pageok
pageok
pageok
Freakonomics:

I know it's a little late in the game, but I just read Freakonomics over the weekend. It's a good read, and I applaud the authors for bringing economic reasoning to the masses (okay, the small percentage of folks who actually read nonfiction). On the other hand, in some ways I was underwhelmed. Some of the "brilliant insights" from the book, while perhaps not in general circulation, didn't strike me as either especially brilliant or original. Within the range of normal, parenting doesn't really affect children's test scores? Old news. Swimming pools in the yard are much more dangerous to children than guns in the house? Anyone remotely familiar with the literature on gun safety would have easily guessed this, as owning a gun is not very dangerous. And the fact that pools pose significant safety problems seems pretty well known, too. Legal abortion may (I'm not convinced that this has been "proven") reduce crime rates? Many years ago, I asked a (conservative)relative what she thought about whether abortion should be legal. She responded, "I want abortion to be legal because otherwise those kids will grow up to mug my kids." Low-level drug dealers don't get paid much? I read an article many years ago in Insight on the News about how teenage residents of slums go back and forth between menial jobs and drug dealing. "Gun buybacks" and other popular quick fixes don't reduce gun crime? Duh! [It's true that Leavitt is rebutting common notions promoted by the news media, but perhaps because of my background writing about "junk science," I'm not exactly shocked that the MSM tends toward illiteracy on scientific, statistical, and technical issues--anyone remember the breast implant fiasco?] (I did find the material on catching "cheating teachers" in Chicago quite interesting, however.)

Sure, Leavitt deserves a lot of credit for using the tools of economics and statistics to get hard(er) data on these matters, and for trying to answer interesting real-world questions instead of retreating into the typical academic economist's world of math. But the plaudits he's received for having such an incredibly original and creative mind strike me as excessive, as very little in the book seemed especially new or original. But an A+ to Leavitt (and Dubner) as popularizers.

Cornellian (mail):
I thought the point of the book was to illustrate economic reasoning as applied to day-to-day things that people wonder about, instead of abstract academic models. It's a sort of Innumeracy for economics. I don't think it was intended to present deep, original insights. It's debatable whether that's even possible in a book intended to popularize.
7.5.2006 12:13pm
Stephen F. (mail) (www):
I read the book a while ago and was underwhelmed for two reasons. First of all, for a book that touts economic solutions to various problems, it's incredibly light on actual supporting data. While the authors may have done this for readability purposes (the "cheating teacher" chapter does seem to bog down under the numbers), an appendix or end notes would have been appreciated.

The other problem I had was that a number of sections in the book just seem to try to answer the wrong questions. A realtor spends more time and effort in selling their own houses than they do yours? They prove that well, though it seems pretty much expected human nature to me. What they don't discuss is the importance of their findings. Would I be better selling my house myself than going to a realtor? Or is the half-hearted effort of a realtor still better than the extra effort of a layman?

Another example: the authors point to the instances of sumo wrestlers barely above qualifying level as losing to barely below qualifying wrestlers a disproportionate amount of time, and takes this as evidence of cheating. There may indeed be cheating going on, but what am I to make of that data? The sumo wrestlers cited are, given their fairly close records, expected to be of close talent. Since one group has a lot to fight for and the other group really has little go gain from a win, you would expect the former group to fight harder than the latter. This isn't cheating, it's just human nature. Are the (few) numbers the author cites so skewed that the human nature element can't account for it? How do the sumo records compare, say, to the records of an NBA team who's locked up a 7 seed against a team fighting for an 8 seed? I don't know, and the authors don't tell me.
7.5.2006 12:20pm
David (mail):
i teach economcs at the college level and read freakonomics when it first came out. i found it to be interesting becaise it related concepts in a fashion that could be better understood by non academics. in addition, it took on some controversial topics and had the data to support the positions, unlike many books which just seem to use more seat of the pants kind of logic. the data was what it was. i think too, that some of the positions were somewhat tongue in cheek, but ultimately proved to be substantive. i feel somewhat vindicated on a few points. cheating teachers are rampant in my school district where i send my kids, not because the students are bad, but because the schools are perceived as being good. it therefore becomes incumbent upon the high school teacher to make certain that the students do well in order to perpetuate this myth. the name theory was one i have thought about for years, anyone ever met a skinny bertha? just as order of birth has some ramifications in later life so too does a person's name. there are definately "strong" names vs. "weak" names, and i feel that they have an impact on a person's self esteem an sucess in later life.

all in all, i do not think that the authors want the reader to take them too seriously, rather just think about certain issues in a different way.
7.5.2006 12:22pm
John Armstrong (mail):

Sure, Leavitt deserves a lot of credit for using the tools of economics and statistics to get hard(er) data on these matters, and for trying to answer interesting real-world questions instead of retreating into the typical academic economist's world of math.


I think [b]Cornellian[/b] correctly characterizes the book as "a sort of Innumeracy for economics" [sic]. As such it is about mathematical reasoning.

So I have to ask: what's so horrible about math that you find it laudable that Leavitt "doesn't retreat into [it]"?
7.5.2006 1:16pm
msmith (mail):
..But the plaudits he's received for having such an incredibly original and creative mind strike me as excessive, as very little in the book seemed especially new or original. ...

