31 of 35 people found the following review helpful:
2.0 out of 5 stars
Not an S-PLUS user? Go elsewhere-, October 30, 2000
By A Customer
This review is from: Introduction to Robust Estimation and Hypothesis Testing (Statistical Modeling and Decision Science) (Hardcover)
Dr. Wilcox apparently works copiously but unaccompanied (36 of the book's technical references attribute sole authorship to him). With all due respect to the author, lack of collaboration can be detrimental to what was intended to be a basic technical treatise - and this may be one of those occasions. The author notes (p. 11) "We stress, however, that many mathematical details arise that are not discussed here. The goal is to provide an indication of how technical issues are addressed without worrying about the many relevant details." Indeed, the reader may find this concentration of effort a bit contrary with the introductory nature implied by the book's title. This impression is reinforced by a subject index perhaps too terse for an introductory or reference textbook (being only 3 pages of regularly-sized font, while a less-useful "author index" takes up almost 4 pages).
For the most part, this lean 296 page book is not so much an "Introduction to Robust Estimation" as it is a tutorial or user's manual for numerous S-PLUS functions relevant to the subject matter. If the reader is not heavily invested in S-PLUS (S-PLUS being a high-level computer language and interactive analysis environment trademarked by MathSoft, Inc.), he can never fully appreciate the contents of this title. For example, in the discussion of median variance (p. 42), the author notes that this estimator is related to the beta distribution, but does not acknowledge that there are several related functions that can take this name (such as the incomplete beta distribution, as well as its ratio). Instead, an S-PLUS function 'pbeta' defines what was meant. One must therefore resort to cited third-party references or a computer to really grasp the basics in these situations (in this case only to discover the terminology was inaccurate). A fundamental reliance on propriety software packages and professional journal articles for basic instruction and accuracy is a characteristic unbecoming of an "introductory" textbook, in my opinion.
Often, software manuals tied to specific libraries or languages become dated. I liked the fact that the companion software was downloadable, rather than provided on a medium that might be incompatible with the user's operating system, such as a 5 1/4" disk (the link in the textbook has been updated to www.apnet.com/updates/ireht.htm). For an S-PLUS owner already familiar with robust statistics, this book would probably rate higher. However, of the four textbooks I currently own on this subject, I regret to say that this title only sees infrequent use. A better alternative for emphasizing basic concepts and theory is "Robust Estimation and Testing" by Staudte & Sheather (ISBN 0471855472).
Help other customers find the most helpful reviews
Was this review helpful to you? Yes
No
5 of 6 people found the following review helpful:
4.0 out of 5 stars
Great book for applied researchers, January 6, 2001
By A Customer
This review is from: Introduction to Robust Estimation and Hypothesis Testing (Statistical Modeling and Decision Science) (Hardcover)
If you want a useful book on modern ANOVA and regression methods, this is an excellent choice. The book begins with some technical background on robust techniques and then focuses on how commonly encountered problems can be addressed. Included are two-way and three-way designs, split-plot designs, a variety of robust regression methods plus some recently developed exploratory tools. The book also contains a description of s-plus functions that can be used to implement the methods covered. The s-plus functions used can be downloaded for free. S-plus is easy to use and learn and I make use of these functions on a regular basis. If you are more interested in theory than applied work, other books, such as Staudte and Sheather, or Huber, or Hampel et al. will be more interesting. But for applied work, this is the only book I've read that covers many experimental designs that are typically encountered in practice. Even some modern advances related to least squares are covered and can make a substantial difference in accuracy as well as the conclusions reached.
Help other customers find the most helpful reviews
Was this review helpful to you? Yes
No
