Frequentist and Bayesian approaches to data analysis: Evaluation and estimation

Abstract

Statistical thinking is essential to understanding the nature of scientific results as a consumer. Statistical thinking also facilitates thinking like a scientist. Instead of emphasizing a “correct” procedure for data analysis and its outcome, statistical thinking focuses on the process of data analysis. This article reviews frequentist and Bayesian approaches such that teachers can promote less well-known statistical perspectives to encourage statistical thinking. Within the frequentist and Bayesian approaches, we highlight important distinctions between statistical evaluation versus estimation using an example on the facial feedback hypothesis. We first introduce some elementary statistical concepts, which are then illustrated with simulated data. Finally, we demonstrate how these approaches are applied to empirical data obtained from a Registered Replication Report. Data and R code for the example are provided as supplementary teaching material. We conclude with a discussion of key learning outcomes centred on promoting statistical thinking.