    Alan E Hubbard and Mark J Van der Laan. Population intervention models in causal inference. Biometrika, 95 0 (1): 0 35--47, 2008

    2008

  17. [17]

    A general approach to causal mediation analysis

    Kosuke Imai, Luke Keele, and Dustin Tingley. A general approach to causal mediation analysis. Psychological methods, 15 0 (4): 0 309--334, 2010

    2010

  18. [18]

    Identification and estimation of local average treatment effects

    Guido W Imbens and Joshua D Angrist. Identification and estimation of local average treatment effects. Econometrica, 62 0 (2): 0 467--475, 1994

    1994

  19. [19]

    Causal inference under unmeasured confounding with negative controls: A minimax learning approach

    Nathan Kallus, Xiaojie Mao, and Masatoshi Uehara. Causal inference under unmeasured confounding with negative controls: A minimax learning approach. arXiv preprint arXiv:2103.14029, 2021

    work page arXiv 2021

  20. [20]

    Identifying causal effects with proxy variables of an unmeasured confounder

    Wang Miao, Zhi Geng, and Eric J Tchetgen Tchetgen. Identifying causal effects with proxy variables of an unmeasured confounder. Biometrika, 105 0 (4): 0 987--993, 2018

    2018

  21. [21]

    Quantifying an adherence path-specific effect of antiretroviral therapy in the nigeria pepfar program

    Caleb H Miles, Ilya Shpitser, Phyllis Kanki, Seema Meloni, and Eric J Tchetgen Tchetgen. Quantifying an adherence path-specific effect of antiretroviral therapy in the nigeria pepfar program. Journal of the American Statistical Association, 112 0 (520): 0 1443--1452, 2017

    2017

  22. [22]

    On semiparametric estimation of a path-specific effect in the presence of mediator-outcome confounding

    Caleb H Miles, Ilya Shpitser, Phyllis Kanki, Seema Meloni, and Eric J Tchetgen Tchetgen. On semiparametric estimation of a path-specific effect in the presence of mediator-outcome confounding. Biometrika, 107 0 (1): 0 159--172, 2020

    2020

  23. [23]

    Learning optimal fair policies

    Razieh Nabi, Daniel Malinsky, and Ilya Shpitser. Learning optimal fair policies. In International conference on machine learning, pages 4674--4682. PMLR, 2019

    2019

  24. [24]

    Semiparametric efficiency bounds

    Whitney K Newey. Semiparametric efficiency bounds. Journal of applied econometrics, 5 0 (2): 0 99--135, 1990

    1990

  25. [25]

    Instrumental variable estimation of nonparametric models

    Whitney K Newey and James L Powell. Instrumental variable estimation of nonparametric models. Econometrica, 71 0 (5): 0 1565--1578, 2003

    2003

  26. [26]

    Sur les applications de la th \'e orie des probabilit \'e s aux experiences agricoles: Essai des principes

    Jersey Neyman. Sur les applications de la th \'e orie des probabilit \'e s aux experiences agricoles: Essai des principes. Roczniki Nauk Rolniczych, 10 0 (1): 0 1--51, 1923

    1923

  27. [27]

    Direct and indirect effects

    Judea Pearl. Direct and indirect effects. In Probabilistic and causal inference: the works of Judea Pearl, pages 373--392. Association for Computing Machinery, 2022

    2022

  28. [28]

    Estimation of direct causal effects

    Maya L Petersen, Sandra E Sinisi, and Mark J van der Laan. Estimation of direct causal effects. Epidemiology, 17 0 (3): 0 276--284, 2006

    2006

  29. [29]

    Proximal learning for individualized treatment regimes under unmeasured confounding

    Zhengling Qi, Rui Miao, and Xiaoke Zhang. Proximal learning for individualized treatment regimes under unmeasured confounding. Journal of the American Statistical Association, 119 0 (546): 0 915--928, 2024

    2024

  30. [30]

    Adaptive proximal causal inference with some invalid proxies

    Prabrisha Rakshit, Xu Shi, and Eric Tchetgen Tchetgen. Adaptive proximal causal inference with some invalid proxies. arXiv preprint arXiv:2507.19623, 2025

    work page arXiv 2025

  31. [31]

    Higher order influence functions and minimax estimation of nonlinear functionals

    James Robins, Lingling Li, Eric Tchetgen Tchetgen, Aad van der Vaart, et al. Higher order influence functions and minimax estimation of nonlinear functionals. In Probability and statistics: essays in honor of David A. Freedman, volume 2, pages 335--422. Institute of Mathematical Statistics, 2008

    2008

  32. [32]

    Identifiability and exchangeability for direct and indirect effects

    James M Robins and Sander Greenland. Identifiability and exchangeability for direct and indirect effects. Epidemiology, 3 0 (2): 0 143--155, 1992

    1992

  33. [33]

    Robins, Lingling Li, Rajarshi Mukherjee, Eric Tchetgen Tchetgen, and Aad van der Vaart

    James M. Robins, Lingling Li, Rajarshi Mukherjee, Eric Tchetgen Tchetgen, and Aad van der Vaart. Minimax estimation of a functional on a structured high-dimensional model. The Annals of Statistics, 45 0 (5), 2017

