The IMS occupies a distinctive position in the statistical profession, focusing on the mathematical and theoretical underpinnings of statistics and probability. While the IMS is not exclusively Bayesian, many of the most important theoretical contributions to Bayesian statistics—including foundational results in decision theory, admissibility, posterior consistency, and nonparametric Bayes—have been published in IMS journals and presented at IMS meetings.
History and Mission
Founded in 1935, the IMS promotes the development of mathematical statistics and probability through publications, meetings, and awards. The society publishes some of the most prestigious journals in the field, including the Annals of Statistics, the Annals of Probability, and Statistical Science. These journals have been the venue for many landmark papers in Bayesian theory.
The IMS is founded to promote the development of mathematical statistics and probability.
The Annals of Mathematical Statistics begins publication, later splitting into the Annals of Statistics and the Annals of Probability in 1973.
Statistical Science is launched, featuring review articles and discussion papers that have included many influential pieces on Bayesian methods.
The Annals of Statistics has published foundational Bayesian work including Berger's contributions to decision theory, results on posterior consistency and rates of contraction in nonparametric Bayesian models, and theoretical analyses of MCMC convergence. Statistical Science has featured influential discussion papers on topics including the Bayesian-frequentist debate, objective Bayesian methods, and the philosophy of prior distributions.
Bayesian Contributions
The IMS's contribution to Bayesian statistics is primarily through the theoretical research published in its journals and presented at its meetings. Key areas include:
Decision theory: The IMS has been the principal venue for research on Bayesian decision theory, including work on admissibility, minimaxity, and the connections between Bayesian and frequentist optimality. Nonparametric Bayes: Foundational results on the Dirichlet process, posterior consistency, and rates of convergence in infinite-dimensional models have appeared in IMS publications. Asymptotic theory: Research on Bernstein-von Mises theorems, Bayesian model selection consistency, and the large-sample behavior of Bayesian procedures has been central to the IMS's Bayesian portfolio.
Awards
The IMS recognizes outstanding research through several awards, many of which have been bestowed on researchers who have made fundamental contributions to Bayesian statistics. IMS Fellows include many of the most prominent names in Bayesian theory, and the Wald Lectures, Neyman Lectures, and Medallion Lectures have frequently featured Bayesian topics.
"The IMS has always been the home of rigorous mathematical statistics. The fact that so much of the best Bayesian theory appears in IMS journals reflects the deep mathematical foundations of the Bayesian approach."— IMS member