The Lady Tasting Tea

W. H. Freeman: New York, 2001. 340 pp. ISBN 0-7167-4106-7. $23.95.

Salsburg intends this book "to describe the statistical revolution in twentieth-century science in terms of some of the people (many of them still living) who were involved in that revolution." He insists on doing this in a "nonmathematical" manner--without graphs, equations, or mathematical notations--though mathematics finds a use as design element by graphically gracing each chapter with an appealing, normal-seeming distribution. What results is another volume of "popular science" that holds its reader in contempt as incapable of handling complex concepts in anything like their essential form. Whether statistics has something profound to say about the universe or has anything to offer as tools to mankind cannot be decided from this nontechnical ramble around a necessarily technical discipline.

Make no mistake, Salsburg's book is probably well enough written; it may be a classic among books of its type. It is that "type" that is bankrupt. Beyond that I could not see. Salsburg did not invent the chatty, factoid-strewn style of reportage; that cannot be held against him. Still, his was the creative decision to adopt that style.

In order to address the charge leveled against me of missing the point, let me acknowledge that I know it to be among the duties of an author to cull the important from an entire herd of noisy facts and to choose what to present to the reader. Yet, additional duties govern the author's responsibilities to the readers of a technical abridgement such as this one. The premise must be to get at the researcher's technical achievements by sneaking up through his (nontechnical) environment and demeanor. The promise must be to describe the intellectual process leading to those achievements. Delivering the description of the man but falling short of making the required intellectual connections cheats the reader of the contractual prize. Insights into creativity and not parlor (or laboratory) dialog are what are sought.

In Salsburg's book, laid out in something-like-chronological order are the developments of the intertwined professional lives of dozens of 20th-century statisticians from Karl Pearson's time to present and the effects of this century's dramatic events on these statisticians' views and venues, as well as Salsburg's commentary on the consequences of his "statistical revolution" and its profound effects on the minds of scientists and the public at large. What goes wrong to ruin this admirable plan of organization is hard to pinpoint but easy to exemplify.

Into the midst of otherwise-admirable narrative intrude irrelevant inanities. As W. S. Gosset's time on earth in the book draws to a close, we read, "When he died at the age of sixty-one, he left his wife, Marjory (who was a vigorous athlete, having been captain of the English Ladies Hockey Team), one son, two daughters, and a grandson." Gosset's wife, despite her impressive athleticism, cannot manage to muscle her way into any other passage in the entire book. This is not a quote cruelly snatched from its context; there can be no context for such a whopper. Further, though perhaps the briefest, it is not the only such pointless digression.

As a statistician himself, and serving to bolster his firsthand credibility, the author has met many of the researchers whose achievements he describes. But, does the patient reader really need to know his dining history with Chester Bliss, the pioneer of probit analysis? As warmly amusing as is the anecdote of the author's 1960s-vintage programmable desk calculator and its abrupt impact printer terrifying the cleaning man one night ("'BRRAAK,' it said"), does it advance the story of the statistical revolution?

Still, as irritating as are these stylistic idiosyncrasies, even more maddening for a physical scientist is the author's repeated and articulated thesis that statistics has supplanted something he calls "deterministic science":

Recall that, before the statistical revolution, the "things" with which science dealt were either the measurements made or the physical events that generated those measurements. With the statistical revolution, the things of science became the parameters that governed the distribution of the measurements. In the earlier deterministic approach, there was always the belief that more refined measurements would lead to a better definition of the physical reality being examined. In the statistical approach, the parameters of the distribution are sometimes not required to have a physical reality and can only be estimated with error, regardless of how precise the measuring system. For example, in the deterministic approach, there is a fixed number, the gravitational constant, that describes how things fall to the earth. In the statistical approach, our measurements of the gravitational constant will always differ from one another, and the scatter of their distribution is what we wish to establish in order to "understand" falling bodies.

This is hogwash. I know I'll be labeled as a hopelessly romantic follower of an "earlier approach", but last time I checked, there was only one world to be measured or understood, and even the ancient Greek philosophers recognized the distinction between the "many" and the "one": the variable manifestations taken on by worldly phenomena that nevertheless can be construed in human understanding as having common causes or relations. The interworkings of the entities of our physical constructs as well as their statistically measured values are of interest to science and engineering. But, how the gravitational constant fits into physical theory and its mathematics is infinitely more important than the skew and kurtosis of its distribution.

The vapid straw man constructed by the author is too easy a target even for a text free of mathematics. Its argumentative structure falls down about its own weak supports.

On second thought, what goes wrong in this book is not so hard to pinpoint. Statistics, its central subject, was chosen rightly as an interesting and important technical human endeavor. Unfortunately, it does not necessarily follow that a nontechnical book about its practitioners will share in that interest or importance. Too much depends on the Muse and her kindliness toward the storyteller.