Thanks!, thought I was the only one. The author's appearance on CSPAN was equally as dull and uninspiring.

Q: In light of the books mega-seller status and its supposed "popularization" of basic concepts in the dismal science and statistics, what does that say about the sorry state of basic education in these fundamental and important subjects?

If in fact people really found the book so fascinating. Sometimes wonder how many of these books actually get read after purchase, especially in the age of the Google synopsis and reviews.

Anyway, thanks for saying it.
7.5.2006 1:24pm
John Armstrong (mail):
msmith: what does it say about the sorry state of basic education that a supposedly-intelligent law professor describes the invocation of mathematical reasoning as a "retreat" from the populace?
7.5.2006 2:41pm
DavidBernstein (mail):
It's not the math that's the problem, it's the theoretical nature of the math, sometimes for what seems to be it's own sake.
7.5.2006 2:58pm
Eric Muller (www):
I think my favorite passage in "Freakonomics" is this little gem, in the book's acknowledgments:
"Of most particular interest to us was Stetson Kennedy, The Klan Unmasked (Boca Raton: Florida Atlantic University Press, 1990). . . . But Stetson Kennedy himself is probably the greatest living repository of Klan lore. . . . The authors visited Kennedy at his home near Jacksonville, Florida, interviewing him and availing ourselves of his extensive collection of Klan paraphernalia and documentation. (We also tried on his Klan robes.) We are most grateful for his cooperation.
Not only did the authors try on Klan robes, but they found it memorable enough to mention.

Sorry, but I just find this weird.
7.5.2006 3:17pm
Joe7 (mail):
I did not read the entire book, but was very dissapointed in what I did read--I've also perused their web site and am even less impressed--for two main reasons:

1) They repeatedly blunder in assuming correlation is causation.

2) Some of their conclusions are simply not supported by the data.

It very much appears to me that they find something startling or contrarian to say and then set out to "prove" it.
7.5.2006 4:53pm
Jonah Gelbach (mail) (www):
A couple of quick thoughts:

1. It's "Levitt", not "Leavitt" (perhaps some are thinking of the sec of hhs?)

2. I should say that I haven't read the book (though I have read at least a couple of the academic papers on which the book is based and am familiar with some of the others). Having said that, I doubt that Joe7 is being fair. One of the reasons that Levitt is so prominent in economics is that he takes very seriously causality and its distinction from correlation.

Most (all?) of Steve's work gets where it does because it involves

(a) asking a question, or asking one in a way, that others haven't asked (this claim notwithstanding Bernstein's criticism on this front), and

(b) finding an often clever way to at least plausibly deal with vexing statistical and behavioral problems (typically called "identification problems" these days) that make it difficult to establish causality.

My guess as a practicing academic economist (who tries not to use more math than necessary to make his point!) is that part (b) is at least as important as (a) to Steve's prominence among economists.

There are real fights going on among economists about whether the style of research that Steve and many others practice can shed much useful light. (This style of work is known as "reduced form" work---and is not, by the way, in any way a new approach. For what it's worth, many critics have a very good point that the retreat from math has come at the cost of a serious loss of precision and clarity. While there are times when math is abused in economics, I think criticisms of "the theoretical nature of the math, sometimes for what seems to be it's own sake", as Bernstein puts it, are made too often and are overblown. Math can be extremely helpful in clarifying behavioral problems and associated econometric issues, and sometimes there just isn't a non-"theoretical" way to do so. In my book, that's why we have specialization. Actually, if I had to add a (c) to Steve's prominence it might be his ability to explain complicated things plainly---but not all complicated things can be so explained.)

As a usual practitioner of reduced form work myself, I think that there's potentially a lot to learn from this approach. I think Steve has, in fact, been one of the most creative folks deploying methods like instrumental variables and differences-in-differences (which may nonetheless be the most abused approach to applied econometric research out there these days).

It is true that there have been a couple of high-profile cases of people challenging Steve's conclusions because of nontrivial coding mistakes that he has admitted to making. Personally I wish he hadn't made those mistakes, which could have been caught with a bit more attention to detail. But at the end of the day, I think that Steve's props within academia (i) are deserved, and (ii) come from finding at-least-plausible creative solutions to very tough causality problems. So I think Joe7 is most likely way off base.

Bernstein's "didn't-we-know-all-that?" point may be right---I don't know if we did or not---but I think you should get some points for writing papers that make a case rather than just having conversations with (conservative) relatives.

Jonah

(Disclaimer: Steve Levitt is a friend of mine and as an editor has treated me well. I am not, however, Steve Levitt having a Mary Rosh moment. Really---check out my website and you'll see.)
7.5.2006 7:11pm
davidbernstein (mail):
Jonah,

Thanks for the post. It's not that I don't think Levitt is an extremely smart guy, has done some very interesting academic work, and wrote a fine book for a popular audience. It's just that I don't think the book, and Levitt, quite lived up to their hype. That would be more easily forgiveable if the "hype" wasn't contained in the book itself, in the form of excerpts from a laudatory NY Times article about him.
7.5.2006 8:22pm
CBB:
The weird asides about what geniuses the authors are was odd, but the book achieved its intended purpose. Saying something doesn't live up to the "hype"--even hype contained in the pages of the book--is always a bit of a nebulous criticism. Saying the book was not groundbreaking or that we already knew most of the conclusions seems to expect a lot from a book that professed to do no more than popularize and compile the past 10 years or so of one guy's already published work.

