Customer Reviews

Applied Linear Regression (Wiley Series in Probability and Statistics)

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

 

 

Weisberg's book is a good introduction/overview of the principal techniques used in linear models. However, in many situations, he leaves out many details and derivations and either directs the reader to the references or says nothing. I would have liked it if he had put in a few more mathematical derivations and had been more thorough in the discussion about collinearity and variable selection.

This book provides an excellent introduction to the application of linear regression models. Starting with the idea of studying "relationships between measurable variables", it develops the concepts behind simple and multiple linear regression, discusses interpretations of the output, proscribes diagnostics, discusses model building techniques, and touches other topics. It covers most of the material contained in Draper & Smith's _Applied Regression Analysis_, but in half the number of pages.

I would wholeheartedly endorse this book as a primary or supplemental text to an introductory course, or as a text for someone who wants to learn about linear regression on their own.

This text is not rigorous, nor does it take a cookbook approach. Little is derived or well justified. This reference will not help you think critically about the underlying methodology. The only reason to use this book is if you want a fluffy, example based approach to regression. Expect to be frustrated.

For someone with an appropriate background in linear algebra, I would recommend Davidson and MacKinnon's "Econometric Theory and Methods," which is a much more satisfying approach to regression.

Please don't waste your money on this book, especially if you have little prior knowledge regarding linear regression. It is poorly written and seldom gives proper explanations for the various aspects of regression analysis. Weisberg wrote it keeping brevity in mind, so don't expect to find many details; this book is definitely not for beginners, and even people I know with more experience have noted that it is often difficult to understand.

An absolutely terrible text. The book rarely offers anything remotely resembling rigorous derivations (almost entirely heuristic arguments), and as often as not has errors in the derivations they do present. On top of that, the R package written for this text by the author (package "alr3") contains errors as well. In order to complete the exercises provided in the text, you actually have to EDIT the provided functions.

Overall, this is an inexcusably bad book.

If it's required for a class, see if you can get it from the library, and buy Applied Linear Statistical Models for reference instead. The notation is a little different, but the notational issues are a small price to pay for actually knowing what's going on.

The explanation is hard to understand.

This book is absolutely terrible; we are using it for a graduate level course in regression, and even with the professor's explanation it is almost impossible to understand. There *have* to be a ton of books out there that are far easier to grasp and far more useful. Even if there aren't this book is worse than nothing.

I am a graduate student and take a graduate level stat class called linear model using this book as a required textbook. Unfortunately I found this book is really not a good fit. The explanation is hard to understand, and the readers can easily get confused. Not worth the price.