Wednesday, July 27, 2005

Freakonomics should teach you to form your own conclusions

I may have misled you. I read Freakonomics by economist Steven Levitt and journalist, Stephen Dubner and put it on my sidebar to tell the world about. I liked the book. The authors take an econometric look at some topics that don't normally get addressed by academics and they do it in a style that is readable and enjoyable. They come to some controversial conclusions the most famous of one is that legalizing abortion in the seventies is responsible for the reduction of violent crime in the nineties because children that would have otherwise been unwanted were never born. I don't think you could come up with a more controversial statement than that. It turns out that it might not be true (see Did Legalizing Abortion Cut Crime by Steve Sailer).

What I took away from the book is that you can and should apply econometrics to your hypotheses and be open to the conclusions even if they may be counterintuitive. The knowledge of why a thing is the way it is may allow you to change it or fix it or simply understand it. This is always with the caution that correlation is not causation. I am a critical thinker myself. I find myself unable to stop applying an almost mathematically rigorous approach to problems around me, at least for setting them up, even if I never have the time to follow through, collect the data and solve them. This applies to something as simple as whether I could recover the heat from the drain water in the shower and save energy to something as complicated as allocation of resources in my work or actual Chemical Engineering.

I should have applied this critical thinking to the conclusions of Freakonomics. The difficulty in a popular books like that is that most people don't have the time to check all of the facts, and I don't have the time to collect the data and review the results myself. So Freakonimics gets points for a catchy title, easy reading on a sometimes boring subject, and the great idea of applying econometrics to a wide ranging set of problems, but it gets some negatives for drawing the wrong conclusions from the data and oversimplifying the topic to make it accessible to the reader. Perhaps they themselves they weren't critical enough in their approach. All of these thoughts were brought to mind today by the review of Freakonomics at Stay Free! Daily.

I am a glutton for the econometrics, so if you would like a better, much more staid but rigorous popular book on the subject you can read Predicting Presidential Elections and Other Things by Ray C. Fair or go to Fair's wesbite at Yale. Here is the presidential vote equation. As always when I find some good things like this I will pass them on. My own book, if I would ever pick a topic and write one, is years and years away.

No comments: