On Multi-Cause Causal Inference with Unobserved Confounding: Counterexamples, Impossibility, and Alternatives

Unobserved confounding is a central barrier to drawing causal inferences from observational data. Several authors have recently proposed that this barrier can be overcome in the case where one attempts to infer the effects of several variables simultaneously. In this paper, we present two simple, analytical counterexamples that challenge the general claims that are central to these approaches. In addition, we show that nonparametric identification is impossible in this setting. We discuss practical implications, and suggest alternatives to the methods that have been proposed so far in this line of work: using proxy variables and shifting focus to sensitivity analysis.

• Publication year

  2019

• Venue

  arXiv.org

• Publication date

  2019-02-27

• Fields of study

  Mathematics, Computer Science, Economics

• Identifiers
  arXiv 1902.10286
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• The Deconfounded Recommender: A Causal Inference Approach to Recommendation

  2018cited by this paper
• Multiply robust causal inference with double‐negative control adjustment for categorical unmeasured confounding

  2018cited by this paper
• Flexible Sensitivity Analysis for Observational Studies Without Observable Implications

  2018cited by this paper
• Multiply Robust Causal Inference With Double Negative Control Adjustment for Unmeasured Confounding

  2018cited by this paper
• The Blessings of Multiple Causes

  2018cited by this paper
• Multiple Causal Inference with Latent Confounding

  2018influential reference
• Implicit Causal Models for Genome-wide Association Studies

  2017influential reference
• Causal Effect Inference with Deep Latent-Variable Models

  2017cited by this paper
• Identifying Causal Effects With Proxy Variables of an Unmeasured Confounder.

  2016cited by this paper
• Measurement bias and effect restoration in causal inference

  2014cited by this paper
• Identifiability of parameters in latent structure models with many observed variables

  2008cited by this paper
• Identification for Prediction and Decision

  2008cited by this paper
• Structural Equation Models with Latent Variables

  2002cited by this paper
• Structural Equation Models with Latent Variables

  2002cited by this paper
• An Introduction to Copulas

  2000cited by this paper
• Multivariate Models and Multivariate Dependence Concepts

  1997influential reference
• Rank, decomposition, and uniqueness for 3-way and n -way arrays

  1989cited by this paper
• Assessing Sensitivity to an Unobserved Binary Covariate in an Observational Study with Binary Outcome

  1983cited by this paper
• Causality

  1938influential reference

• Counterfactual Situation Testing: From Single to Multidimensional Discrimination

  2025cites this paper
• Towards a more robust comparative oncology: a Bayesian reanalysis of Peto’s paradox and discussion of comparative cancer risk studies in vertebrates

  2025cites this paper
• Single-Cell Proteomic Technologies: Tools in the quest for principles

  2025cites this paper
• Simultaneous Estimation of Multiple Treatment Effects from Observational Studies

  2025cites this paper
• A/B testing under Interference with Partial Network Information

  2024cites this paper
• Controlling for discrete unmeasured confounding in nonlinear causal models

  2024cites this paper
• Few-Shot Learner Parameterization by Diffusion Time-Steps

  2024cites this paper
• Counterfactual Situation Testing: Uncovering Discrimination under Fairness given the Difference

  2023cites this paper
• Eliminating confounder-induced bias in the statistics of intervention

  2023cites this paper
• Consistent End-to-End Estimation for Counterfactual Fairness

  2023cites this paper
• CHEMIST: an R package for causal inference with high-dimensional error-prone covariates and misclassified treatments

  2023cites this paper
• The Blessings of Multiple Treatments and Outcomes in Treatment Effect Estimation

  2023cites this paper
• Estimating Treatment Effects from Irregular Time Series Observations with Hidden Confounders

  2023cites this paper
• Debiasing Recommendation by Learning Identifiable Latent Confounders

  2023cites this paper
• A qualitative, network-centric method for modeling socio-technical systems, with applications to evaluating interventions on social media platforms to increase social equality

  2022cites this paper
• Deep Causal Reasoning for Recommendations

  2022cites this paper
• Exploring polypharmacy with artificial intelligence: data analysis protocol

  2021cites this paper
• Identification of Latent Variables From Graphical Model Residuals

  2021cites this paper
• Do-calculus enables causal reasoning with latent variable models

  2021cites this paper
• Adversarial Visual Robustness by Causal Intervention

  2021cites this paper
• Learning latent causal graphs via mixture oracles

  2021cites this paper
• M3E2: Multi-gate Mixture-of-experts for Multi-treatment Effect Estimation

  2021cites this paper
• Estimating Causal Effects of Multi-Aspect Online Reviews with Multi-Modal Proxies

  2021cites this paper
• Deconfounded Visual Grounding

  2021cites this paper
• Multi-treatment Effect Estimation from Biomedical Data.

  2021cites this paper
• Do-calculus enables estimation of causal effects in partially observed biomolecular pathways

  2021cites this paper
• Treatment Assignment Outcome Pre-treatment Covariates ( c )

  2020cites this paper
• Identifying Effects of Multiple Treatments in the Presence of Unmeasured Confounding

  2020cites this paper
• Causal Intervention for Weakly-Supervised Semantic Segmentation

  2020cites this paper
• Misspecification and unreliable interpretations in psychology and social science.

  2020cites this paper
• Targeted VAE: Structured Inference and Targeted Learning for Causal Parameter Estimation

  2020cites this paper
• Limited Functional Form, Misspecification, and Unreliable Interpretations in Psychology and Social Science

  2020cites this paper
• Targeted VAE: Variational and Targeted Learning for Causal Inference

  2020cites this paper
• Improving Mental Health Services: A 50-Year Journey from Randomized Experiments to Artificial Intelligence and Precision Mental Health

  2020cites this paper
• Comment on “Blessings of Multiple Causes”

  2019cites this paper
• Counterfactual Reasoning for Fair Clinical Risk Prediction

  2019cites this paper
• Comment: The Challenges of Multiple Causes

  2019cites this paper
• Mitigating unobserved spatial confounding when estimating the effect of supermarket access on cardiovascular disease deaths

  2019cites this paper
• Multiple Causes: A Causal Graphical View

  2019cites this paper
• Comment: Reflections on the Deconfounder

  2019cites this paper
• The Blessings of Multiple Causes: A Reply to Ogburn et al. (2019)

  2019cites this paper
• Discussion of "The Blessings of Multiple Causes" by Wang and Blei

  2019cites this paper
• Counterfactual Prediction Methods for Causal Inference in Observational Studies with Continuous Treatments

  2019cites this paper
• The Blessings of Multiple Causes

  2018influential citation
• A causality-inspired adjusted plus-minus model for player evaluation in team sports

  year unknowncites this paper
• 1 5 O ct 2 01 9 Comment on " Blessings of Multiple Causes ”

  year unknowncites this paper