Don’t buy Bayesian Data Analysis . . .

Posted on by

. . . yet!

BDA3 is coming soon, and it’s bigger and badder than ever. It’s got Gaussian processes and weakly informative priors and HMC and VB and EP and Stan. It’s got all-new material on WAIC and several new chapters on nonparametrics. It’s got the birthday data!

Screen Shot 2013-06-25 at 10.19.59 PM

I’ll have some future posts with more information, but for now I’d just like to thank Hal Stern for helping to keep this project going, John Carlin for providing the stone that got this soup started, Don Rubin for intellectual inspiration, David Dunson for getting us moving on nonparametric Bayes, and Aki Vehtari for closing, getting the job done.

Thanks also to the Front Row All-Stars and Alex, Alex, and Vince for helping me think a lot of this through.

P.S. You can order the book here and there’s an ASA member 20% discount code AJL01, giving a price of $55.96. Standard shipping is free from that website and orders will be fulfilled as soon as the stock is in the warehouse, i.e. before Amazon even has copies.

39 thoughts on “Don’t buy Bayesian Data Analysis . . .”

  1. I have the second edition, but I’m definitely interested on the third edition. Hope that the price be affordable.

    • Similar situation, bought mine at the beginning of the year. I’ll probably wait a while before getting the third.

  3. Congratulations! I remember taking my first course in BDA with the authors in the JSM-1998. I have been using the book for teaching (in the previous editions) since 2005. I am looking forward to this new edition!!

    P

  4. I pre-ordered it a few weeks ago at the US amazon.com site for $66.45. Amazon says it will list for $69, has 800 pages, and is due to be released on Nov 15.

    • I don’t know about the price, and I can’t say when the physical copies will come out (I hope it is before November!), but it will be approximately 675 pages, not 800. It’s reformatted so there’s more text per page than in the previous editions.

  5. I just pre-ordered from Amazon, and with their free shipping they claim I’ll have it on November 19. (Thanks for the tip, @dab.)

    Not sure how you achieved a higher TPP (text-per-page) than the second edition, but perhaps you used a Bayesian layout algorithm with tighter white-space priors? ;-) (And I hope it has section numbers at the top of each page, to make it easy to jump to a section. That’s my only gripe about the second edition.)

  6. You posted about parallel coordinates earlier (but comments are closed there). On SIGMOD just this week in NYC, there is a 3d variant of it. Nothing spectacular, just a demo of the basic idea as a proof of concept.

  9. Pingback: Stop, collaborate and listen | Sam Clifford

    • October? Please. My order was placed on May 18, 2011.

      It’s actually quite entertaining to get the “extended release date” emails from Amazon every few months. In fact, I just received one today that says my estimated arrival date is January 2014. November indeed! The fact that I’ve gone the past 2 years without a copy of BDA while still doing BDA tells me one thing: I do not need this book. But I’ve come too far to give up now.

      • Just got notified from Amazon that the book should arrive in November now. Obviously a good sign, after 2 years of extensions!

  11. Question: Have a background in stats from many moons ago. Steeped in some of the frequentist traditions, aware of / and have used MLE. Not a statistician by trade but want to be conversant with Bayesian methods, particularly hiearchical modeling, etc. Is your book or Kruschke’s the best place to start?

    • Doug, I found that books such as Kruschke’s and Peter Hoff’s “A First Course in Bayesian Statistical Methods” were an easier introduction to Bayesian statistics than BDA because they have a gentler learning curve and also there is R code (and BUGS code for Kruschke’s text). (I like that Gelman and Hill’s ARM text has R code.) But I am certainly planning to get the new edition of BDA. IMHO, part of the value of BDA is that it introduces particular Bayesian topics which go beyond the introductory Bayesian textbooks. For example, I look forward to the treatment of the Expectation Propagation in BDA3; a proper understanding of this topic has so far eluded me.

