| Summary: | We consider a confidence parametrization of binary information sources in terms of appropriate likelihood ratios. This parametrization is used to construct tools for Bayesian belief updates and for equivalent comparisons of binary experiments. Firstly, we provide a Bayesian Update Diagram, which enables a decision maker to generate belief updates, largely without the use of formulas. Secondly, we present a Confidence-Augmented ROC Space for sharp comparisons between binary experiments for a class of balanced decision problems. The resulting confidence order is stronger than the Blackwell informativeness order. |
