Got a blog tip? Contact us

« Picture of a Disaster | Main | Wal-Mart pitches Brooklyn »

Thoughts on Freakonomics

Freakonomics, the best-seller by economist Steven Levitt and his wordsmith, Stephen Dubner, has been getting a lot of attention and I can see why: Levitt is a creative thinker whose practical concerns extend beyond the question of how to make money. An economist that isn't deathly boring? Yes, yes, sign me up. (Most of the work in this scattered collection is quantitive social science generally rather than economics in particular, but still...)

Yet something about this breezy read, with its interesting anecdotes and surprising conclusions, made me suspicious. Perhaps it stems from this statement in the introduction:

Morality, it could be argued, represents the way that people would like the world to work -- whereas economics represents how it actually DOES work. Economics is the science of measurement.

Okay. If economics is a science, I am a walrus. Without getting bogged down in an old debate about "hard" science v. "soft (social) sciences" (a question of whether measuring the interplay of chemicals is equivalent to measuring human behavior) let me just say up front that I'm of the Neil Postman school, which holds that social scientists are "scientists" in quotes. Like cultural critics, historians, and journalists, they are essentially story tellers. The use of numbers and quantifiable data can add greatly to human knowledge but attempts at empiricism shouldn't be categorically privileged over other means of truth-telling (in matters of human behavior, that is). In other words, counting and data collecting can be as misleading as the expert opinions Dubner and Levitt purport to debunk.

One need look no further than Levitt's widely reported studies on the impact of legalizing abortion for evidence of just how fuzzy social "science" can be. Levitt provides compelling evidence that Roe v. Wade, by preventing unwanted pregnancies, is largely responsible for the drop in crime throughout the US in the 1990s. But the authors fail to acknowledge -- or even mention -- that a large body of evidence contradicts their thesis.

Though violent crime over all dropped in the 1990s, it actually increased among teenagers -- the group that would have been affected by Roe v. Wade -- and especially black teenagers. (African Americans are three-times more likely than Anglos to have abortions.) And even the assumption that Roe cut the number of unwanted births is debatable; illegitimate births maintained a steady climb after Roe, for example.

For a thorough and well-written examination of the abortion argument, see Did Legalizing Abortion Cut Crime by Steve Sailer (a professor of economics at the University of Chicago and a conservative).   The point isn't that legalizing abortion did or did not cut crime, but rather that Freakonomics grossly oversimplifies complex social matters... and in a deceptive, Gladwellian kind of way.

Another example: in chapter 4, the authors argue that the education, class-status, and age of birth parents largely determines how well their children do in school and that particular parental behaviors are more or less irrelevant. "It isn't so much a matter of what you do as a parent," the authors write, "it's who you are."

I'm sure there's a good deal of truth to this, but Levitt and Dubner gloss over the fact that demographic data such as class and income level are much easier to measure than behavior. To quantify parent habits, they relied on surveys, on self-reported data. It's widely understood in the ad industry that people don't respond accurately or honestly to queries about personal matters, so why take these surveys at face value? If you ask a parent whether they spank their children, take them to museums, or read to them ever day, what do you think they're going to say? (Dubner and Levitt acknowledge the unreliability of survey data with the spanking question, but not the others.)

So while it may very well may be that birth parent demographics are primary predictors of student success, you can't so easily write off parent behavior. Just because you can't reliably measure something doesn't make it irrelevant.

Dubner and Levitt conclude the book by predicting it will encourage readers to ask questions of self-appointed experts and others touting some party line. Hopefully those questions will begin by examining some of the authors' own arguments.

RELATED: For more commentary on Freakonomics, see My New Cuddly Pet is a Smith & Wesson