CAST Seminar: Andrew Gelman

Join us at the CANSSI Ontario STatistics Seminars (CAST) with

Andrew Gelman

Professor, Departments of Statistics and Political Science
Director, Applied Statistics Centre
Columbia University

Talk Title

Hierarchical Bayesian Modeling as a Way of Life

Suggested Readings

[2004] Standard voting power indexes don’t work: An empirical analysis:  https://sites.stat.columbia.edu/gelman/research/published/gelmankatzbafumi.pdf

[2006] Multilevel (hierarchical) modeling: What it can and cannot do:  https://sites.stat.columbia.edu/gelman/research/published/multi2.pdf

[2006] Bayesian measures of explained variance and pooling in multilevel (hierarchical) models:  https://sites.stat.columbia.edu/gelman/research/published/rsquared.pdf

[2018] R-squared for Bayesian regression models:  https://sites.stat.columbia.edu/gelman/research/published/bayes_R2_v3.pdf

[2022] A proposal for informative default priors scaled by the standard error of estimates:  https://sites.stat.columbia.edu/gelman/research/published/default_prior_zwet.pdf

Abstract

Hierarchical Bayesian modeling is useful for partial pooling in meta-analysis, small-area estimation, and many other areas in which we want to make dense inferences from sparse data.  Although these methods have been around for several decades, many little-known issues and open questions remain.  We discuss several of these, including meta-analysis with a single study, empirical models for the signal-to-noise ratio, the relation between error variance and group size, anthropic priors, the search for a theoretical justification of weakly informative priors, different measures of influence, why we should care about R-squared, the Vermont/Wyoming problem, different ways to model varying treatment effects, the unification of causal methods in econometrics, and the possibility of a new level of abstraction for expressing multilevel models with interactions.

Speaker Profile

Andrew Gelman is a professor of statistics and political science at Columbia University. He has received the Outstanding Statistical Application award three times from the American Statistical Association, the award for best article published in the American Political Science Review, the Mitchell and DeGroot prizes from the International Society of Bayesian Analysis, and the Council of Presidents of Statistical Societies award.

Andrew has done research on a wide range of topics, including: why it is rational to vote; why campaign polls are so variable when elections are so predictable; the effects of incumbency and redistricting; reversals of death sentences; police stops in New York City, the statistical challenges of estimating small effects; the probability that your vote will be decisive; seats and votes in Congress; social network structure; arsenic in Bangladesh; radon in your basement; toxicology; medical imaging; and methods in surveys, experimental design, statistical inference, computation, and graphics.