Statistics on the Table: The History of Statistical Concepts and Methods (review)

Statistics on the Table: The History of Statistical Concepts and Methods. By Stephen M. Stigler (Cambridge, Mass., Harvard University Press, 1999) 488 pp. $45.00.

Stigler has collected a wonderful set of essays written over the past quarter century, revised and supplemented them with new work, and provided a rich set of stories, lessons, and vignettes in the history of statistics. He takes his title from a challenge by Karl Pearson to critics of his 1910 study about the effect of parental alcoholism on their children. Ethel Elderton and Pearson had collected data from the Edinburgh Charity Organization Society and a school for the "mentally defective" in Manchester to test the widely held notion that the children of alcoholic parents were affected physically and mentally by their parents' alcoholism (16-17). Elderton and Pearson could not identify any effect. The published results were met with disbelief, and a firestorm of criticism from economists, medical experts, and temperance advocates. Pearson responded to his critics that moral or theoretical arguments did not discredit the results of the study. He demanded "statistics on the table" from his critics; different data and improved methods of analysis, not diatribes, would be required to discredit his results.

Stigler uses the title's metaphor to frame his approach to the history. He organizes the volume into five parts with individual essays about episodes on the invention of the conceptual tools and mathematical techniques that have come to make up modern statistics. Part I tells the detailed story of Pearson's challenge before proceeding to chapters about the work of Lambert Adolphe Quetelet concerning the "average man" and the development of techniques for aggregate data by economists William Stanley Jevons and Francis Ysidro Edgeworth. Part II treats the conceptual issues surrounding Francis Galton's "invention of correlation and his discovery of the phenomenon of regression to the mean" (6). Part III returns to the seventeenth century for the debates about the uses of mathematical probability in medicine and philosophy. Part IV begins with the presentation of "Stigler's Law of Eponymy: No scientific discovery is named after its original discoverer" (7). The section illustrates the law with a series of vignettes on well-known, and lesser-known, issues surrounding the priority of invention, including Bayes Theorem, the invention of least squares, and the origins of maximum likelihood. Part V reviews "Questions of Standards," the biggest of which are the varying meaning and claims of the "normal curve" and "normative terminology" (361, 403).

In all the essays, Stigler brings his experienced eye and capacity for retrieving and explicating the contemporary local and intellectual context and technical detail of the many issues involved in the development of modern statistics. The reading is a treat. Stigler brings a fundamental sympathy and respect to his protagonists, even while he reveals his sometimes prickly characters behaving in high-handed, petulant, or venal ways. He treats the many episodes of academic intrigue fairly, and he effectively brings out the irony, humor, and frustrations of the great men involved in great debates. For the most part, he stands aloof from his subjects, and lets his constructions of his protagonists' dilemmas convey his larger point about the intimate relationship between the development of the technical and conceptual aspects of statistics and the social policy questions that these methods were intended to address.

In the essay that provides the volume's title, Stigler concludes that Pearson clearly won the argument about the requirements for demonstrating the impact of parental alcoholism on children, despite his now discredited eugenicist predilections. But Stigler also points out that Elderton and Pearson overlooked or downplayed the "one real effect he did find, the evidence for fetal alcohol syndrome" (50).The paths of scientific innovation are twisted and confusing, but Stigler's history helps us understand how we got where we are, and how much we owe to the pioneers.

Margo Anderson
University...