Lies, Damn Lies, and...well, you know

A while back, I mentioned Steven Levitt's Freakonomics, including some of his more interesting assertions (poor economics of selling drugs, parenting techniques that matter, etc.) One of his most controversial points is that the real cause of the drop in crime in the 90s was Roe v. Wade. His story is in-depth, of course, but the gist is that unwanted children have the highest risk of a life of crime. He looks at a wide range of data throughout the U.S. to prove this effect is real. In fact, he asserts that it is far more important to the drop in crime than traditional views, including more cops on the street, "innovative" policing tactics, changes in drug use patterns, and others.

Enter Franklin Zimring, author of The Decline of American Crime. To be clear, I have not read this book, I only listened to an interview on public radio's Forum with Michael Krasny. So this entry isn't about the book...it's about how even reasonable, academic, smart, diligent people can come up with such different views. Zimring looks specifically at the Roe v. Wade assertion and finds it wanting. In a gentle way, he even accuses Levitt of one of the greatest sins for an academic: confusing correlation with causality. Yes, Zimring says, the timing is right to believe that reducing unwanted children would cut the crime rate, but the U.S. isn't the only story. He decided to check for the same effect outside the U.S.--and didn't see it. Then he makes a variety of other contrasting claims: hospital admittances for drugs actually went up in this period, birth rates for young, unwed mothers went up, etc. Zimring says there is some modest effect, but explicitly says it is not even close to the most important.

I sure don't know the answer and I guess I could write this off with the positive view that "I'm confused at a higher level." But it is less than satisfying....