Bayesian methods provide a unifying framework for all of statistics, ranging from the test of simple hypotheses and the estimation of parameters to decision making using loss functions and the model-averaged prediction of key outcome variables. Bayesian methods are principled in the sense that they adhere to the laws of probability theory, meaning that all uncertainty is taken into account and that no contradictory statements can occur. At the same time, Bayesian methods are unparalleled in their flexibility, and uniquely allow researchers to make statements about the relative evidence for hypotheses and parameter values – and to seamlessly update those statements as data accumulate.
We wish to improve current research practices and consequently our goal is to promote Bayesian methods and facilitate their use in concrete applications. Through this blog we hope to deliver Bayesian stories and ideas that are fun, interesting, and educational.