This book falls into some strange no-mans-land where definitions, proofs, and the underlying statistical theory are skipped, but formulae are stated in a symbolic language complex enough that it is hard to imagine any readers that could understand the book yet not desire more mathematical rigor. The authors do give extensive references to material they don't explain--on the other hand, for $90 most readers probably want the explanations. The authors also present many of their own opinions about the relative merits of certain statistical quantities. While their opinions, based on their years of experience as researchers, may have some merit, all too often they are forced to admit that other researchers, with comparable credentials, have opposing opinions! In that case, why include the opinions _in a textbook_? Debate in scientific journals. Explain in textbooks.

The organization at a chapter level is clear enough, but within chapters the exposition is ad-hoc. This is probably a side effect of the authors' commitment to eliminating explanations and proofs--there's no need to organize since they are just enumerating formulae and opinions.

I used this book for an undergraduate class in college. It was on the bookshelves of most of my Statistics professors (older editions of course). I found it to be a very intuitive and strait forward introduction in Regression Analysis. It contains all the basics needed to understand the bread and butter of Regression.

Draper and Smith have long had a reputation for an outstanding book on regression analysis written at an elementary to intermediate level. I have long had a copy on my bookshelf and continue to purchase the revisions. They are careful to keep the book current by always incorporating new advances. This edition includes many of the recent advances in regression diagnostics as well as a description of the bootstrap approach to regression problems. Those interested in regression graphics should consult the book by R. Dennis Cook. More on the bootstrap can be found in my book "Bootstrap Methods: A Practitioner's Guide" or the other fine books by Efron and Tibshirani, Davison and Hinkley, and Lunneborg.