Can Educational Cyberinfrastructure Empower Nonstatisticians with Bayesian Methodology? (303143)

Keywords: Bayesian, statistics, R, Shiny, consulting, collaboration. frequentist

The Bayesian Paradigm has transformed the statistical landscape over the last several decades, which is old news to statisticians. Outside of our field, however, Bayesian approaches are not necessarily considered standard practice. Technical barriers prevent many scientists from understanding or implementing modern Bayesian approaches, and the existing infrastructure (in the form of education, software, etc.) has had a longer time to develop for classical methods. In this talk I will briefly review a few Bayesian analogs to classical approaches. Then I will demonstrate point-and-click software (developed in R using the Shiny application) that allows the user to (a) import or simulate data, (b) perform a variety of standard classical and Bayesian analyses, and (c) compare differential performance and interpretation of the various approaches. The ultimate goal of this work is to develop open-source educational cyberinfrastructure to assist non-statisticians who want to learn about Bayesian methods and also statisticians who seek resources to educate students and researchers about these methods.