Bayesian Data Analysis

by Andrew Gelman, John B. Carlin, Hal S. Stern

Bayesian Data Analysis describes how to conceptualize, perform, and critique statistical analyses from a Bayesian perspective. Using examples largely from the authors' own experiences, the book focuses on modern computational tools and obtains inferences using computer simulations. Its unique features include thorough discussions of the methods for checking Bayesian models and the role of the design of data collection in influencing Bayesian statistical analysis.
Bayesian Data Analysis offers the practicing statistician singular guidance on all aspects of the subject.

Genres: Mathematics, Textbooks, Science, Reference, Nonfiction, Academic, Technical

552 pages, Hardcover

First published January 1, 1995

Reviews

[Hyokun Yun · Rating 4 out of 5]

My impression on this book from what people around told me before actually reading it was that this book is the canonical textbook for those who want to get into Bayesian statistics. After having read this book from cover to cover, however, I do not think it is a good idea to start learning Bayesian statistics with this book, as it covers very wide range of topics and therefore does not get into much technical depth for most of them. I think this book is ideal for someone like me who has very basic understanding of Bayesian statistics but would like to get some exposure to a variety of existing tools in the literature so that when some of them become needed at certain point of my career, I can reopen this book and follow its references to learn enough to actually use them.

The stance of this book is very practical, and it is great to get a glimpse of how these grand-master Statisticians approach data analysis. First few chapters regarding the underlying philosophy of Bayesian statistics is also very insightful.

I was somewhat disappointed with changes in the third edition though. The addition of Gaussian Processes and other advanced topics is conspicuously advertised here and there, but I found these new chapters to be relatively poorly written compared to those from the previous edition; notations are not consistent with previous chapters, and clarity of writing is disappeared. It was a stupid idea to buy the new edition while having the second edition.

[Michael Culbertson]

A canonical reference text. For a more friendly and accessible introduction, I'd recommend Statistical Rethinking: A Bayesian Course with Examples in R and Stan.

[Radovan Kavický · Rating 4 out of 5]

This book's biggest strength is its introduction of most of the important ideas in Bayesian statistics through well-chosen examples. These are examples are not contrived: many of them came up in research by the authors over the past several years. Most examples follow a logical progression that was probably used in the original research: a simple model is fit to data; then areas of model mis-fit are sought, and a revised model is used to address them. This brings up another strength of the book: the discussion and treatment of measures of model fit (and sensitivity of inferences) is lucid and enlightening.

[Tinwerume · Rating 4 out of 5]

I just skipped around doing exercises as a self-test. They're not particularly difficult, they're reasonably comprehensive, and there's an (incomplete) solution set. A lot of them are computational, which is useful if you're an undergrad but I would skip them if you're already familiar with data analysis.

Main text is less rigorous than I'd like (there are *some* proofs in the appendices but they're kind of lacking), but it seems fine as a first text.

[Gavin · Rating 3 out of 5]

Ongoing. Made it to Chp 12 this time.

Disappointingly dry (compare his extremely fun and salacious blog), but wise.

[Terran M · Rating 4 out of 5]

This is a challenging but rewarding book on Bayesian statistics. Before you get to any kind of computerized methods, you're going to have to get through a substantial amount of somewhat tersely presented calculus with conjugate priors; this level of rigor is both the strength and weakness of this presentation. If I could do it all again, I wouldn't read this book first - I'd read Kruschke, and then Gelman and Hill, and then come back to this book.

n.b. they're on the 3rd edition now

[Phyllis · Rating 4 out of 5]

This is the textbook for my Bayesian Data Analysis book. This book contains lots of real data analysis examples, and some example are repeated several times through out the book, for example a 8-school SAT score example appears in both single-parameters models and in hierarchical models. I really like how this book teaches you how to solve real problems.

[Bing Wang · Rating 5 out of 5]

Skipped the last chapter.

Very detailed bayesian tutorial. Part I, II and IV are more practical for everyday's work. Should combine this text book with some practical coding exercise together. Such as pymc3 in python.

Help to organize/connect/compare many knowledge knots like mixed effect model vs hierarchical model, kriging vs gaussian process.

[Kirill · Rating 4 out of 5]

Finished this book as a part of BDA course. Is is a dense and readable overview of modern Bayesian data analysis. 4 stars rating is mainly because some chapters of this book need an update because they use outdated computational approaches for which we have better alternatives now. Nonetheless, it is still one of the best resources to get knowledge about the topic.

[Huong · Rating 4 out of 5]

More of theory and general example. It would be great to have more coding coach.

[Bryan · Rating 5 out of 5]

The crown jewel of statistics books

[Ngaafare Ricardo · Rating 5 out of 5]

Addicted to numbers now

[Sean · Rating 5 out of 5]

Greatest book to read for anyone who wants to learn about Bayesian statistics, Andrew Gelman is the foremost expert in this.

[Ian]

good followup to casella and berger

[Dylan · Rating 5 out of 5]

It feels weird to add a textbook onto goodreads, but I'm making an exception because I owe this book my gratitude, and I'm paying it back with 5 stars. Used this to prepare for applied work, and I'm very grateful. It's definitely introductory, prioritizing breadth over depth, but it does a great job at getting you situated, so that you're ready to read other material. Despite it being introductory, my guess is that it's intended for those with other prior statistical background (but I can't judge that personally).

Definitely supplement by going through the backlog of Gelman's blog, particularly [this list](https://statmodeling.stat.columbia.ed...), all that will help flesh out his view of applied work (and how it differs from some others, but I quite like Gelman's approach).

[Tobias Wolfram · Rating 4 out of 5]

In the end I'm a bit torn. Everybody describes it as the bible of bayesian statistics and it indeed covers a wide range of topics and by that supersedes all other general textbooks on bayesianism I know of. A great amount of literature is given after each chapter for everyone who wants to learn more about specific stuff. As a first time introduction on the other hand it's an incredible bad choice - if you don't know the basics you are going to have a bad time.

[Jerzy · Rating 3 out of 5]

Gelman has complained about sloppy notation in other books, which is weird because the notation here seems extra sloppy -- no bold or anything to differentiate between scalars and vectors, which can be a pain sometimes. But if you already have at least a tiny bit of exposure to basic Bayesian stats in the usual notation, this seems to be a good and accessible way to expand your knowledge.

[Jim N · Rating 3 out of 5]

challenging for a poor mathematician

[Lurino · Rating 5 out of 5]

the definitive textbook for bayesian statistics. almost everything is here, but it's definitely not a gentle introduction.

[Nick · Rating 4 out of 5]

classic. must have. I have owned several editions.

[Ram Das · Rating 5 out of 5]

An excellent book for bayesian analysis

[Haseeb Shehzad]

The best on the topic in the market.