Freakonomics: A Rogue Economist Explores the Hidden Side of Everything

#25
Title: Freakonomics: A Rogue Economist Explores the Hidden Side of Everything
Authors: Steven D. Levitt and Stephen J. Dubner
242 pages (book club edition)
Publisher: William Morrow
Year: 2005
Genre: Economics/Culture
242 pages
+ A fun foray into seemingly dissimilar questions about society, readable
- Cumbersome transitions at times, dismisses other arguments in suspect ways

This was a good-enough non-fiction read, though I think it does illustrate the idea that a bestseller may be appealing without being rigorous. As companion pieces, read Gladwell's The Tipping Point: How Little Things Can Make a Big Difference and Blink: The Power of Thinking without Thinking; Gladwell addresses some topics in common with Freakonomics and though one may also question his accuracy, he is a better writer.

Accuracy is a bit hard to evaluate here, since the methodology and statistics aren't described. My impression is that there is an over-reliance on correlation and that at times it is confused with causality. In addition, the justification for calling this research "economic" rather than "sociological," for example, seems to be the conversion of what we might understand as "psychological motivation" into "economic incentive." I'm not opposed to this, but when I consider how to design some of these correlational studies from my perspective as a psychologist, I wouldn't do anything different (other than underscore more firmly in my popular reporting that correlation is not causality). I question what appears to be mystique-building on the authors' part in this regard.

Each chapter is preceded by an annoying, self-aggrandizing excerpt about Levitt from one of Dubner's articles. I found these really offputting and was glad to learn that they have been removed from the revised edition, apparently because I was not alone in finding them irksome.

Like all reports of statistics describing a large number of participants, these reports provide, at best, generalizations about how the majority of those surveyed or observed behave (or so I assume--measures of central tendency were not reported, nor was the degree of significance in most cases). As anyone who does not have 2.3 children knows, statistical samples tell us about a fictional person. I remind you of the joke about the three statisticians who go deer-hunting. They spy a magnificent buck and the first statistician exclaims, "It's mine!" BANG! Her bullet goes two yards to the left of the deer. "No, it's mine!" calls the second. BANG! Her bullet goes two yard to the right, and the third statistician yells, "Bullseye!" This book would be more interesting, and more useful, if it told us something about the range and tails of the distribution in each study, giving the reader a better understanding of human experience, or, as our unique and individual experiences are known to statisticians, "error."