Using Multivariate Statistics

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.

I used this text for my multivariate stats class. T & F set up each chapter for describing an individual method and then gives several examples for each statistical method described. A floppy disc with the data sets used in the text is available with the text so one can use a computer, with the appropriate programs (SAS,SPSS,Lisrel), with the text as a guide to get to the correct answers. A step by step description and interpretation is given for each problem in case you get lost in the process. This helps immensely in understanding what's going on in the problem; as multivariate stats can get real confusing after awhile. The only complaint is that the previous edition's cover had the author in a pixelated belly dancing pose - I was hoping that the obscured cover would be in focus in this edition.

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!

Tabachnick & Fidell have produced that rare thing, a technical tome that is helpful to junior and senior statisticians, and that is even fun, at times. I've used the 2nd edition of this text since it came out in 1989, during my student days and, now, professionally. It is the first place I turn when I need a clear review of data screening techniques, when I am debating the merits of one analytic approach over another, when I need to understand the fundamental formulas underlying a technique, and when I want to see how others have written up data from the technique I've selected. The structure of the book is particularly helpful in addressing these questions; there is a chapter on most of the common statistical techniques, and each chapter has sections describing the research questions for which the technique is appropriate, limitations of the approach, its fundamental equations, miscellaneous issues, a comparison of programs (which becomes dated too quickly, of course), and examples. All of these comments concern the 2nd edition, but I'll make the wild assumption that the authors have retained the structure and improved the content in the 3rd edition...and I'll buy it.

This is a wonderful book that includes examples from major software packages. It is easy to read and understand. I used this book in an independent study class for my graduate degree. Although I only met with the teacher a few times, I had a thorough understanding of multivariate statistics because of this book.