Modelling binary data. 2nd ed.

*(English)*Zbl 1041.62058
Chapman & Hall/CRC Texts in Statistical Science Series. Boca Raton, FL: Chapman and Hall/CRC (ISBN 1-58488-324-3/pbk). xiii, 387 p. (2003).

This is the second edition of a book on binary data, which has become a standard text for practitioners and students in need of a straightforward and clear introduction to regression modelling of binary data. Since the first edition, which appeared in 1990, the field has benefitted from methodological and computational advances, mainly mixed effects models and exact methods of inference, and this is reflected in the new edition.

The focus is on traditional logistic regression modelling in the first six chapters, with emphasis on model checking and strategies for model building. Chapter 7 discusses specific problems in modelling data from epidemiological studies, like measures of association, confounding, interaction, and cohort studies. Case-control, and matched case-control studies are introduced.

Chapter 8 presents mixed effects, or multi-level, models for binary data, and chapter 9 introduces the methodogy, based on conditional likelihood, for exact inference in logistic regression. Chapter 10 wraps up the theory by discussing some topics not covered in earlier chapters, like ordered categorical data, proportions and percentages, binary time series, and multivariate binary data. The final chapter covers computer software suitable for modelling binary data. All data sets used in the book are available in electronic form via the world wide web.

The focus is on traditional logistic regression modelling in the first six chapters, with emphasis on model checking and strategies for model building. Chapter 7 discusses specific problems in modelling data from epidemiological studies, like measures of association, confounding, interaction, and cohort studies. Case-control, and matched case-control studies are introduced.

Chapter 8 presents mixed effects, or multi-level, models for binary data, and chapter 9 introduces the methodogy, based on conditional likelihood, for exact inference in logistic regression. Chapter 10 wraps up the theory by discussing some topics not covered in earlier chapters, like ordered categorical data, proportions and percentages, binary time series, and multivariate binary data. The final chapter covers computer software suitable for modelling binary data. All data sets used in the book are available in electronic form via the world wide web.

Reviewer: Göran Broström (Umea)