Bayesian Methods: A Social and Behavioral Sciences Approach, Second Edition (Statistics in the Social and Behavioral Sciences),   ISBN:9781584885627

Customer Rating: 

I bought this book because it's required for a class. It was shipped in great shape and very fast. Thank you!

Customer Rating: 

I have edited this review into two parts because amazon does not allow us to do second reviews of a book even though the second review is on a new and substantiakky reviseed edition. First i will give my review of the second edition and then I will indicate where the review of the first edition begins.

I reviewed the first edition of this book when it came out some 5 years ago or so. I liked the first edition but I like the second edition even more. Don't fooled by the subtitle. It gives the impression that this might be an elementary book designed for social science majors. Although Professor Gill does research in the social sciences and wants to popularize Bayesian methods in that community, this is far from an elementary book. In the second edition he has expanded the coverage of MCMC methods and hierarchical Bayesian models. Every technique is illustrated with examples and MCMC methods are developed in R as well as in BUGS. I paricularly liked the use of the Florida Palm Beach County voting data from the 2000 Presidential election to illustrate the Bayesian techniques and various types of sensitivity analysis.

Although many of the examples come from the social sciences the methods Gill presents are applicable to a wide variety of problems including clinical trials data and many engineering problems.

Some background in probability and statistics is needed to appreciate the advanced techniques employed. Dr. Gill includes approximately 1300 refrences in the back of the book. This is a testiment to the explosion of research papers on MCMC methods. Of all the books that cover MCMC methods I think this is by far the best as it gives a lot of good practical advice and teaches the reader how to use CODA and R to test the validity of the MCMC results. In addition to the many chapters inn the text the appendices are like chapters in themselves. One chapter from the first edition was moved to become an appendix.

My remaining remarks are the remarks I wrote about the first edition.
Jeff Gill is a statistician and a programming geek. He writes code in R and S. This book is an introduction to Bayesian methods for social scientists with the primary goal of making Bayesian methods accessible and used in that discipline.

I discovered Jeff when I took a course from George Casella on Markov Chain Monte Carlo (MCMC). Jeff helped George teach the ins and outs of BUGS and BOA and CODA all common and important tools for the implementation and understanding of MCMC. In the course Jeff presented material from examples in this book (which was not yet out when I took the course). I knew then that I wanted to get this book first chance I got!

I am a statistician and this is a great reference for statisticians and biostatisticians who are also finding Bayesian methods and MCMC very useful. The book is designed for social scientists but is good for everyone wanting to do sophisticated Bayesian analyses!

For an authoritative and more detailed review of the book I recommend you also read Andrew Gelman's amazon review of this book. He is an expert on Bayesian methods and has pulbished books on the subject as well as on the MCMC technique.

Customer Rating: 

The thing that makes this book excellent for the social sciences is not so much its examples (although these are real social-science examples, which can be hard to find in statistics textbooks) but the tone, a sort of theoretically-minded empiricism that is hard for me to characterize exactly but strikes me as a style of writing, and of thinking, that will resonate with the social science readership.

Compared to a more mainstream Bayesian data analysis book such as Carlin and Louis or our book, Gill has more on the history (addressing questions such as why has Bayes suddenly seemed to become more popular) and a lot on hypothesis testing, which is a big issue in social science, where a standard research paradigm is that falsifiable research hypotheses are set up and then put to the test.

One great feature of this book is its use of examples where real prior information is used. Not just convenient noniformative priors, but real discussion of how prior information comes in to the analysis. As a related point, summaries such as that on page 64 are particularly useful in comparing Bayesian and classical approaches to statistics. This kind of thing is great for a class: if students disagree on these things, it can spark useful discussion.

The presentation of results is largely done in a standard social science manner; for example the table on page 121 presents posterior intervals to three decimal places ([6.510:11.840], etc.), and the table on page 126 presents variable names in all-caps (EXTENT, DIVERSE, etc.). This isn't how I would do it, but it does place things closer to what is usually done in social science, which can be a virtue here.

Gill's book also has a fairly theoretical treatment of computational issues, actually more theoretical than our book, which might seem surprising (I'd think that, if anything, social scientists would be less likely to want to see heavy Markov chain theory), but makes sense for a couple of reasons. First, Gill himself does research in statistical computation and can give the readers the benefit of his insights. Second, social scientists, not being mathematicians themselves, do want to see the rigorous mathematical foundations of their methods. It's fine for me to just describe methods and sketch proofs in my book, because much of my audience is statisticians who will know where to follow the more detailed derivations if they need to, but Gill is connecting with the social science students who might not want see this anywhere else--and the good ones will want some rigor.

P.S. Given that Gill does talk about history, I would've liked to have seen a bit more discussion of the applied Bayesian work in the "dark ages" between Laplace/Gauss in the early 1800s and the use of the Gibbs sampler and related algorithms in the late 1980s. In particular, Henderson et al. used these methods in animal breeding (and, for that matter, Fisher himself thought Bayesian methods were fine when they were used in actual multilevel settings where the "prior distribution" corresponded to an actual, observable distribution of entities (rather than a mere subjective statement of uncertainty)); Lindley and Smith; Dempster, Rubin, and their collaborators (who did sophisticated pre-Gibbs-sampler work, published in JASA and elsewhere, applying Bayesian methods to educational data); and I'm sure others. Also, in parallel, the theoretical work by Box, Tiao, Stein, Efron, Morris, and others on shrinkage estimation and robustness. These statisticians and scientists worked their butt off getting applied Bayesian methods to work _before_ the new computational methods were around and, in doing so, motivated the development of said methods and actually developed some of these methods themselves. Writing that these methods, "while superior in theoretical foundation, led to mathematical forms that were intractable," is a bit unfair. Intractable is as intractable does, and the methods of Box, Rubin, Morris, etc etc. worked. The Gibbs sampler etc. took the methods to the next level (more people could use the methods with less training, and the experts could fit more sophisticated methods), but Bayesian statistics was more than a theoretical construct back in 1987 or whenever.

OK, OK, I guess that was a bit too blog-like. Anyway, Gill's book is excellent and I highly recommend it to social scientists and students who want more motivation and theory beyond what they'd get in a straight how-to-do-it applied statistics text.

Customer Rating: 

Gill's book is ideal for psychologists and others who tend to recieve minimal statistical training but want to learn (1) why to pursue Bayesian methods when p-values seem to be doing okay, and (2) how to start with those methods. You need to know about various continuous and discrete probability distributions, and also what an expectation is (I'm serious, many psychologists don't), so if all you know are the mechanics of t-tests, ANOVAs and regression, you have to buff up your theory a bit before reading this book.

Nevertheless, the book walks you through things and includes code in every chapter, which is rather helpful. Psychologists wishing to use this book will need to download R and Bugs (freely available for Unixes, Macs, and Windows), which involves programming and is not as simplified as SPSS.

Customer Rating: 

I do not usually like mathematical or statistical books "for Social Scientists". This is not a complete exception. It is very useful as an introduction to Bayesian statistics but examples and applications are neither interesting nor real problems of the discipline. Therefore, the "applicable" side of the book diminishes its value and makes it only a handbook for those with low background in mathematics.
On the other side, it includes many examples of estimation with BUGS and WinBUGS. This is clearly a book for a short introductory course on Bayesian Statistics since it soon needs of a more comprehensive and depth reference if you really want to go through the methodology.