ABSTRACT
This chapter describes diagnostic tests that are binary: a positive test indicates that the person has the disease, and a negative test indicates that the person does not have the disease. It suggests that there are only two states of the disease in question, namely diseased or non-diseased, and that the true disease status of each subject is known. The chapter considers agreement between a diagnostic test on one side, and a gold standard—representing the truth—on the other side. It focuses on agreement for categorical data. Cohen’s kappa—which is a measure of agreement between two raters—is by far the most used single measure of agreement in this context. The chapter also describes missing data mechanisms. It analyzes the data with both equally spaced scores and the mid-interval scores {0.5, 5, 17, 39} $ \{0.5, 5, 17, 39\} $.
