My primary research area is in Bayesian model selection. Specifically, I investigate the case where the levels of a categorical predictor fall into latent groups. I have developed a fully Bayesian model selection approach of clustering the data according to the levels of a categorical predictor to reveal latent group-based fixed effects, heteroscedasticity, and/or hidden interactions. Through the use of mixture g-priors and fractional Bayes factors, I test for both the presence and structure of such clustering. This method is broadly applicable to the class of linear models that include categorical predictors. Currently, I am working on developing an R package for this implementation and investigating Bayesian model averaging in this context.
Thomas Metzger joined the statistics faculty in 2019. He earned bachelor's degrees in mathematics and actuarial science, as well as a master's degree in secondary mathematics education, from The Ohio State University in 2010-11.