Data Analysis Using Regression and Multilevel/Hierarchical Models (Analytical Methods for Social Research) Paperback – Bargain Price, 14 Jun. 2007

I have been studying a lot on statistical modelling in the last few years and this is the best textbook I have ever come across. To cut the long story short, it is the first time I have seen fully explaining choices for coefficients (eg why, even though a coefficient is non-significant, would a modeller keep it? Most of the other books / articles just rely on the metrics like AIC etc) and other details like the use of "offset", overdispersion etc.It is a very much needed contribution to statistical modelling both for explanation and for prediction. Thank you.

This book has to be one of the best, if not the best, book upon statistics that I have ever read. Throughout I found the descriptions of the models and the mathematical explanations clear and easy to follow. The book also builds up nicely, first discussing classical regression before going onto describing multi-level models and ultimately to fitting them using Winbugs. Also, the book explains how to interpret the co-efficients one gets from models. I really enjoyed this as some stats books do not really explain these numbers to the reader properly in my opinion.Furthermore, as all the analyses are carried out in R, with scripts provided in the book, even seemingly complex models can easily be fit by R users. The ability to call winbugs using R is also a huge bonus and will aid researchers seeking a more Bayesian approach to statistics. For those who do not use R details of how to fit the models in other statistical packages are given in the appendix. My only criticism of the book would be this: Some of the R scripts available on the website are a bit messy and it can be difficult to find specific bits of code.Overall I felt compelled to give this book five stars because it has taught me more about statistics then any other book on the subject.

My old professor of linear statistics used to rage against the emergence of microcomputers that allowed any fool to run a regression. Gelman and Hill take this a step further: They actually encourage readers to upload some data and run commands in R or WinBugs -- without properly understanding what is going on inside those commands, without properly understanding the mathematical structure of the estimators or the models estimated. Gelman and Hill thus train a new generation of inexpert applicators of ill-understood statistical tricks.The book is unbalanced. The title has regression and multilevel models at equal footing, but much more space is devoted to the regression models. There are better textbooks on regression analysis. An introduction to multilevel models this book is not.The book is poorly organized. The reader has to wade through rafts of examples, often taken from the rather mundane work in political science by the first author, to find the occasional, unclearly-marked methodological point.

I read this book looking for an accessible and comprehensive treatment of multilevel models. The topic of social science appealed because this area offers different examples yet has used multilevel techniques widely. See Bryk and Raudenbush for example. The issues I have with this book is that it is over long. It could easily have been made shorter. The writing is often terse and there are clearer ways to put the point across.The real weakness with the book is the website and support materials. The website is very poorly organized and although the authors provide some R examples and code their instructions for getting up and running with these packages to work through the examples is far from clear. Contacting the authors is similarly a waste of time.If you want to understand how to run regression models with R the best book is John Fox's The R and S-Plus Companion to Applied Regression. Whose website and customer facing skills are also markedly superior. For getting into multilevel models fox's appendix on his website is well worth reading. To understand Longitudinal models Bryk and Raudenbush or Singer book is better written.I like the idea of using R and Winbugs but the authors just haven't packaged the practicalities well.