One of the nice things about the book, and Levitt in general, is one of the same things that has inspired so many challenges and criticisms: That he is both the academic researcher trying to break the new ground and do the cutting edge technical research, and also the popularizer. Typically the academic can deflect any criticism arising from the popularizer's work by saying that it doesn't "quite capture the complexity" or whatever else, but Levitt doesn't have that luxury.
7.5.2006 8:45pm
bellisaurius (mail):
I don't think we give enough credit to popularizers in science and engineering. I mean, if it weren't for Asimov, I wouldn't be in my job right now. Had I read levitt's work when I was younger, I might have given economics a better shot.

The problems he looks at are just so interesting, and he does have a talent for storytelling, like the "How superman beat the klan" story. I too knew a lot of the general statistics before, but looking at it that way seems to be a little like saying watching MacBeth loses something because I know how it ends. it's the process that's intruguing.
7.5.2006 9:05pm
Toby:
My daughter. committed to studying either English or Sociology at the University of Chicago, has consented to take an econ course or two vecause of this book.

This makes me a fan, even though I felt like underwhelmed as described above when I read the book.

I also noted that my wife, a BA in French, always felt me a jerk when I espoused some of the opinions that come along with this work. She, a smart woman if nearly innumerate, found his arguments compelling.

Again, a thumbs up to the effects of this book
7.5.2006 10:10pm
Mp (mail):
The book is fairly interesting, but I don't see why people are trying to hype it so much. It isn't really cutting edge research -- social scientists have been doing this kind of thing for decades (though the cheating part is pretty original). They have used the same types of statistical methods (call it econometrics if you wish; psychometrics, econometrics, biometrics, etc. all are just basically subsets of the larger body of statistical theory) that hundreds of others have used, though in some ways that are questionable.

The abortion-crime piece is a good example. This was debunked, redone with "new &improved data," and recently debunked again. The omission of fixed effects and the use of raw counts on the first run that the book uses is a mistake that is sophomoric (which is why he is criticized by some of "cooking the books" so to speak), not something I would expect to see from someone who just won the ASA award. IMO, it uses a very demanding statistical model to explain a very parsimonious theory - something that shouldn't need this complex of a model. Steve Sailer pointed out in the original slate debate that if you look at the homicide offense rates by age for the first teen &young adult cohorts that fit into post-Roe, their homicide rates are several times that of those from before! It ironically reminds me of Lott's work.
7.6.2006 12:44pm
Jack Marshall (mail) (www):
Bingo. The book is, in fact, a fraud; hardly "dazzling," and often tedious, banal or wrong. "Eureka! Experts sometimes use their expertise to gull customers!" Has Levitt never used an auto mechanic! "Amazing! Inner city crime gangs are organized like General Motors!" Did he never see "The Godfather"? And the absurd conclusion Levitt reaches about how Superman hurt recruitment for the Klu Klux Klan shows that he is embarrassingly ignorant of basic psychology.

Sabermetrician Bill James has done this kind of analysis for years better, wittier, and without discounting the impact of ethical considerations, which Levitt seems to believe don't exist. The fact that so many reviewers fell all over themselves over such a thin and unextraordinary book has depressing implications for the state of the American intellect.
7.6.2006 2:47pm
Colin (mail):
"[T]he absurd conclusion Levitt reaches about how Superman hurt recruitment for the Klu Klux Klan shows that he is embarrassingly ignorant of basic psychology."

How does it show that?

"Levitt seems to believe [ethical considerations] don't exist."

Why do you say so?
7.6.2006 6:13pm
Closet Libertarian (www):
The insights in the book are probably not that new for this crowd, but would be for the general population. I applaud the book on that account: getting economic thinking into the public mind.

I do have an issue with the way the book raises questions. You might notice a connection between cheating teachers and sumo wrestlers after the fact but no one would start research in that manner. Future versions would do well to add a chapter about how economic analysis is really done.
7.7.2006 10:55am
Francisco:
Finally someone says what's been on my mind since I read this book. Especially the comment above that highlights how self-praising it is. The first sentence in the book is illustrative: "The most brilliant young economist in America..."!
7.7.2006 11:57pm
Shirley (mail):
Jonah Gelbach's post above (in his disclaimer section)mentions Mary Rosh. John Marshall, could your name also be a pseudonym for John Lott, Jr. as Mary Rosh is? I ask this because the writing style is the same. John Lott as Mary Rosh is well known in the world of economics as a defamer of Steven Levitt and as a self-promoter anonymously through the use of "Mary Rosh" as his signature. See http://www.whoismaryrosh.com/. Also interesting is http://timlambert.org/lott.
7.8.2006 5:55pm