Superfreakonomics has come in for some criticism by Andrew Gelman among others, and Stephen Dubner’s response has certainly not helped. So I was amused when checking the deceased blog of frequent Gelman-target Satoshi Kanazawa that the “scientific fundamentalist” also thought poorly of the sequel. He doesn’t talk much about the book getting things “wrong” (though he says the logic for doctor’s ignoring risks to their patients seems inconsistent), but does think that it’s too padded with interviews due to the lack of new research from Levitt that could fill an entire book. Gelman also thinks the reliance on other researchers (particularly, friends of Levitt that he trusts) is problematic, but because it leads to uncritical acceptance of ideas he shouldn’t be so confident in. I’ve only read first book, so I can’t give my opinion, but I didn’t think the “drunk walking” example was so bad. Often enough, the distance to be traveled is a constant. Admittedly, the data on drunk walking may be sketchy.
On an unrelated note the Monkey Cage has a post on the representation of groups in D.C. Professionals like scientists & lawyers are better represented than manufacturing workers, but college students are less represented than gun owners. The low representation of Catholics may be the result of them leaving responsibility to the Church, whereas Jews form lots of organizations to pursue various goals. The rate at which each group pays attention to news and votes is also important. This sounds important for public choice theory, but I don’t recall them discussing it too much other than in terms of a concentrated industry being able to extract a small amount from each member of a vastly larger indifferent public. Something like the “New Class” idea that Kenneth Anderson always goes on about could tie in to that.
UPDATE: Matt Grossman continues his guest-blogging about interest groups at the Monkey Cage. Apparently he’s got a book coming out, and a lot of this material is from there. His recent posts have attempted to explain why the media seems to lean liberal, and to argue that (despite the influence of unrepresentative interest groups) our elections look like “one man, one vote“.
UPDATE 2: OrgTheory on lobbying and other political matters.