Of Stories and Statistics. - Review - book reviews

"An integration of stories and statistics, or more generally, of the literary and the scientific, can be energizing."

And John Allen Paulos does just that in this educational and entertaining book - he energizes the reader to fill the gap between stories and statistics, between narratives and numbers, between tales and truth. Writes Paulos, "Describing the world may be thought of as an Olympic contest between simplifiers - scientists in general, statisticians in particular - and complicators - humanists in general, story-tellers in particular. It is a contest both should win" [italics mine].

Paulos informs the reader in a detailed, yet sometimes humorous, manner in which both storytellers and statisticians can win in describing our world. He reminds us that when you listen to stories you want to suspend disbelief to be entertained, whereas in evaluating statistics you are inclined to suspend belief so as not to be beguiled. And yet if the gap between the two can be bridged, if people can find the hidden mathematical logic of stories, and if narratives can be added to statistics, the world would be a better understood place.

Paulos writes wittily. The reader must be on his or her proverbial toes to get the full effect of Paulos's intelligence. He writes of ordering an extra medium ice cream, he mentions Keillor's Woebegone Effect in which everybody is above average, and he describes young Shepsel, the neighborhood's perfect sharpshooter (Shepsel shoots first, and then he draws the target around the bullet hole). He discusses the "belly-shaped curve" and gives a good logical probabilistic explanation for Murphy's Law by using socks (yes, socks). Another example of Murphy's Law Paulos explains is the "waiting time paradox," how delays in supermarket lines and doctors' offices are caused by mathematical variation, not some cosmic malevolence. Furthermore, he explains how you can become a "psychic" to your friends by using cards. It is not a magic trick, but a mathematical trick using probabilistic coupling.

Paulos's greater contribution, however, is not in explaining how numbers can be used in a comical way but instead in a rational, logical way - a way in which statistics can add to stories. "Subjective viewpoints and objective probabilities have at times a troubled relationship that resembles that between informal discourse and formal logic," he writes. "Applications of probability and statistics require a story, a context, or an argument in order to make sense." He says applying probability and statistics is much more a matter of comprehending the situation, of creating informal arguments, and of building comprehensive narratives than of substituting numbers into formulas. For instance, he contends that no matter how accurate the statistics are on, say, affirmative action, to be truly meaningful they must be imbedded in a story that helps advance the argument. He adds further that when one understands the idea of freedom of expression, "It should not be measured by how likely the average person is to be silenced, but by how likely someone with something to say is."

Paulos also notes how stories can misuse statistics. He relates how the collection and distribution of social statistics affects the quantities being measured. For example, on religious surveys people tend to overestimate their activity while on sex surveys they tend to underestimate it. Other abuses: the average American adult has one testicle and one ovary, and the average resident of Dade County, Florida, is born Hispanic and dies Jewish. Paulos takes on the believers in the Bible codes by offering a humorous Bill/Monica code in the U.S. Constitution and a Roswell/UFO code in a verse in the King James version of Genesis. He writes that "common sense underscores the inanity of basing any political, spiritual, or sexual judgments on these contextless numerological oddities."

Finally, Paulos believes that if people are comfortable with the insignificance of most coincidences (the statisticians) or if they insist on always finding meaning behind the coincidences (the storytellers), it reveals much about their personality and world outlook. Both, however, should contribute since the integration of both can energize both. His concluding paragraph is worth quoting in full:

How we can maintain a place for the individual, protected from the overweening claims of religion, society, or even science, is an increasingly important unsolved problem. Its solution, I have no doubt, will require simply and pragmatically accepting the indispensability of both stories and statistics and of their nexus, the individual who uses and is shaped by both. The gap between stories and statistics must be filled somehow by us.

Mark Durm is a professor of psychology at Athens State University, Athens, Alabama.

COPYRIGHT 1999 Committee for the Scientific Investigation of Claims of the Paranormal
COPYRIGHT 2000 Gale Group