Building and Applying Logistic Regression Models
Chapter 6 extends the discussion of Chapter 5 on logistic regression to topics about the building and application of such models. It discusses strategies for model selection, and introduces logistic regression diagnostics such as residuals. It considers special methods of inference for stratified tables, such as Mantel–Haenszel methods. It discusses sample size and power considerations. Alternative models for binary data, the probit and complementary log-log models, also are discussed. The chapter finishes with an introduction to conditional logistic regression and exact distributions for small-sample inference.