Freakonomics

There’s nothing like a 7-hour travel day to catch up on reading. I managed to crack open and finish Freakonomics by Steven Levitt and Stephen Dubner. Levitt is a lauded young economist from the University of Chicago who likes to use his statistical superpowers to answer interesting questions. Dubner is a New York Times journalist who likes to write about the answers to interesting questions. Hence, the collaborative book.

Freakonomics is a collection of (mostly) unrelated questions grouped into chapters by their statistical similarity. How else can one coherently write about the Klu Klux Klan and real estate agents in successive paragraphs? Even if the statistical technique is several universes away from my own skills, the logic is largely sound. More importantly, the point of these exercises is well taken: Conventional wisdom is conceived from biased self-interest and rarely reflects what actually happens in the world.

In taking great care to define their area of interest, Levitt and Dubner make a clear distinction between morality (the way the world should work) and economics (the way it actually does). The authors seem to let this stand as their excuse for being distant in looking at the world and letting the statistics speak for themselves. However, at times they are biased and/or limited in their view.

The two most controversial ideas in Freakonomics are:

1. Abortions are most responsible for reduction in crime
2. Parents have no effect on improving their kids’ education

The authors clearly danced around the politics of the first statement (which is my summarization, not a quote). They re-stated the difference between morality and economics, emphasizing that it is strictly economics seeing this correlation between Roe v Wade and the 1990s crime plunge. They were playing Solomon in some thought exercises, and not advocating for or against abortion. If any ideology leaked through, it probably erred on the side of conservatism (especially when trying to calculate the trade-off between aborted lives and those snuffed out through homicide). All that summersaulting aside, the research seems very sound. It doesn’t mean abortions are the key to crime prevention, but it does recognize a positive, unnaticipated effect of those butterfly wings. More practically, it points yet again to the underlying cause of crime: socioeconomic depression. And there are many ways we can work as a society to address that.

The second controversy was sparked by the authors’ inability to remain dispassionate. Early in the chapter titled, “What Makes a Perfect Parent?” they write:

Theirs is a gesture of love, surely, but also a gesture of what might be called obsessive parenting. (Obsessive parents know who they are and are generally proud of the fact; non-obsessive parents also know who the obsessives are and tend to snicker at them.)

It it easy to see that anyone doing any of the activities Levitt/Dubner associates with “obsessive parenting” might be a bit turned off by that choice of phrase, and by the picture of the knowing “non-obsessive parents” snickering amongst themselves. So right from the start, the authors have alienated a sizeable readership, needlessly.

The real problem with this later chapter, though, is that the underlying question is flawed. Or more correctly, the implied definition is flawed. The bulk of the parenting chapter examines the data from a 1990s study by the U.S. Department of Education, the Early Childhood Longitudinal Study (ECLS). In this study, a carefully-selected cross-section of American students was tracked from kindergarten through fifth grade. There was also an exhaustive questionnaire that accompanied each kid, asking about details such as parenting and the number of books on the home shelf. Levitt used this data to extract a list of eight things that mattered, and eight that did not, when parents try to improve their child’s education. The conclusion was that the things that mattered the most were environmental factors that were already established by the parents before the child was even born. The things that did not were the stuff parents do to help a kid’s education. All great stuff … except the definition of educational impact was success at standardized tests*.

Giving benefit to doubt, these writers probably meant to be clearer about the limits of their study. They are not talking about perfect parenting, as the the title of the chapter suggests. They are talking about the impact parents have on test scores. That may be a fairly obvious distinction, but it is not obvious from the text of the chapter. When they write:

“… [A] mother who stays home from work until her child goes to kindergarten does not seem to provide and advantage. Obsessive parents might find this lack of correlation bothersome—what was the point of all thoe Mommy and Me classes?—but it is what the data tell us.”

The assumption is that the only reason to stay home and take Mommy and Me classes is to improve test scores. Levitt and Dubner may recognize and willingly admit to this limitation, but they don’t write about it. They just snicker.

There is also the matter of the 16 items they included for comparison. In the questionnaire, there were some very important questions about mobility, asking parents how long the child has been at the current address and how many moves they have made in her lifetime. The closest Freakonomics got to including that was the inclusion among the “Doesn’t Matter” list: “The child’s parents recently moved into a better neighborhood.” That isn’t quite the same as “The child moves around a lot” or even “The child has attended more than one school.” Student mobility is a big obstacle to learning, and studies have tied mobility to lower performance in schools. Something that might make Levitt giddy can be found in an Australian study on the subject, which recognized that different children respond in different ways to mobility. In fact, the study also suggests (in passing) that mobility and poor performance might be an indicator of socio-economic status. Keeping with the premise of the book, it seems that it would be within Levitt’s reach to study the impact of mobility as detected both by questionnaires and actual movement of students during the course of the ECLS study. That didn’t happen, and as a result the parenting chapter seemed very incomplete.

The most important sentence in the book comes technically before it begins, in the explanatory note:

As Levitt sees it, economics is a science with excellent tools for gaining answers but a serious shortage of interesting questions.

The book is filled with interesting questions and, for the most part, well-conceived answers. Snickering aside.

---

* There are some studies that say educational reform has little to do with social mobility in moving from class to class, which is the primary motivation for accountability through testing in the first place.