By writing this book, I have divided the women and men who have contributed to the field into two groups: those whom I mention and those whom I do not mention. The first group may object to the fact that I have tended to describe only a small part of their work. The second group may object that I have mentioned nothing at all about their work. A decent respect for the feelings of both groups requires that I explain my methods of choosing what to describe and what to leave out.
The first group of omissions is due to the fact that modern science has become too extensive for any one person to know about all its ramifications. As a result, there are areas of research in which statistical methods have been used but of which I am, at best, barely aware. In the early 1970s, I ran a literature search on the use of computers in medical diagnoses. I found three independent traditions. Within any one of them, the workers all referred to one another and published in the same journals. There was no indication that any of the scientists in each of these groups had the slightest knowledge of the work being done by the other groups. This was within the relatively small world of medicine. In the wider range of science in general, there are probably groups using statistical methods and publishing in journals that I have never heard of. My knowledge of the statistical revolution results from my reading in the mainstream of mathematical statistics. Scientists who do not read or contribute to the journals I read, like the engineers developing fuzzy set theory, may be doing notable work, but if they do not publish in the scientific/mathematical tradition of which I am aware, I could not include them.
There are omissions even of material I do know about. I did not set out to write a comprehensive history of the development of statistical methodology. Since this book is aimed at readers with little
or no mathematical training, I had to choose examples that I could explain in words, without the use of mathematical symbols. This further limited the choice of individuals whose work I could describe. I also wanted to keep a sense of connection running through the book. If I had had the use of mathematical notation, I could have shown the relationships among a large collection of topics. But without that notation, the book could easily have degenerated into a collection of ideas that do not seem to have any connection. This book needed a path of somewhat organized topics. The path I have chosen through the immense complications of twentieth-century statistics may not be the path that others would choose. Once chosen, it forced me to ignore many other aspects of statistics, as interesting as they might be to me.
The fact that I have left someone out of my book does not mean that that person’s work is unimportant, or even that I think it unimportant. It only means that I could not find a way to include that work in this book as its structure developed.
I hope some readers will be inspired by this book to look into the statistical revolution more thoroughly. It is my hope that a reader may even be inspired to study the subject and join the world of statistical research. In the bibliography, I have isolated a small group of books and articles that I think are accessible to the nonmathematician. In these works, other statisticians have tried to explain what excites them about statistics. Those readers who would like to examine the statistical revolution further will want to read some of these.
I wish to acknowledge the efforts of those at W. H. Freeman who were instrumental in producing the final polished version of this book. I am indebted to Don Gecewicz for a thorough job of fact checking and editing; to Eleanor Wedge and Vivien Weiss for the final copyediting (and further fact checking); to Patrick Farace who saw the potential value of this book; and to Victoria Tomaselli, Bill Page, Karen Barr, Meg Kuhta, and Julia DeRosa for the artistic and production components of the effort.