The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century
There is a fascinating book that I don’t think has been written yet, on the social history of statistics. This book isn’t it.
The Lady Tasting Tea by David Salsburg is entertaining for a geeky value of entertaining, but it doesn’t live up to its subtitle. The book is a set of biographical sketches of the people who pioneered statistical techniques such as analysis of variance, significance tests, sampling methods. It mentions a few sentences on the impact of stats on experimental design, clinical trials, epidemiology, and other scientific topics, but doesn’t go into any depth on the impact on scientific practices or discoveries.
The anecdotes about the careers of the pioneering statisticians raise interesting questions about the relationship between statistics and modern industry. The book’s heroes work in agriculture, measuring yield and pesticides; industrial process control, monitoring the production of beer and cotton. They also contributed to social policy, working out theory and politics of eugenics; measuring economic activity for the new deal.
It’s interesting reading these stories in the context of the debacle of modern industrial food production. Controlling the variations in batches of Guinness led to the bland hegemeny of Budweiser; the study of yields in England led to the pesticide and fertilizer treadmill, soil decline, and big dead zones in the ocean. Chester Bliss’ pesticide experiments showed that at any dose of pesticides, some bugs survived. In the short run, his calculations led to effective doses; in the long run, to pesticide-resistant bugs. Incremental progress and quality control that seemed so rational and positive turned out to have counterproductive results.
It’s interesting that the heroes of Salsburg’s book are so obscure compared to the scientists, entrepreneurs, and policymakers of the last two centuries. Statistics appears as the servant to science and politics and industry. In the 20th century there’s an aspect of Tom Lehrer’s apolitical rocket scientist Werner von Braun — “once the rockets are up, who cares where they come down.”
The story I’d really like to read would be a comprehensive social history of statistics — the relationship between statistics and the evolution of modern society and industry. There are some interesting-looking books about Eighteen Century public health and the emergence of statistical thinking in the 19th century. How did the trends continue into the 20th century? A number of Salsburg’s subjects created departments of statistics in the mid-20th century, presumably to meet a growing need. It would be interesting to see a graph of where those students went to work in industry and government. Was the answer just “everywhere”? Or is the adoption of statistical methods uneven, and does this tell any interesting stories?
Summary: Salsburg’s book adds some interesting biographical spice to names and terms that many people know only from menu items in math programs. But don’t expect in-depth history.