January 2005, Volume 1
New from Donald J. Wheeler
January 2005, Volume 1
New from Donald J. Wheeler
Traditional textbooks fail to provide an overall approach for the analysis of data, while typical training classes focus on techniques without telling you how to choose between those techniques for a particular analysis. Dr. Wheeler's Guide to Data Analysis is the remedy for both of these problems. Here, the various techniques are organized according to the type of analysis problem, making it easy to select an appropriate analysis technique. Moreover, this is the first text to integrate the techniques of SPC with the traditional techniques of statistical inference. By placing these various analysis techniques side by side and using them on the same data sets, the reader can see how to gain the maximum insight with the least effort.
An Interview with Dr. Wheeler regarding his new book, Guide to Data Analysis.
You said that this book has taken you longer to produce than many of your other books. Why is that? The scope of this book is broader than that of most of my other books. This book goes into traditional areas of statistical inference that are not even mentioned in my other books, and considers these traditional areas from a perspective that is unique. This broader scope made it necessary to do more preparatory work, and to develop ways of integrating these various topics into a coherent overall package that would be useful to the readers. What is it that will make this book useful to your readers? This book differs from traditional statistical textbooks from the very first chapter. There I begin with a contrast between the way in which statistical theory is developed and the way in which data are analyzed. Since this book is for practitioners, it uses the data analysis approach rather than the traditional, developmental, approach. I think that most readers who have studied statistics out of the traditional textbooks will find this shift in perspective extremely helpful. One of the reviewers commented that he wished that he had had this material when he was writing his master’s thesis. What are some of the other differences? Another way in which this book is different is its organization. Rather than following the development of descriptive statistics, probability theory, and statistical inference, this book begins with a section that is closer to exploratory data analysis than it is to inference. Once this foundation has been laid, it proceeds to present techniques for analyzing data according to the nature of the problem. For example, what type of response variable is being used?, Were the data collected under one condition? two conditions? or three or more conditions? Then, for each of these various problems a series of analysis techniques are given and illustrated, with their strengths and weaknesses noted. So this book is organized from the point of view of the user. This will help the user choose a technique for data analysis, but how will the user know what problem to tackle? Perhaps the most profound difference in this book is found in the third part. After providing the reader with a reasonably complete tool kit, this part of the book outlines how to use the techniques effectively. The data analysis approach used in this book provides a foundation for a coherent philosophy of data analysis. Rather than leaving the reader with a grab-bag of tools, part three provides the reader with the big picture. How does this book do this? Specifically, Part Three begins with a systematic way of characterizing the status of a process. Following this, the traditional measures of process performance and capability are used to determine the potential savings to be obtained from different improvement efforts. By using monetary units, rather than statistical values, you will find it much easier to work on the right projects. Where does the conversion to monetary units come from? The conversion from process status to monetary units is illustrated in Chapter 14 and explained in Chapter 15. It is based upon the quadratic loss function, which is a widely accepted idea that has proven its worth in all kinds of applications. The mathematical foundation is sound and many companies have demonstrated that the conversion works in practice. I call this conversion the Effective Cost of Production. Are there any other aspects of this book you would like to mention? Yes, an extension of the Effective Cost of Production results in a rigorous explanation of why you should want to operate in the Six Sigma Zone. Moreover, unlike other explanations, this explanation provides you with a roadmap to achieving this happy state of affairs. There has been some controversy about some aspects of Six Sigma programs. What does this book have to contribute on these topics? Chapter 17 deals with several problems by discussing why they are problems and providing solutions to those problems so that they no longer need to be an obstacle to effective data analysis. Guide to Data Analysis is available now!