There is no unifying theme to this book, and the book will even get into it. Each section is an investigation into some problem and compares data that you might not otherwise think to compare. This is the work of a man who likes to get his hands dirty looking through data for connections and trends. However, no correlations are made until he finds a method to control for other variables -- such as trying to establish the link between money and politics by concentrating on rematches between the same two candidates.
There are some interesting moments, some infuriating and some a little dull. The authors go into detail about the impact of Roe v. Wade on violent crime 15-20 years later, the economics of dealing crack (and why crack dealers still live with their moms), but also look into things like the impact of a person's name on their future employment. (That latter chapter, while interesting, seemed to go on a little too long for me.)
I read the 2006 revised hardcover, which contained additional material. In particular, the original article that spawned the book itself, some other articles published after the book and some blog posts.
Overall, it was an interesting read. I'm not running out to buy another one, nor am I ready to concede the war on crime was won because of abortion, but it was an interesting exercise.