From the Bookshelf
Ashley Faber and Stanley Wasserman
From time to time, your editor looks over the books, some technical and some not, that have landed on our bookshelves but have not been sent out for full reviews. What follows is an annotated list of the best books that we have seen over the past two years.
Fox, J. (1997), Applied Regression Analysis, Linear Models, and Related Methods, Thousand Oaks, CA: Sage Publications.
A revision of the author's 1984 book, Linear Statistical Models and Related Methods, this wonderfully comprehensive book focuses on regression analysis and linear models. Fox introduces the idea of regression in a nonparametric, exploratory sense rather than immediately jumping into least squares or maximum likelihood estimation. The text includes a lengthy, and quite thorough, discussion of regression diagnostics. Other topics include graphical representations of data, variable transformations, use of dummy variables, analysis of variance models, and the geometry of the linear model. In the latter part, Fox introduces more advanced topics including time series regression, nonlinear and robust regression, logit models, and bootstrapping and cross-validation. Each chapter concludes with exercises, chapter summaries, and a recommended reading list. Data are also available on the Web. We enthusiastically recommend this book --- having used it in class, we know that it is thorough and well liked by students.