|
| Author(s) |
Montgomery, D.C. Peck, E.A. |
|---|---|
| Title | Introduction to linear regression analysis (Second edition) |
| Publisher | Wiley |
| Year of publication | 1992 |
| Reviewed by | Anatoly Zhigljavsky |
The title perfectly corresponds to the contents of the book which is actually an introduction to linear regression analysis. It touches the following topics: basics, testing for model adequacy, diagnostics for departures from the standard regression assumptions, multiple linear regression, polynomial regression, quantitative factors, variable selection, multicollinearity and some others. However, a number of important questions such as the use of a prior information and the analysis with missing data are out of the scope of the book.
In fact, quite a lot of high-grade books devoted to linear regression analysis have been published during the last 20 years, e.g. N. Draper, H. Smith, Applied regression analysis, Wiley, 1981; S. Weisberg, Applied linear regression, Wiley, 1985; S. Chaterjee, B. Priece, Regression analysis by example, Wiley, 1977; G.A.F. Seber, Linear regression analysis, Wiley, 1977; H. Toutenburg, Prior information in linear models, Wiley, 1984.
Compared to them, the present book has neither an outstanding level nor a specific style or topic selection. Furthermore, the authors seem to be not very active researchers in the field. I explain the publication of the book by the two main reasons: (i) the first edition had been well accepted by the statistical community, and (ii) there is a stable demand on good simple statistical books.
The book has been designed as an upper-level college textbook and, undoubtely, it can be useful for students of various specialities (I exclude the mathematical specialities from the list.) It seems that the book will be also well accepted by engineers, econometrists and many others dealing with applied regression analysis. This is due to that the exposition is very clear and simple, and the book has an index as well as quite a good updated list of references.
The book is very well produced, it is not expensive for the libraries collecting statistical books. A very simple decision of mine is to recommend purchasing the book to such libraries. As for the individuals, my advise is to compare the book, before choosing it, with the books on linear regression analysis which they have in their personal libraries.