Bayesian inference for causal effects: The role of randomization
Open accessAbstract
Causal effects are comparisons among values that would have been observed under all possible assignments of treatments to experimental units. In an experiment, one assignment of treatments is chosen and only the values under that assignment can be observed. Bayesian inference for causal effects follows from finding the predictive distribution of the values under the other assignments of treatments. This perspective makes clear the role of mechanisms that sample experimental units, assign treatments and record data. Unless these mechanisms are ignorable (known probabilistic functions of recorded values), the Bayesian must model them in the data analysis and, consequently, confront inferences for causal effects that are sensitive to the specification of the prior distribution of the data. Moreover, not all ignorable mechanisms can yield data from which inferences for causal effects are insensitive to prior specifications. Classical randomized designs stand out as especially appealing assignment mechanisms designed to make inference for causal effects straightforward by limiting the sensitivity of a valid Bayesian analysis.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Counterfactual explanation generation appears highly relevant to sequential tasks in robotics such as automatic planning.
“Contfactual explanation generation appears highly relevant to sequential tasks in robotics such as automatic planning [94], [157], [161].”
From A Survey of Contrastive and Counterfactual Explanation Generation Methods for Explainable Artificial Intelligence (2021) · cites this paper as [157]
Sukkerd et al.'s objective is to find an optimal policy that minimizes the expected cumulative cost of reaching a goal state over all closed policies.
“Hence, their objective is to find an optimal policy ''that minimizes the expected cumulative cost of reaching a goal state over all closed policies''.”
From A Survey of Contrastive and Counterfactual Explanation Generation Methods for Explainable Artificial Intelligence (2021) · cites this paper as [157]
The explanation is believed to justify the rejection of the policies alternative to the optimal one.
“The explanation is believed to justify the rejection of the policies alternative to the optimal one.”
From A Survey of Contrastive and Counterfactual Explanation Generation Methods for Explainable Artificial Intelligence (2021) · cites this paper as [157]
Sukkerd et al. formulate the planning problem as the shortest stochastic path problem and develop the corresponding problem solver to obtain a contrastive explanation.
“For instance, Sukkerd et al. formulate the planning problem as the shortest stochastic path problem and develop the corresponding problem solver to obtain a contrastive explanation [157].”
From A Survey of Contrastive and Counterfactual Explanation Generation Methods for Explainable Artificial Intelligence (2021) · cites this paper as [157]
Multiple-sentence counterfactual explanations provide end-users with specific details for the given decision or explain multiple decisions at once.
“In contrast, multiple-sentence explanations provide end-users with specific details for the given decision [153], [156], [157], [161] or explain multiple decisions at once [146].”
From A Survey of Contrastive and Counterfactual Explanation Generation Methods for Explainable Artificial Intelligence (2021) · cites this paper as [157]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 19 · Cited by 2,503
Research works printed but not linked: 3
- BAILAR, J. C. (1976). Patient assignment algorithms—an overview. Proc. 9th International Biometric Conf. 1 189-206. The Biometric Society, Raleigh.
- CAMPBELL, D. T. and ERLEBACHER, A. (1970). How regression artifacts in quasi-experimental evaluations can mistakenly make compensatory education look harmful. In The Disadvataged Child, Compensatory Education: A National Debate, 3 (J. Hellmuth, ed.). Brunner/Mazel, New York.
- DE FINETTI, B. (1963). Foresight: its logical laws, its subjective sources. Studies in Subjective Probability (H. E. Kyburg and H. E. Smokler, eds.). Wiley, New York.
- Authors
- Rubin, D. B.
- Corresponding author
- Rubin, Donald B.
- ORCID
- Rubin, Donald B.
- Venue
- The Annals of Statistics
- Publisher
- Institute of Mathematical Statistics
- Published
- 1978-01-01
- Volume
- 6
- Issue
- 1
- Pages
- 34-58
- ISSN
- 0090-5364 (ISSN-L)
- 2168-8966
- OpenAlex
- W2064790857
- Cited by
- 2,503
- References
- 19
- Citations by year
- 2026: 52 · 2025: 103 · 2024: 85 · 2023: 92 · 2022: 111 · 2021: 107 · 2020: 135 · 2019: 118 · 2018: 109 · 2017: 82 · 2016: 106 · 2015: 133 · 2014: 118 · 2013: 95 · 2012: 101
- FWCI
- 6.06
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- The Annals of Statistics (publishedVersion)
- Project Euclid (Cornell University) (submittedVersion)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
- Physical Sciences › Computer Science › Artificial Intelligence › Bayesian Modeling and Causal Inference
- Keywords
- causal inference, Bayesian inference, randomization, randomized experiments, assignment mechanisms, prior sensitivity, predictive distribution, causal effect estimation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
