This book has proven incredibly helpful on so many occasions that I highly recommend it. Here are some examples:

Years ago, when I had a problem to deal with to improve probe yields on some major products, I presumed that a certain technique existed. I searched through several libraries, and finally found this book - and it had exactly the technique I needed.

Months later, an associate expressed frustration with his efforts to model a complex reliability problem - I showed him this book, and his eyes lit up as he saw the technique he needed.

Years later, when I needed to develop a statistical model for on time delivery, I again referred to this book to derive relationships between the standard deviation and average yields - but it also provided the information I needed on using Poisson statistics for process and assembly yield, and gave a helpful model for obtaining the cycle time distributions for a series of processes, such as from order through manufacturing and shipping to delivery of final product.

S programmers refer to this as "the white book", and it is a key reference for understanding the methods implemented in several of S-PLUS' high-end statistical functions, including 'lm()', predict()', 'design()', 'aov()', 'glm()', 'gam()', 'loess()', 'tree()', 'burl.tree()', 'nls()' and 'ms()'.

It's apparently out of print, but it shouldn't be.

Even with the recent arrival of S-PLUS releases that incorporate S version 4 and many of the ideas discussed in "the green book" (<>, also by John Chambers), this classic S reference is an indispensable tool for the serious statistician. It needs to be reissued--with a white cover, of course.

Here are the titles of the chapters, for reference:

1. An Appetizer

2. Statistical Models

3. Data for Models

4. Linear Models

5. Analysis of Variance: Designed Experiments

6. Generalized Linear Models

7. Generalized Additive Models

8. Local Regression Models

9. Tree-Based Models

10. Nonlinear Models

A. Classes and Methods: Object-oriented Programming in S

B. S Functions and Classes

References

Index

I recently started study about Pattern Recognition. This book is so well organized.

- Introduction to statistical pattern recognition

- Basic approaches to supervised classification via Bayes' rule and estimation of the class-conditional densities.

- Discriminant function approach to supervised classification.

- Techniques of exploratory data analysis.

- Additional topics on pattern recognition including performance assessment.

Especially, this book contains URL which concerned with topics. It is very useful!!

Variation is inherent in many processes common in companies and societies. Edwards Deming, one of the architects of Japan's industrial success, regarded the understanding of variation as one of the most important managerial skills. Unfortunately only few books on management treat that subject at all, and these that do, most frequently give it only a very superficial coverage.
The reviewed book is the only one I know that gives a full explanation of both the practice and the theory of statistical process control (SPC) - the way to understand variation. As a manager and TQM coach I fully recommend the book to managers and management students, as a mathematician I recommend it to students and researchers in statistics.
Readers without mathematical background should not be scared by many mathematical formulas in the book that explain why SPC works. They may very well skip the reading of equations without any loss in understanding of how SPC works and how to use it in practice.
In my opinion the book of James R. Thompson and Jacek Koronacki should be regarded as a must for any business school library and the bookshelf of any manager.

Andrzej Blikle
Professor in mathematics and computer science
Member of Academia Europaea
President of "A.Blikle Ltd."

This textbook is well-conceived and well-executed. The authors have a strong background in science as well as a keen interest in making technical subjects accessible to a broad audience. They have succeeded in presenting statistics with a minimum of mathematics. Emphasis is placed upon the application of these concepts to real life situations taken from news sources, the internet and individual experiences. The graphics, sidebars, focus sections and case studies are timely, relevant, engaging, lucidly written and enlightening. A one-page epilog provides a nice summary statement to the entire textbook. A glossary, index, answer key , appendices and suggestions for further reading are also included.
This is an excellent textbook for a general education student whose career would not require the direct use of statistical tools. None of the usual tables are discussed or present in this textbook, but all of the main topics are treated in a thorough and thought-provoking fashion. The authors' discussion of correlation and causality is especially good, although their definition of causality in the glossary as " the relationship present when one variable is a cause of another " is circular. Instead, the authors should reiterate their assertion from the main text that a "cause" is a "physical mechanism" based upon a "physical model" which is generally acceptable as a scientific explanation.

Disclaimer: I was a student of Dr. Van Ness and assisted with some minor research for the early chapters in the book.

The text deals with measurement error, ie, situations where there may be error in the measurement of the independent variable as well as the dependent. Focus is given to problem-solving and illustration of key points rather than the most rigorous mathematical proofs. One must be strongly conversant with at least undergraduate mathematical statistics to grasp most of the text.

However, it's an intriguing field with a number of real-world applications that become apparent in the text.

