A range of books, some excellent, have been published on the analysis of epidemiological data. While some focus mainly on the underlying epidemiological concepts glossing out the details of the underlying statistical methodology others provide an advanced course on statistics requiring a strong grounding in statistics from the reader. Statistics in Epidemiology has just the right balance between these two approaches.

The book has been developed from a graduate course in statistical methods for epidemiology taught in the School of Public Health at Berkeley, focusing mainly on the analysis of epidemiological studies with binary outcome. It progresses logically from elementary analysis to more complex logistic regression models. Alternatives and extensions to logistic regression models are briefly outlined at the end. Using the same examples throughout the book, the author helps the reader to compare and contrast the different approaches and to appreciate their similarities and their limitations. The excellent introduction to the concepts of causal graphs and their use in understanding and controlling for confounding is a great strength of the book. At the end of each chapter the author provides comments and further reading, pointing to relevant statistical textbooks and articles. The carefully selected problems and exercises (solutions to which are provided at http://www.crcpress.com/e_products/downloads) help the reader to gain an in-depth understanding.

Overall, the book is well written, comprehensive, and well structured. The author, through step by step development of the concepts, has successfully shown the hazards of going straight to a complicated analysis, before gaining a proper appreciation of the data. I personally found it very pedagogical and instructive. I therefore strongly recommend it to all graduate students and researchers in epidemiology or relevant fields. The book is a good reference for teachers giving similar courses.