Regression Methods in Biostatistics: Linear, Logistic, Survival, and Repeated Measures Models (Statistics for Biology and Health)

I have owned this book for a couple of weeks. In that short time it has proven very useful to me.The authors use an easy-to-follow writing style and don't get too bogged down in theoretical, statistical formulas. It is full of useful figures that illustrate the points being made. Note: although the authors rely on Stata for creating their printouts and figures, this is not a book on how to use Stata. You don't get the feeling that you have to learn Stata in order to follow along. I have found that most of the Stata diagrams are very similar to the diagrams created in SPSS, and probably SAS and R for that matter.Although I am reading the book from beginning to end, I have already gleaned some useful information from advanced chapters, thus suggesting that it is a good reference book. For instance, I was frustrated by the lack of coverage on interpreting log transformed data (in multiple regression) in other stats books. I was pleased to discover that this book covers this issue in a clear and concise manner. I am also pleased that the authors have included a chapter on generalized linear models.This is a very good book for people working in health care research. The authors talk to the reader and explain things in a lucid manner (I have read several stats books that do not do this, so it is a refreshing change). The authors also provide many practical examples to clarify the issues. A background in the basics of statistics is required.

Regression Methods in Biostatistics is clearly a very well-organized book, covering topics from simple linear regression theory and methods, to the more complex survival analyses. The material is especially recommended for students who have just completed introductory biostatistics and statistical programming, and are looking for practical applications of their skills (of course, for those looking for more thorough practice, it is recommended that those individuals take more advanced biostatistics courses). Relevant examples are abundant throughout the chapters, and the authors are also very thoughtful in providing a website ([...]) where one is able to download the data (in all types of files) used in all the examples in the book, as well as for the practice problems. One drawback to this book, however, is the authors' reliance on only STATA to present the modeling examples; this is incredibly useful for primarily STATA users (the authors provide tips on STATA codes) but not particularly helpful for SAS users, for example (though it is certainly not a very huge learning barrier).

Overall a very excellent, broad yet detailed overview of regression and statistical methods for parsing meaning and substance from different epidemiologic and/or other health-related investigations. One caveat: the writing is extremely verbose and geared toward analytic, mathematical parsing of meaning in context of data graphical overlays. Can be understood by any functional graduate student with robust quantitative skills, but is still a bit awkward/stilted in how the information is conveyed with numbering of tables, graphs, etc., in reference to textual explanations. Other than that, kudos. Very helpful.

You can actually read this book - which is surprising given the subject. I'm a grad student taking two Biostats courses for a master's degree. This book is great and conceptual.

Love this book! provides lots of details and examples. Great refresher from the time i took courses in statistics.

not for biginner.it is very difficult

One of the best books for beginners. The text has good examples and detailed explanations

Vittinghoff is very verbose in explanations of the methods within, but this is very useful to newcomers in the field. The examples are robust and coded in a number of common statistical programming environments.

Reading this book from cover to cover for background to a MSc course.On pp52, para 2 : quoted:"The second [M-H] test result ...addresses the null hypothesis that the ... odds ratio for the association ...is different than one".Apart from the poor grammar, shouldn't this be that null hypothesis is that the OR is equal to one?It follows, that the conclusion drawn from the M-H test in that paragraph is also wrong.I'm not a statistics expert like the writers, but at the very least this paragraph is extremely confusing on a hugely important statistical point.

This is a great reference for applied health researchers. It is packed with practical suggestions and has a rich bibliography to expand on a given topic. I particularly appreciate the discussion on confounders, propensity scores, predictor selections (or non-selection) and causation. The issue of multiple testing and its perils is often brought up. Some topics appear to have been cut short: for example, when talking about collinearity or prediction.One aspect that differentiates this book from similar ones are the examples. Most topics are illustrated with examples in Stata. And there lies also the limitation: all the code is in Stata, so if this is not one's package of choice, one would miss some of the practical implementations. Nonetheless, the material is still useful and of high quality. Overall this book is a perfect complement to Steyerber (2009) and Harrell (2001).