general

Many statistical books, particularly multivariate ones,are, - let's face it- boring, heavy-going and ugly. Formulae are used as if to terrify and frighten, not to illustrate or show, and textual passages are all too often chock full of outright fibs such as 'it will be clear that', or, 'the reader will see that' or - a sure sign that something truly monstrous is to follow - 'it is obvious that'. Not so Tabachnick and Fidell's great book. Everything is clear and well set-out, with lots of real examples. Unlike many textbooks, it does not assume that you got your data out of a textbook, it shows how to analyze real data, how to check statistical assumptions and what to do about violations of those assumptions. Although some of the information on statistical packages may now be a little dated, the program inputs and outputs are supplied, as are written-up reports of the results, similar to how they would appear in a socio-behavioral journal. I've been using the various three editions of this truly wonderful book since 1985 and think it should be on the desk of every professional, aspiring or sometime data analyst or statistician!

This is the 'statistical bible' for graduate students or professors performing multivariate statistics in their research. It explains every concept, provides examples, and explains different statistical programs. All of this information is provided in easy to understand terminology. It is worth the steep price tag because I use it all the time. As a researcher, I perform multivariate stastics frequently and this book is a great resource.

There's a reason why some students leave out the first and soften the second 't' in the word 'statistics'. Well, not if working with this book.
Doing a lot of multivariate analyzes in the course of my research and teaching research-methods to graduate students, I repeatedly find this book a treasure I would not want to do without. The advice is well organized, hands-on with guidelines to follow and explanations on how to run and interpret analyzes in SPSS, it's sufficiently detailed (though not overbearingly so) and quite complete, and clearly written and easy to understand (honestly, whoever tells you differently hasn't read any of the alternatives).
Yes, the price is scary - unfortunately, since that price is the only thing keeping this book from being a best-seller and a student-favourite. However, if you plan on doing any multivariate stats for a little bit longer than a few months (longer than your University Library will be happy to lend this book to you, that is), it pays off. This book is excellent, and often, after searching through my SPSS handbooks, my Stevens and Cohens (all very good books themselves), I return to this one, finally finding the information I need, and in the digestible format that I need. A definite keeper.