Multilevel Models for Binary Outcomes

In this chapterbinary outcomes in clustered data; logistic regression; the multilevel logistic model in three views, its latent-response formulation, and its graph; interpretation with odds ratios, predicted probabilities, the latent-scale ICC, and the median odds ratio; estimation by PQL, Laplace, adaptive quadrature, and MCMC, with a simulation; conditional versus marginal effects and GEE; random slopes and the scale problem; inference, fit, and diagnostics; reporting.