|
Abstract:
|
A frequentist describes the data (or outcomes) conditioned on the unknown parameters while a Bayesian describes the parameters conditioned on the observed data. The two approaches are shown to complement each other in two applications, namely, (i) diagnostic test kit, and (ii) hypothesis testing. In the first application, the frequentist evaluates the sensitivity and specificity, while the Bayesian considers the positive and negative predictive values. In the second application, the frequentist controls the type I error rate while the Bayesian evaluates the false discovery rate. Recently, Ng (2007) argues that simultaneous testing for noninferiority and superiority increases the false discovery rate for superiority, although the type I error rate is controlled.
|
