Stochastic Models with Applications to Genetics, Cancers, AIDS and Other Biomedical systems, 2 nd ed. (Tan Wai-Yuan) Silke Rolles Methodological Developments in Data Linkage (Katie Harron, Harvey Goldstein and Chris Dibben, eds) Donna Pauler Ankerst MICHAEL FRIENDLY AND DAVID MEYER. Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data. Boca Raton, FL: CRC Press.This book makes a very useful contribution by focusing on graphical methods for portraying discrete data and the results of fitting models to such data. Existing books on the analysis of discrete data pay relatively little attention to graphics. This book's main emphasis is on categorical data and logistic and loglinear models for them, but two chapters deal with count data, including dealing with the common existence of overdispersion and zero-inflation. The authors show three types of plots: Data plots, Model plots, and Data+Model plots that combine the two, showing how well the model fits the data and portraying the uncertainty of estimated response means. The book has three sections: Getting Started introduces graphical methods for categorical data, working with the data in various forms (e.g., types of contingency tables), and fitting and graphing discrete distributions with useful displays, such as "rootograms," "Ord plots," and "Poissonness plots," and extensions due to Dave Hoaglin and John Tukey for other distributions. Exploratory and Hypothesis-Testing Methods presents displays and plots for two-way contingency tables, mosaic displays for multiway contingency tables, and plots such as biplots for correspondence analysis. Model-Building Methods presents plots relevant for standard models for discrete data, emphasizing logistic regression and its extensions for multinomial data, loglinear models for contingency tables, and generalized linear models for count data and their extensions. The 11 chapters each have exercises for practicing the methods and their graphic displays.As the section outline suggests, the book covers the most popular statistical methods for analyzing discrete data. Among the types of graphical displays not presented are classification trees, graphical models for conditional independence structure, and depictions of estimates for models with large numbers of predictors (such as lasso estimates as functions of a smoothing parameter). Discrete modeling methods not covered include quasi-likelihood methods, such as generalized estimating equations for marginal models with multivariate responses, generalized linear mixed models, and Bayesian inference. But I believe it was sensible for the authors to emphasize graphics for basic methods, such as contingency table analysis and ordinary logistic regression, as the book already contains an impressive amount of material and will be very useful for most discrete-data analyses conducted by applied statisticians. Also, the authors consider many nonstandard models within these general classes, such as ordinal loglinear models and specialized mo...