Thursday 5 July 2018

Statistical Revolution in the 20th Century



David Salsburg (2001), The Lady Tasting Tea. Henry Holt Paperbacks, New York, pp. 340

This book is a tour de force of the statistical revolution in the twentieth century that transformed science from a deterministic branch of inquiry into something more probabilistic as we know it today. In the introductory chapter, Salsburg paints a picture of nineteenth century science that introduced a ‘clock-work’ universe in which a small number of mathematical laws were used to describe reality and predict future events using a set of formulas and their associated measurements. This ‘god-less universe’ was a fundamental disruption from the earlier conceptions of creationism and shook popular culture into understanding and accepting its temperament. Gradually, into this new universe, ‘error functions’ were introduced when reality deviated from the predicted models or when human mistakes crept in. The addition of error function and probabilistic thinking about the world around us was the major contribution of ‘statistical’ thinking in science that later led to major developments including the development of computers on the one hand, and public policy tools such as census on the other.

Lives and Works of Statisticians
This book tells the story of how statistics transformed the philosophical foundations of science in the last century, through the lives and works of major statisticians. On the criterion of selecting the scholars described in the book, Salsburg confesses that he chose major contributors of mathematical statistics who are accessible non-mathematically to a lay audience. In 29 chapters, the author has given a bird’s-eye view of pioneers such as Galton, Pearson, Fisher, Gosset and Neyman, sympathetic portrayal of extraordinary lives of geniuses and polymaths in the likes of Kolmogorov and Tukey and unearthed the stories of quiet but significant contributions of Samuel Wilks, Isidore Good and FN David. An entire chapter deals with the contribution of women statisticians in addition to their stewardship that comes through many chapters throughout the book.

The book builds up anticipation through its riveting narrative of the most well-known figures in history. We begin in Galton’s laboratory and see him at work in regression and skew distribution followed by Karl Pearson as he expands and sets up great institutions such as Biometrika and an entire school of philosophical thought. Then, enter Gosset and his ‘student’s T-test’ till the imitable Fisher comes in to break old orthodoxies and liberate many strands of inquiry. The gentle giant Neyman and the extra-ordinary foresight of Kolmogorov, especially his philosophical questions, are dealt with sensitivity and understanding. One thing I found missing was a last chapter that could have shown the way ahead in terms of exciting new work done by young scholars and the directions that the discipline in taking in the 21st century.

The book is at once a collection of brief biographical sketches of statisticians, broad-brush history of institutions and a short reference guide to important academic work in modern statistics. Written with sympathy and erudition, Salsburg’s work is a warm introduction to both the science-history buff as well as to anyone about to embark on a serious academic journey in statistics.