A Subjective Bayesian Approach to Address Unchecked Assumptions in a Causal Analysis
Abstract
Causal analyses using observational data rely on identifying assumptions that cannot be verified from the observed data alone. When such unchecked assumptions fail, effect estimates may be biased and uncertainty may be understated. We propose the Adjusted Model for Unchecked Assumptions (AMUA), a subjective Bayesian framework that propagates uncertainty about the joint validity of key causal assumptions through a two-component posterior mixture. One component corresponds to the standard observational causal model in which the assumptions hold (referred to as the "assumption-hold component"); the other represents a flexibly defined model that allows the causal assumptions to fail (referred to as the "assumption-fail component"). The mixture weight π is specified by the analyst as the prior probability assigned to the assumption-hold component. This formulation leads to a new diagnostic, πmin: the smallest value of π for which the 95\% HPD set for the average treatment effect (ATE) excludes zero. In simulations with a spatially structured unmeasured confounder, we evaluate bias, interval length, and coverage as π varies. The results show that acknowledging assumption uncertainty can widen intervals, and improve coverage relative to analyses that treat the assumptions as certain. In a county-level application relating long-term PM2.5 exposure to COVID-19 mortality, AMUA shows how posterior conclusions depend on prior confidence in the identifying assumptions.
Related
- A General Approach for Assessing Sensitivity to Violations of Causal Assumptions
- Where Causal Discovery Breaks on Real Data: Measuring the Assumptions One Property at a Time
- Dangers of Bayesian analyses and how to address them
- Dirichlet-Tree Allocation Priors for Bayesian Causal Forests
- Graph-Split Bayesian Causal Forest for Spatial Heterogeneous Treatment Effect Estimation
BibTeX
@misc{kim2026subjective,
title = {{A Subjective Bayesian Approach to Address Unchecked Assumptions in a Causal Analysis}},
author = {Jaiyool Kim and Indrabati Bhattacharya and Jonathan R. Bradley},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.31587},
eprint = {2610.31587},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.31587},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - A Subjective Bayesian Approach to Address Unchecked Assumptions in a Causal Analysis
AU - Jaiyool Kim
AU - Indrabati Bhattacharya
AU - Jonathan R. Bradley
PY - 2026
DA - 2026/10/22
PB - arXiv
DO - 10.48550/arxiv.2610.31587
UR - https://arxiv.org/abs/2610.31587
AB - Causal analyses using observational data rely on identifying assumptions that cannot be verified from the observed data alone. When such unchecked assumptions fail, effect estimates may be biased and uncertainty may be understated. We propose the Adjusted Model for Unchecked Assumptions (AMUA), a subjective Bayesian framework that propagates uncertainty about the joint validity of key causal assumptions through a two-component posterior mixture. One component corresponds to the standard observational causal model in which the assumptions hold (referred to as the "assumption-hold component"); the other represents a flexibly defined model that allows the causal assumptions to fail (referred to as the "assumption-fail component"). The mixture weight π is specified by the analyst as the prior probability assigned to the assumption-hold component. This formulation leads to a new diagnostic, π_min: the smallest value of π for which the 95\% HPD set for the average treatment effect (ATE) excludes zero. In simulations with a spatially structured unmeasured confounder, we evaluate bias, interval length, and coverage as π varies. The results show that acknowledging assumption uncertainty can widen intervals, and improve coverage relative to analyses that treat the assumptions as certain. In a county-level application relating long-term PM_2.5 exposure to COVID-19 mortality, AMUA shows how posterior conclusions depend on prior confidence in the identifying assumptions.
ER -