Customer Reviews

Applied Regression Analysis and Multivariable Methods

5 star:		(4)
4 star:		(0)
3 star:		(1)
2 star:		(3)
1 star:		(3)

 

 

I have used this book from first edition in the early 1980's in grad school to lastest edition in 2003. A very good book, designed to be used with statistical software. The authors understand the field and provide a thorough yet concise perspective of mainly regression and ANOVA models. An absolute must book for the applied statistician.

I used this book for a second level statistics course for my Master's degree in Epidemiology. I liked it!

All the underlying math you want to know is sitting on the pages, clearly explained though examples with computer output and graphs. I worked through the problems in the text without difficulty and reproduced their work. I understood what I was doing. Each chapter is followed by a series of problems. You probably want to get a solutions manual if you want to check your answers.

The material covered includes: Univariable and multivariable linear regression, correlations including multiple partial, ANOVA, ANCOVA, Polynomial Regression including orthogonal polynomials, dummy variables, selecting best regression equation, and introductions to repeated measures ANOVA, maximum likelihood methods, and logistic regression.

Now that I feel that I have these basics under control, I would like a book on "approaches" to data and dealing with "difficult" data. This book contains one chapter on regression diagnostics -- not enough. But I guess that is the next step....

Other readers have commented on other books addressing the same topic, unfortunately I have not read those other books. However, I am certain that you will learn from this book, and when you are done, you will be ready for more.

(Did I mention that I signed up for a course with Dr. Kleinbaum on analysis of matched data?)

Only the questions with answers in text have solutions. Solutions not complete enough to help if you are having trouble with the concept. Very disappointing, big waste of money. Useless. Wish I could return it

The ideas put forth in this book are not articulated in a way that I found to be useful. Also, even when you do understand the concepts, the actual work can be quite challenging to do (impossible).

For instance, the homework questions are all based on output you get from SAS. The thing about that is, there is not a single line of SAS code in this book. In other words, if you don't have handouts from your teacher or prior knowledge of the SAS language, you can do absolutely nothing with the formulae gleaned from the text. These formulae are merely conceptual, to help you understand what the computer is doing for you. That is to say, the material is bound strongly to SAS, it constantly makes reference to SAS output but doesn't tell you how to make SAS DO ANYTHING. It was mandatory for me to buy this book and, in fact, it's the most expensive text I've EVER purchased. Used $170.

On top of that, the actual questions are worded in the most absurd fashion, causing much perplexity. Even worse, the answers in the appendix are of odd questions only... but get this: it only contains SELECTED answers, i.e maybe HALF of the odd questions in the book. And, trust me, you will NEED some hints.

One would be better served by copying someone else's actual homework questions and finding a book that just tells you how to do things with SAS or some other software program, you can find the formulae associated with the processes online if you need conceptual understanding.

Pair this book with an instructor who does not speak fluent English for ultimate fun....

I'm giving it Two Stars because it's the Christmas season...

I am lucky that I had such a great professor for my statistics class. Unfortunately, although he's great at teaching the material in class, he doesn't have great acumen in choosing textbooks. This book honestly made me way more confused than I otherwise would have been and is one of the worst textbooks I've ever had. First, the layout jams a lot of formulas without giving you enough specifics on how and when to apply them. Second, the authors have a talent for making straight forward stuff confusing.

Case in point is the section on assumptions about error distributions that need to be satisfied in order to do most regression analysis. Among those assumptions are equal variance and normality. These are not hard to represent in a diagram. Yet the diagram the authors provide shows an example of a case in which some, but not all, of the assumptions are violated. Furthermore, it doesn't say in the caption which ones are violated and which ones are not. In fact, it doesn't even mention in the caption that there are some violations. There's a short line in the text that tells you that. It would have been much clearer to present either a picture of a case in which all assumptions are violated, or one in which no assumptions are violated. And it would have been tremendously helpful to make that clear in the caption. As a caveat, I should mention that I used the second edition instead of the third, so this particular diagram may have been changed. But my classmates who used the third edition despised it as much as I did, so I suspect that many of my comments still apply.

I ended up finding a similar book in the library by Terry Dielman and using that instead. That book covered basically the same material but was far better organized and practical.

This book arrived with two advertising stickers attached. One was a large sticker on the spine of the book that was a third to half the length of the spine and wrapped around to the front and back covers. The second sticker was smaller but was affixed over a marring of the cover. I was able to remove the stickers, but only after some effort and there is still some residual stickiness.

Nowhere in the description was there any indication that the book would come with difficult to remove advertising. I feel the covering of defects with a sticker to be questionable.

I am very surprised by the poor rating this book received. This book is a definite keeper as a reference book on multivariate regression. Personally, I enjoy Statistics, but I even found multivariate and logistic regression difficult to grasp...especially for a "linear" and "continuous" thinker like me. This textbook has a LOT of details, and would be very confusing if you are trying to learn this on your own. One does have to spent time reading the chapters slowly and carefully, since the book is not an easy read, however, the information is excellent. I don't know how other classes are setup, but my class was taught using this textbook and SAS (SAS output), which are nicely presented in the book. When used in conjunction with good professor lectures and familiarity with SAS programing, I thought it was an excellent reference book. The problems at the end of the chapters were difficult to manually work through, but my professor used the datasets for our homework assignments (mainly using SAS software) and that works out very nicely. As a standalone book to learn from, without the aid of good lecture materials (including SAS codes) or some SAS knowledge, this book would not be a good introduction to multivariate, logistic regression etc. If you are interested in Logistic Regression "Applied Logistic Regression" by David W. Hosmer and Stanley Lemeshow is excellent. Multivariate regression and Logistics are advance Statistics and therefore the concepts are difficult grasp, explain, and teach. This book really helped me understand this concept, I highly recommend this book.

I was surprised when I took this class that I did not like this book. All of my epid profs highly recommended this text. You have to flip back and forth when you are doing the problems because they list the SAS output that you need once in one chapter. I know it saves space, but this can get maddening. I did not find the examples clearly written at all-the lack of fit test, and some of the stuff about partial F tests could have used a few more sentences.

No, I cannot write a glowing review of the text, but the only reason for the second star is because I also had a terrible instructor for this course. Perhaps if Dr. Kleinbaum had taught this to me, I would have a different perspective of this book. However, I was pretty much teaching myself this material, and this book is not designed for that.

This book is really difficult to understand because there are no real explanation. There are a ton of symbols and formulas that are not explained at all. Mostly, this book just shows how certain formulas were derived from others. Our prof did not like this book either, so she gave us her own notes...but even as reference, it is terrible!

The most complete and cristal clear exposition of multivariate analisys I ever read.