Introduction to Applied Bayesian Statistics and Estimation for Social Scientists

Statistics for Social and Behavioral Sciences

by Scott M. Lynch

This book outlines Bayesian statistical analysis in great detail, from the development of a model through the process of making statistical inference. The key feature of this book is that it covers models that are most commonly used in social science research - including the linear regression model, generalized linear models, hierarchical models, and multivariate regression models - and it thoroughly develops each real-data example in painstaking detail.

387 pages, Hardcover

First published January 1, 2007

Reviews

[Silvia · Rating 4 out of 5]

This book is awesome. I can't remember having read such a well-written statistics book lately. I've read the first 6 chapters (there are 10) and I could read through them pretty quickly, while listening to music at the same time, and often found myself thinking along the lines of 'I really want to send fanmail to this author' or 'why didn't I read this *before* I read books A, B, C, (etc.) which never explained this so clearly even though they spent so many pages on it?!'

The text has a clear and logical structure, the author explains everything very well, and uses clear and concise language. There are examples throughout the text (mostly the same data is used to show the effect of a slightly different analysis technique) but this never takes away from definitions or general explanations. And the examples are easy to follow, partly because the same data is used and added to when this is possible (instead of using a wholly different example each time).

I usually skim over examples because I prefer to understand things in the abstract, instead of having to remember numbers from previous graphs etc. Often, books with many examples are a little lacking (imo) in stating clearly which things are unique to this example data and which are generally the case (or perhaps not generally, but for specific types of data, etc.). Also, they often make it necessary to keep looking back at what those data were again, or I feel slowed down by having to study new sample data for each new example. But Lynch always mentions briefly the relevant points about the data for the current paragraph, so you don't have to look back, and he doesn't come up with a whole new dataset for each example.

The author also makes it very clear, both at the beginning and throughout the text, what he does and does not cover, and why, and where additional information may be found.

As this is the first book I read about Bayesian statistics, I can't give a well-informed opinion about the depth or difficulty of this book. Of course it's intended as an introduction, and with an emphasis on applying the techniques, as the title makes clear. That does not mean that there are never any complicated-looking formulas. But I found that, even without really looking at those formulas (because I was reading this for fun, not trying to understand every detail), I could understand what was done with them and all the points made in the text. Formulas and graphs are not left to 'speak for themselves', everything is always explained in the text, and mathematical terms are described in ordinary English.

As I mentioned, I only read the first 6 chapters (out of 10), but I assume that the others will be just as well-written.

In conclusion, I really recommend this book to anyone interested in an introductory/applied book about Bayesian statistics for social sciences. It's excellent!

[Dani Arribas-bel]

Only read the first 6 chapters, but really good introduction into the basic notions. Code examples and social science context help a lot to understand ideas.

[Philip Parker · Rating 4 out of 5]

A better introduction to winbugs would have been nice. Still this can easily be found elsewhere.

[Charlien Tania]

Foolproof textbook for any frequentist student who wants to learn bayesian approach.