      Also, have a look at this link.
      http://stats.stackexchange.com/questions/125/what-is-the-best-introductory-bayesian-statistics-textbook

  13. Does it contain any information / suggestions regarding Bayesian Network Modelling?
    If so I would be very interested :)
    Thanks,
    Jennifer

  15. Why the last graph doesn’t have December and January “bookending” the main graph? That was the idea worked out on this very blog.

  16. Bad news on the ordering front. I just got this email from Amazon on my pre-order:

    We’re writing about the order you placed on June 26, 2013. Unfortunately, the release date for the item(s) listed below has changed, and we need to provide you with a new delivery estimate based on the new release date:

    Gelman, Andrew “Bayesian Data Analysis, Third Edition (Chapman & Hall/CRC Texts in Statistical Science)”
    Estimated arrival date: December 30, 2013

  17. From Amazon earlier today:

    “Hello,

    We’re writing about the order you placed on March 12, 2013 (Order# xxx-xxxxxxx-xxxxxxx). Unfortunately, the release date for the item(s) listed below has changed, and we need to provide you with a new delivery estimate based on the new release date:

    Andrew Gelman, et al “Bayesian Data Analysis, Third Edition (Chapman & Hall/CRC Texts in Statistical Science)”
    Estimated arrival date: January 02, 2014 – January 06, 2014

    We will make every effort to get the delayed item(s) to you as soon as possible…”

  18. Hello all

    I am writing on behalf of the publisher, Chapman & Hall/CRC. We are planning an expedited production schedule for the book, so are hoping to publish sometime in November. As soon as I receive the final manuscript from Andrew, the schedule will be nailed down and our website and Amazon will show the correct publication date.

    To order the book, you can visit http://www.crcpress.com/product/isbn/9781439840955 and use the ASA member discount code AJL01 to receive a 20% discounted price of $55.96. Standard shipping is free from our website and orders will be fulfilled as soon as the stock is in the warehouse.

    Regards,
    Rob Calver
    Senior Acquisitions Editor, Statistics
    Chapman & Hall/CRC

    • No e-book version? I don’t like having to carry around paper books any more.

      Out of curiosity, where does the time go between Andrew sending you the manuscript and books being available? I thought he was sending you photo-ready copy and there wasn’t going to be a round of copy editing.

    • I’d like an e-book version – preferably in kindle and google books formats. The last ebook edition was only available in e-book form as a Borders books (or maybe it was B&N) DRM pdf. I had to keep a copy of adobe’s crappy digital editions software on my computer just to read it.

    • I want ePub format.

      Moreover since I already own the previous edition, I’d like to buy an “upgrade”, not the full version.

        • I once did something similar for an online report. In the old days we simply had pictures and recordings for all possible parameterizations, so no actual processing was done in the browser…

          I doubt an iPad can run Stan but Stan could sit in the cloud. More generally, BDA could be more like software with patches for errata, updates, upgrades, chats, Stack Exchange support, associated MOOC, etc…

          You can push this really far.

        • Fernando:

          I know what you’re saying but I think a book has something to offer as well. I find it useful to be able to flip back and forth and find things in physical space.

          Regarding the mook, we don’t yet have that. But I do have slides from when I last taught the class. I’m revising them to fit with the new edition and then will post them on the web to help others who will be teaching classes using the book.

        • Of course I agree with you. I am a book sniffer, and eBooks don’t have scent.

          But is nice to imagine new things.

    • I agree with others. I won’t even buy physical books anymore. Bad for the environment and bad for my back. I work both from my home and my office. I’m not lugging physical books between the two. I also travel a lot, and refuse to carry a bag full of heavy books with me. Plus, I annotate my books and often look back at them after months or years. It’s great to always have them available in the cloud for reference wherever I am. With so many Bayesian books available now, I would easily buy a slightly inferior book if it was available in an ebook format over a superior book that was only available in physical form. If you want to do other versions please go ahead, but at least do a Kindle print replica version as the print replicas formats work well for technical books.

Comments are closed.