    2017

  34. [34]

    Estimating causal effects of treatments in randomized and nonrandomized studies

    Donald B Rubin. Estimating causal effects of treatments in randomized and nonrandomized studies. Journal of Educational Psychology, 66 0 (5): 0 688--701, 1974

    1974

  35. [35]

    On asymptotically efficient estimation in semiparametric models

    Anton Schick. On asymptotically efficient estimation in semiparametric models. The Annals of Statistics, pages 1139--1151, 1986

    1986

  36. [36]

    Optimal treatment regimes for proximal causal learning

    Tao Shen and Yifan Cui. Optimal treatment regimes for proximal causal learning. Advances in Neural Information Processing Systems, 36: 0 47735--47748, 2023

    2023

  37. [37]

    A minimax learning approach to off-policy evaluation in confounded partially observable markov decision processes

    Chengchun Shi, Masatoshi Uehara, Jiawei Huang, and Nan Jiang. A minimax learning approach to off-policy evaluation in confounded partially observable markov decision processes. In International Conference on Machine Learning, pages 20057--20094. PMLR, 2022

    2022

  38. [38]

    Theory for identification and inference with synthetic controls: a proximal causal inference framework

    Xu Shi, Kendrick Qijun Li, Myeonghun Yu, Wang Miao, Arun Kumar Kuchibhotla, Mengtong Hu, and Eric Tchetgen Tchetgen. Theory for identification and inference with synthetic controls: a proximal causal inference framework. Journal of the American Statistical Association, pages 1--23, 2026

    2026

  39. [39]

    Counterfactual graphical models for longitudinal mediation analysis with unobserved confounding

    Ilya Shpitser. Counterfactual graphical models for longitudinal mediation analysis with unobserved confounding. Cognitive science, 37 0 (6): 0 1011--1035, 2013

    2013

  40. [40]

    Proximal causal learning of conditional average treatment effects

    Erik Sverdrup and Yifan Cui. Proximal causal learning of conditional average treatment effects. In International Conference on Machine Learning, pages 33285--33298. PMLR, 2023

    2023

  41. [41]

    Semiparametric theory for causal mediation analysis: efficiency bounds, multiple robustness, and sensitivity analysis

    Eric J Tchetgen Tchetgen and Ilya Shpitser. Semiparametric theory for causal mediation analysis: efficiency bounds, multiple robustness, and sensitivity analysis. Annals of Statistics, 40 0 (3): 0 1816, 2012

    2012

  42. [42]

    arXiv preprint arXiv:2009.10982 , year=

    Eric J Tchetgen Tchetgen, Andrew Ying, Yifan Cui, Xu Shi, and Wang Miao. An introduction to proximal causal learning. arXiv preprint arXiv:2009.10982, 2020

    work page arXiv 2009

  43. [43]

    Conceptual issues concerning mediation, interventions and composition

    Tyler VanderWeele and Stijn Vansteelandt. Conceptual issues concerning mediation, interventions and composition. Statistics and its Interface, 2: 0 457--468, 2009

    2009

  44. [44]

    Blessing from human- AI interaction: super policy learning in confounded environments

    Jiayi Wang, Chengchun Shi, and Zhengling Qi. Blessing from human- AI interaction: super policy learning in confounded environments. Journal of the American Statistical Association, pages 1--14, 2026

    2026

  45. [45]

    Proximal survival analysis to handle dependent right censoring

    Andrew Ying. Proximal survival analysis to handle dependent right censoring. Journal of the Royal Statistical Society Series B: Statistical Methodology, 86 0 (5): 0 1414--1434, 2024

    2024

  46. [46]

    Proximal causal inference for marginal counterfactual survival curves

    Andrew Ying, Yifan Cui, and Eric J Tchetgen Tchetgen. Proximal causal inference for marginal counterfactual survival curves. arXiv preprint arXiv:2204.13144, 2022

    work page arXiv 2022

  47. [47]

    Proximal causal inference for complex longitudinal studies

    Andrew Ying, Wang Miao, Xu Shi, and Eric J Tchetgen Tchetgen. Proximal causal inference for complex longitudinal studies. Journal of the Royal Statistical Society Series B: Statistical Methodology, 85 0 (3): 0 684--704, 2023

    2023

  48. [48]

    Fortified proximal causal inference with many invalid proxies

    Myeonghun Yu, Xu Shi, and Eric J Tchetgen Tchetgen. Fortified proximal causal inference with many invalid proxies. arXiv preprint arXiv:2506.13152, 2025

    work page arXiv 2025

  49. [49]

    On identification of optimal dynamic treatment regimes with proxies of hidden confounders

    Jeffrey Zhang and Eric Tchetgen Tchetgen. On identification of optimal dynamic treatment regimes with proxies of hidden confounders. Observational Studies, 12 0 (1): 0 1--15, 2026

    2026

  50. [50]

    Causal Inference with a Hidden Treatment

    Ying Zhou and Eric Tchetgen Tchetgen. Causal inference for a hidden treatment. arXiv preprint arXiv:2405.09080, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024