Identifying causal effects with proxy variables of an unmeasured confounder

DOI10.1093/BIOMET/ASY038zbMATH Open1506.62451arXiv1609.08816OpenAlexW2962685611WikidataQ104507139 ScholiaQ104507139MaRDI QIDQ4562744FDOQ4562744

Authors: Wang Miao, Zhi Geng, Eric J. Tchetgen Tchetgen 

Publication date: 18 December 2018

Published in: Biometrika (Search for Journal in Brave)

Abstract: We consider a causal effect that is confounded by an unobserved variable, but with observed proxy variables of the confounder. We show that, with at least two independent proxy variables satisfying a certain rank condition, the causal effect is nonparametrically identified, even if the measurement error mechanism, i.e., the conditional distribution of the proxies given the con- founder, may not be identified. Our result generalizes the identification strategy of Kuroki & Pearl (2014) that rests on identification of the measurement error mechanism. When only one proxy for the confounder is available, or the required rank condition is not met, we develop a strategy to test the null hypothesis of no causal effect.

Full work available at URL: https://arxiv.org/abs/1609.08816

Recommendations

• Proxy variables and nonparametric identification of causal effects
• The identification of direct and indirect effects in studies with an unmeasured intermediate variable
• Identification of average effects under magnitude and sign restrictions on confounding
• Bounding the effects of proxy variables on instrumental-variables coefficients
• Bounding the Effects of Proxy Variables on Regression Coefficients

zbMATH Keywords

proxyidentificationmeasurement errorconfoundernegative control

Mathematics Subject Classification ID

Applications of statistics to biology and medical sciences; meta analysis (62P10) Causal inference from observational studies (62D20)

Cited In (27)

• Proximal Learning for Individualized Treatment Regimes Under Unmeasured Confounding
• Semiparametric Proximal Causal Inference
• Double Negative Control Inference in Test-Negative Design Studies of Vaccine Effectiveness
• Estimating treatment effects in the presence of unobserved confounders
• Half-trek criterion for identifiability of latent variable models
• An introduction to proximal causal inference
• A self-censoring model for multivariate nonignorable nonmonotone missing data
• A Particle Method for Solving Fredholm Equations of the First Kind
• Proxy variables and nonparametric identification of causal effects
• Measurement bias and effect restoration in causal inference
• Identification of average effects under magnitude and sign restrictions on confounding
• The identification of direct and indirect effects in studies with an unmeasured intermediate variable
• Off-policy evaluation in partially observed Markov decision processes under sequential ignorability
• Online causal inference with application to near real-time post-market vaccine safety surveillance
• Counterfactual inference with latent variable and its application in mental health care
• Assumption-lean falsification tests of rate double-robustness of double-machine-learning estimators
• Regression‐based negative control of homophily in dyadic peer effect analysis
• De-confounding Causal Inference Using Latent Multiple-Mediator Pathways
• Sensitivity to Unobserved Confounding in Studies with Factor-Structured Outcomes
• Identification and estimation of causal effects in the presence of confounded principal strata
• A confounding bridge approach for double negative control inference on causal effects
• Proximal causal inference without uniqueness assumptions
• Bounding the Effects of Proxy Variables on Regression Coefficients
• Comment: Reflections on the Deconfounder
• Confounder detection in high-dimensional linear models using first moments of spectral measures
• Identifying Effects of Multiple Treatments in the Presence of Unmeasured Confounding
• Semiparametric Inference of Causal Effect with Nonignorable Missing Confounders

This page was built for publication: Identifying causal effects with proxy variables of an unmeasured confounder

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4562744)