Posted: 9/30/2005 8:35:14 AM EDT
[#9]
Quoted:
Quoted:
I'm part way through Levitt's book (the audio version, narrated by his coauthor, Dubner I think, who has a crappy voice, get the hardcopy instead) and some of the conclusions are interesting, but some are reaching a bit. The readers' reviews on Amazon go into a little detail dissecting some of Levitt's "conclusions" as being tenous at best... suspect that the editors were pushing for some "wow factor" on a dry subject, and playing the race card is a sure ticket to the NYT Bestseller list.
|
This reminds me of the concerted effort to discredit John Lott's books. Any error or perceived error was going to negate the force of his argument and they really jumped on his anonymous e-mail as some proof of incompetence.
|
You've lost me. Don't believe I have insinuated that Levitt is somehow incompetent. There are some thoughtful reviews of his book that address specific conclusions he draws - there are quite a few, here is one: Too often articles or books written by economists are couched in arcane verbiage and statistics. Levitt avoids this, and the result is a very interesting, thought-provoking review of several social myths. Unfortunately, it is also greatly over-simplified and in error on key points.
He begins by summarizing the status of crime in the mid-1990's - high, and projected to go much higher with the coming "teenager boom." Instead, it began a long, steep decline. The most common "explanations" were "roaring economy," "gun control," and "innovative policing." Levitt then goes on to summarize data that attempt to refute them all. For example, a good economy might decrease economic crimes, but why did violent crimes drop even more? Further, why didn't crime also fall during the booming '60s? As for innovative policing, Levitt reports that the declines began prior to this initiative, and that its prime contribution was through adding policeman (accounting for about 10% of the drop). Similarly he refutes the logic for crediting increased rights of citizens to carry guns, and gun buy-backs, while the drop in crack prices is credited with 15% of the drop.
Levitt then reports the results of Romania's strong anti-abortion posture in the 60s - a large contingent of resented children, many of whom became serious problems when they grew up. Finally, the "shocker" - Levitt presents various data that provide a solid case for concluding that the drop in crime was primarily due to Roe v. Wade making abortions available to lower-income women - many of whom would have had problems raising the unwanted children. Another important factor was the greatly increased incarceration rate - accounting for about 1/3 of the decline.
ON THE OTHER HAND, in 1993 there were 1,927 murders in NYC, and only 1,181 - the year after significant police reforms were introduced by Commissioner Bratton. In 2003, there were 596 - a level that had not been reached since the early 1960's. The rapidity of this drop cannot possibly be explained by increased abortions! Additional evidence is provided by reviewing murder rates by age groupings (per Professor James Q. Wilson). In the late 1990's, roughly a quarter-century after Roe, the murder rate was falling for offenders aged twenty-six and older - a class of offenders much too old to have been affected by Roe one way or the other. As for the youngest offenders, those between sixteen and twenty, their murder rate had jumped up in the early 1990's, probably because of involvement in the crack cocaine trade. Again, no Roe effect. Finally, Freakonomics' claim that increases in police staffing helped is countered by the fact that NYPD staff DECREASED post 9/11, while the murder rate has continued to fall. Then there is the problem of explaining why New Orleans' murder rate is about ten times that of New York City.
Other topics addressed by Levitt include documenting cheating associated with "high-stakes" (eg. potential job loss, raises, school closure) pupil testing (estimated at about 5% in Chicago Public Schools), documenting and explaining the lack of drug traffic profits for most of those involved (rakeoffs by those at higher levels). Another interesting and useful topic covered is how society often misplaces efforts into low-payoff efforts to protect children (eg. child-resistant packaging, flame-retardant pajamas, avoiding being seated near front-seat airbags, and keeping their children out of homes with guns), instead of the much higher-payoff of keeping children away from homes with swimming pools.
Throughout the book, Levitt carefully summarizes supporting data, while also informing readers of how similar data are often misused. His "bottom-line," so to speak, is for the reader to become more aware of the effect of incentives, and the frequent lack of factual bases for conventional thinking.
Another area where the authors go astray is in reporting achievement similarity between African-American and Caucasion children upon entering school - after controlling for key background variables. The inference is that low African-American pupil achievement in subsequent grades is simply the result of poor schools. Reality, however, is not so simple. While Levitt does point out that blacks exhibit serious self-defeating behaviors in school, he also fails to report that even in schools with strict management of behavior and very strong black parental support, a racial achievement gap does develop and grow over the years. (This result is similar to the finding that U.S. pupils generally score well in early-grade international comparisons, but then steadily deteriorate vs. their counterparts as they advance through the grades.)
An interesting, though questionable, two-hour read. However, in addition to the questionable assertions already pointed out, I wish that it had also covered the evidence regarding the linkages between education expenditures and pupil achievement (none), and healthcare spending vs. patient outcomes (negative - the more errors a hospital makes, the greater its reimbursement) - far too much money is wasted in these areas. (See John Wennberg's data on variations in Medicare spending, Hanushek's on relationships between class size and pupil achievement.) On the other hand, they'd probably get those wrong also!
|
There are over 500 reviews of this book currently up, doubtless some laud his every word while others take issue, I don't see citing statistics as some sort of a hatchet job, and whether it is Levitt or his publisher that is pushing the racial conclusions (all the way to #6 on Amazon's list) there is a strong history of the literary industry doing just that to push sales. It is a business after all, and controversy moves their product... (not unlike the media) One thing is for sure. Whatever constructive point Bennett may have wanted to make will only get through to those people who already had the same opinion. The media, both sides of it will only be waging partisan politics and the "firestorm" will take the place of reason.
Thank god Bennett didn't say something like "Hitler DID make the trains run on time", .......
|
Did you mean Mussolini? (making the trains run on time)
|