Open Access
Translator Disclaimer
January, 1978 Bayesian Inference for Causal Effects: The Role of Randomization
Donald B. Rubin
Ann. Statist. 6(1): 34-58 (January, 1978). DOI: 10.1214/aos/1176344064


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.


Download Citation

Donald B. Rubin. "Bayesian Inference for Causal Effects: The Role of Randomization." Ann. Statist. 6 (1) 34 - 58, January, 1978.


Published: January, 1978
First available in Project Euclid: 12 April 2007

zbMATH: 0383.62021
MathSciNet: MR472152
Digital Object Identifier: 10.1214/aos/1176344064

Primary: 62A15
Secondary: 62B15 , 62C10 , 62F15 , 62K99

Keywords: Bayesian , causality , experimentation , inference , missing data , Randomization

Rights: Copyright © 1978 Institute of Mathematical Statistics


Vol.6 • No. 1 • January, 1978
Back to Top