rblog

Freakonomics : A Rogue Economist Explores the Hidden Side of Everything
Authors: Steven D. Levitt, Stephen J. Dubner

What a fun book! In my ongoing discussion of books written by professors and books written by journalists, I may have discovered the magic answer - a book written by both. Freakonomics is written by a star economics professor at the University of Chicago (Levitt) and a journalist (Dubner). The result is both readable and credible.

The book applies economic analysis to a variety of real-world problems (e.g. why did crime rates drop in the 1990s - shockingly, the authors demonstrate that this was due to legalized abortion!). There may have been the slightest political bias, but in general, the authors approach the problem with an open mind and let the data lead where it may. For example, it turns out that successful parenting is less about what you do than who you are - again, based on a study of the data.

This book is less about economics than it is about statistics and determining meaningful correlations from masses of data. As always, it is about asking the right question and designing the right "experiment" - once you have that, the answers will speak for themselves. The book's appeal and the reason for its title is the wide variety of topics covered - from parenting and crime rates mentioned above to the economics of drug dealers and why swimming pools are more dangerous than guns.

My favorite section of the book was the discussion of the crime rate drop in NYC during the late 80s and through the 90s. This topic was also covered in Malcolm Gladwell's "The Tipping Point", a book that underwhelmed me. I was pleased to see that Leavitt essentially disproved Gladwell's theory and presented a much stronger argument (intuitively speaking) with far more data. Strangely, Gladwell has a quote on the cover of the book (showing that he clearly didn't read it in depth).

One issue I would raise with the book is that there is no real discussion as to how the statistics were collected and whether there were alternative data collection methods that would have subtlely influenced the outcome. One famous example of this was an election survey in the Truman-Dewey race. The phone survey showed voters strongly for Dewey, but of course Truman won. The bias was that the voters were only people who owned phones! In an era where phones were still relatively new, this was a hidden bias in the data. Given that statistics can be used to prove almost anything, I think that special care needs to be taken to make every step of the process as clear and transparent as possible.

Recommended to everyone who enjoys thinking and wonders about the world.