Customer Reviews

Applied Regression Analysis, Linear Models, and Related Methods

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

 

 

I am now using this textbook for a graduate statistics course. I personally do not find it to be the most accessible book for those who are not already highly schooled in statistics, linear algebra, and calculus. There is an attempt at the back of the book to introduce you to the only linear algebra and calculus you "need" to understand the book. But I think the book continues to go far beyond what is accessible to someone being introduced to this information for the first time.

I think the book is probably excellent if you are already familiar with regression, calculus, and linear algebra. However, for those who are not, I would recommend Paul Allison's "Multiple Regression: A Primer" to get OLS and Pampel's "Logistic Regression: A Primer" to get Logistic. These books introduce you to the same concepts but without all the extra stuff that most people won't use anyway.

Dr. Fox makes an excellent contribution to the student community across geographies. The text is an excellent balance between theory and practical applications of the linear regression methodology. The author is extremely clear in explaining not only simple and multiple linear regression, but also topics such as bootstraping, logistic and other regression techniques for non normal response variables. The book do not fall down near your toes: the topics are covered in a depth that is amenable for a PhD student.
It is very interesting also to look at the many side comments and suggested readings that the author introduces many times in the book. I congratulate Dr. Fox for this clear, understandable and easy to follow text.

Dr. Fox has written in a thoughtful original manner. Example, pretty much all regression books starts out with a graph of simple linear regression model statisfying all the strong assumptions that went into it. Dr. Fox starts out by showing graph of data that violates every single assumption. This is the sort of innovative and creative approach that shows what is best about this book. Dr. Fox has a deep conceptual understanding of this material.

The book doesn't get 5 stars becuase a significant flaw. Dr. Fox (or perhaps the publishers) wanted every kind of student to be able to read this book. Both students with advanced and also students with no statistical/mathematical expertise and sophistication. The result is a fragmented text. For instance, the geometrical interpretation of least squares fit is not integrated into the initial discussion (it comes 130 pages later!). If it was integrated then many of the derivations and discussions would be far simpler and intuitive. This sepration allows a student with no linear algebra background to read this text but it also wastes the time of the advanced students who have to wait for the more simpler and intuitive approach.

As a student, I have taken the opportunity to read many texts on research methodology. This book is one of the best.

I appreciated the balance between technical information and readability. The book offers the reader the ability to gain in depth knowledge about regression analysis and linear models, yet presents the information in an understandable way.

Other texts have proven too simplistic to answer important questions or too complex to understand. This book is a nice balance of both features.

This book takes an unusual start. It begin with the assumption that regression usually has to the data and illustrates how the assumption can be violated, illustrates why graphical analysis is important for data analysis and, in chapter 4, explains how to "fix" the violations of the requirement to the data, before actually starting to explain regression models. I find this unusual approach very insightful. Moreover, difficult parts are marked with an asterisk and can be left out if this is more convenient for the reader.

Although some math is required, I find this book very understandable throughout due to its focus on application. The book covers linear models and some extensions (for the large part of the book) and also Logit- and Probit models for nominal data (in Chapter 15). Chapter 16 deals with bootstrapping, and the appendices give some introduction to statistical and mathematical requirements that the book poses.

Overall, a good buy for people who apply regressions (as the title says), probably not so much for those who are in statistics, math, or econ.

This book is a useful first book in linear regression. Fox covers the basics effectively, and the book nicely complements Fox's guide to Regression in R and S Plus.

My biggest problem with the book is Fox's meandering prose style. Few of his paragraphs start with topic sentences, and the formatting of the text means it's almost impossible to figure out what the take-away points are. (It would have been nice for Fox to reverse-engineer his chapters from the end-of-chapter summaries, which are the most helpful guide to his book, and actually happen to be organized.

For a more organized and more clearly written guide to regression, see Basic Econometrics by Gujarati.

While this book is no doubt useful to students with a solid background in math and statistics, I certainly would NOT recommend it as an introduction to regression analysis. The explanations tend to be far too complex and inaccessible for most graduate students in the social sciences. Put it this way: after having spent almost $100 on this book, I've had more success finding useful explanations of some of the book's topics on Google than I have had by reading the book itself.