BMER-Project-Papers

Our work seeks to develop, apply, and disseminate a Bayesian alternative to the frequentist (classical) statistical paradigm used in rigorous empirical research in the education sciences. We investigate Bayesian approaches motivated by problems in large-scale observational and longitudinal studies. Our past work, funded by IES (Grant #R305D110001), has focused on developing Bayesian methods for propensity score analysis, missing data problems, and general issues of model uncertainty. Our current work, also funded by IES (Grant # R305D190053), focuses on the idea of Bayesian dynamic borrowing as a means of utilizing historical data in current analyses. We are also developing approaches for national and international education growth modeling using Bayesian methods. This latter work was supported by an initial seed grant from the International Association for the Evaluation of Educational Achievement (IEA).