Ever since the legal cases against the Census Bureau in 1980, statisticians have played a much more visible role testifying and providing depositions in legal cases. Due to the adversarial role lawyers play in trials we are seeing each side pit their expert statistician against the other. Many times the statistical evidence is confusing to the jury and/or the judges and the testimony tends to cancel out with verdicts being decided by other means. It is important to keep things simple. In the case about undercount adjustment for the Census in 1980, eminent statisticians argued on both sides. There was no right or wrong answer. Everything hinged on what statistical models you are willing to believe. Unfortunately, such cases revive the old adages that make statisticians kringe, "you can prove anything with statistics" and "lies, damn lies and statistics".

With the advent of DNA evidence, statisticians are asked to compute matching probablities to determine the likelihood that a suspect is the person whose DNA was found at the crime scene. The results can be overwhelming but even a statistician with expertise in DNA matching can be tripped up by clever high priced lawyers. Such was the case when Bruce Weir testified on national television in the O. J. Simpson case.

Joe Gastwirth has contributed to the statistical research applied to legal problems over the past 20 years at least and he has published a book on the subject. In this volume, he compiles a number of case stories and statistical issues in legal cases told by many very capable statisticians including Alan Izenman, Jay Kadane, Bruce Weir, Seymour Geisser, Don Rubin, Joe Gastwirth himself,David Pollard and Scott Zeger. These are all fascinating tales that will especially be appreciated by lawyers and statisticians. But this is also worthwhile reading for the general public. Read the preface, where Gastwirth gives you a synopsis of these articles.

One of my favorites is the article by Seymour Geisser who tells a sad tale about how statistical issues relating to problems in the analysis of DNA evidence is covered up by the FBI. This is taken to the extent of influencing the refereeing process for journal publications, a shocking tale!

Unfortunately even though DNA evidence can be as conclusive as a fingerprint, human error in processing the evidence can create doubt about the matching process or even pursuade a jury that evidence was planted or a defendant frame. Such things are possible and defense lawyers now exist who are up to the task of creating such doubt as was done masterfully by Johnny Cochran and Barry Scheck in the O.J. trial.

If you have Sokal and Rohlf's Biometry book it pays to get these tables. They are referenced throughout the book and it is convenient to have them to do most of the problems.

One thing to watch out for is to make sure that the edition of the tables matches the edition of the text. I have the second edition of the book and the third edition of the tables. The authors removed eight tables in the third edition and instead of using the numbering system of the second edition, they switched to letters floowed by double letters after they ran through the alphabet.

There are many useful tables and explanations are given, so they can be used independently of the text. However, if you get the text you will want the tables. The tables are referenced in numerous interesting and instructive examples in the text.

In the Preface to the Second Edition the authors say that the tables are there for pedigogical reasons only and they chose not to include them in the text because in these days pocket calculators can often be used as replacements for tables. Nevertheless without realizing it the reader does become dependent on these tables to get a full understanding of the examples.

If you get the book get the tables also. If you just want to have a reference set of statistical tables they are useful but I much prefer the "Pocket Book of Statistical Tables" by Odeh and Owens.

I have used this book in college for my lab courses. In these courses I had to perform statistical analysis of the data. There was one little problem. I did not have formal training in statistics. Well, this book will tell you in a concise way what you need to know about statistical analysis of your data. The topic of uncertainty (Chapter 5) is extremely helful and to the point. If you want to know how to analyze your data accurately and don't have much time, this is the book you need to read.

Herbert A. David turned 70 in 1995. Much of his long career was spent as chairman of the Department of Statistics at Iowa State University. David is well known for his authoritative texts on pairwise comparison methods and order statistics.

He is known as H. A. at Iowa State so as not to confuse him with his colleague Herbert T. David who also is a Professor of Statistics at Iowa State. In fact at the end of this book Herbert T. David write a very interest review of the life and career of Herbert A. David.

H. A. David made major contributions to the theory and application of order statistics, biostatistics and the design of experiments. This is reflected in the topics chosen by the distinguished statisticians that contributed articles, most of whom are students or colleagues of David.

Noel Cressie write on a generalization of Akaike's information criterion for model selection. Dunnett talks about applications of the multivariate t distribution. Galambos and Xu discuss multivariate Bonferroni-type inequalities. Kale and Sebastian provide some interesting examples of distributions that are symmetric and have kurtosis equal to 3 (the same as for the Normal Distribution) but are non-normal. Some of the densities have very unusual shapes. These are a few of the papers under the general category of "General Distribution Theory and Inference". The articles are all entertaining and interesting and some contained discussion of David's contributions to statistics. Other anecdotes and appreciation letters were combined in the last chapter in this volume.

The volume includes six papers on general distribution theory and inference, six on the distribution theory of order statistics, five on the use of order statistics for statistical inference and applications, three on analysis of variance and experimental design and four on biometry and biomedical applications.