(cross posted on Natural Blogarithms)
I just finished reading the book by Steven D. Levitt and Stephen J. Dubner. It had been recommended to me by a couple of different folks but the most recent recommendation came from listening to the TWIT podcast where Leo Laporte recommended it as an Audio Book. The TWIT podcast, in fact all TWIT podcasts are now recommending a different Audible book. I had only had a cursory introduction to the topic of this book but when Leo mentioned it, he gave a very good overview that convinced me to pick it up the last time I was at the library with my kids. I no longer pay for a subscription with Audible, so I figured I might as well read it the old fashion way.
Steven D. Levitt is an economist who has a tendency to look with an economist’s perspective at everyday issues in society. For example:
- What do schoolteachers and sumo wrestler’s have in common? (Answer: Cheating, evidenced by the numbers)
- How is the KKK like a group of real estate agents? (Answer: Thrive off of misinformation)
- Why do drug dealers still live with their moms? (Answer: Because drug dealing is not nearly as profitable as conventional wisdom tells us)
Stephen Dubner helps to paint the answers to these questions through data collected and analyzed by Levitt in a very understandable way. I enjoyed the book and would recommend it because it helps to provide a mindset to answering the tough questions a society faces. Of course, as a mathematician, I like the idea of looking at correlations between data sets and I especially like the care with which the authors explain the difference between a correlation between two factors and a cause-effect relationship between those two factors.
Like I said, I would recommend this book as a very interesting read to anyone who likes to question conventional wisdom.
I have two favorite sections in the book. The first is the chapter on parenting (Do we really need parents?). In this chapter, the authors show the correlation between things like reading with your child and the performance in school. In fact, they give a list of several things, of which half are strongly correlated with school performance (such as having a lot of books in the home). The other half are not correlated with school performance (such as reading with the child). It paints a very interesting picture, suggesting that it is not what you do as a parent as much as it is who you are.
In another section, Dubner gives an overview of Levitt’s research in explaining the surprising drop in crime rate over the last decade. As recently as the late 80s to early 90s crime was on the rise with nothing but doom and gloom in the forecast but suddenly, without any explanation at all, crime began to decrease. Levitt reviewed many of the theories that popped up trying to explain this change in trend but in looking at available data, none of those explanations could account for the sizeable change. Suggestions like changes in policing strategies and the “bursting of the crack bubble” could not account for the dramatic reduction that was taking place. Levitt proposes an idea that is intriguing but a little bit disturbing. He suggests that it is abortion. Roe v. Wade allowed an enormous percentage of unwanted pregnancies to end in abortion instead of children being raised in lower income families with a mother who sees that child as an inconvenience. Both of these circumstances can be strongly correlated to criminal behavior later in life. Plus, it can be shown that a majority of abortions occur in these settings. Thus, thousands, if not millions of potential criminals are no longer being born. The authors are careful not to suggest that morality of abortion can be determined by this unpredictable, yet beneficial, result for society. In fact, the rightness or wrongness of abortion has nothing to do with this result. It simply happened and the data seems to support their claim. Of course, there are all sorts of immoral actions that can lead to beneficial results for society. For example, if the death penalty became the consequence for all criminal behavior even down to shoplifting, wouldn’t crime also decrease? Still, it is a very interesting point to be made and one I had never heard before.
Overall, I was pleased with the read. It is fairly short, taking me only a few days to finish. I look at difficult issues a little differently after having read the book, often asking myself about what the data would suggest, or simply wandering about the relationships between certain actions and the consequences. Intuition might tell me one thing, but the actual results may be completely different.