BAYESIAN HYPOTHESIS TEST AND BAYES FACTOR

Bayes factors are the cornerstone of Bayesian hypothesis testing. In contrast to classical p values, the value of a Bayes factor has a direct interpretation in terms of whether or not a hypothesis is true: It represents the factor by which data modify the prior odds of two hypotheses to give the posterior odds. Unfortunately, the values of Bayes factors often depend on the prior densities assigned to the model parameters inherent to null and alternative hypotheses. In addition, the calculation of Bayes factors usually involves the evaluation of high dimensional integrals. For these reasons, Bayes factors are employed less frequently than Classic hypotheses are. This paper provides a brief review of Bayesian hypothesis testing.