Summary
This document is a technical report that outlines a procedure for comparing two receiver operating characteristic (ROC) curves in a paired sample context. It describes the scenario where each subject has a known condition value and test scores from two diagnostic tests. The report details the generation of various statistical measures, including difference tests, equivalence tests, non-inferiority tests, and confidence intervals for the difference in the area under the ROC curve. It explains the methodologies for ROC curve analysis, specifically focusing on empirical (nonparametric) and binormal ROC curve estimation. The document also discusses the interpretation of ROC curves in diagnostic testing, emphasizing the importance of the area under the ROC curve (AUC) as a measure of diagnostic accuracy. The report includes formulas for estimating AUC using both empirical and binormal methods, as well as variance calculations for these estimates. Overall, it provides a comprehensive framework for researchers to compare the diagnostic performance of two tests using ROC analysis.
