Talk:Steven Levitt/Archive 1

abortion effect
RichFarmer has twice added a large blockquote (so that is consists of about half of the main article text) of, and externally linked to, the Economist magazine on Levitt and Donohue's paper on Legalized abortion and crime effect. I cut down the original posting to one sentence linked to the internal article and reverted the second. This is a bio article, not an article on the effect of abortion upon crime rates. There already is such an article, Legalized abortion and crime effect, and extensive information about it should go there. (Note that RichFarmer and I appear to be working ourselves up over there, so I'm hardly an uninterested party in this.) The fact that the the working paper that the Economist article is discussing, and Levitt's reactions to it and the Economist article, are not mentioned, basically makes the extended quote using the Economist's storied attitude little more than a slam in my mind. Outside opinions are welcome. - BanyanTree 13:22, 13 January 2006 (UTC)


 * BanyanTree has twice cut out two additions that I made on Levitt. BT only provides an explanation for one of those cuts.  He eliminates both the discussion from the Economist magazine and the discussion of other programming mistakes that he was accused of making in the American Economic Review.  Many bios have similar quotes.  The Economist magazine does a good job of providing a clear discussion and most economists will agree that the magazine is reputable.  No response can be given for the second cut because no explanation for cutting it is given. This is important to understanding Levitt, if he is a careful researcher, and how he responds when people have discovered errors on multiple occasions.RichFarmer 23:41 13 January 2006


 * Ummm.. how? He has responded to it, but you have not given equal time to his response.  Nor to the actual analysis that the Economist is referring to.  You have chosen an extended passage showing the Economist's haughty attitude (Note that I have a subscription to the Economist, so am not arguing that it is wrong, but it does make a point of being as snide as possible, which is amusing but I would never put an extended quote in a Wikipedia article.)  You have not added an analysis of his reaction to criticism; you have added a sarcastic interjection from a article without referring to his written response to it.  Besides the fact that extended block quotes taking up a third of the total article that are externally linked is bad, it neither gives context nor rebuttal.  Let me repeat: this is a bio article, not the article on abortion and crime, and extensive additions on that subject skew this article.  This discussion should take place in the relevant article.


 * Well I am glad that you have an subscription to The Economist, then you know that they have a very good reputation for carefully explaining academic economics articles. They regularly go through recent research.  THe magazine explains the problems in a straightforward way.  Later in the article the magazine gets tough, but given Donohue and Levitt's claim that now the data set needs to be changed, this seems justified, but not the part that was quoted.  Note that I tried compromising with a shorter quote.  You also have not responded to the previous programming errors in other Levitt work.  He also made changes in the data there also.  Strikes many people as a pattern here.  Readers should be given this account.RichFarmer 3:41 14 January 2006

I hope the changes I just made lay this issue to rest. A short, neutral summary is all that is required here - details should be (and are) in Legalized abortion and crime effect. The short quote from the Economist I put in is much more apposite. And unless there's a reason for removing the AER paragraph, leave it in. Rd232 talk 09:33, 14 January 2006 (UTC)

Levitt's response is uncited, and I want to read it. The article makes it sound like Levitt conceded he was wrong. I didn't think that was correct. —Preceding unsigned comment added by 68.197.200.140 (talk) 22:46, 7 July 2008 (UTC)

RichFarmer continues to revert the page to a version he/she likes as opposed to one that is factually correct and portrays Levitt and his career fairly. Compromise has nothing to do with factual correctness. The current version of the article includes the Economist quote as well as discussion of the two mistakes that Levitt has made. And like some other users, I do not understand why the Summary of Works has been removed.

RichFarmer keeps reinserting this claim:


 * Similarly, Ian Ayres and Steven Levitt's paper entitled "Measuring Positive Externalities from Unobservable Victim Precaution: An Empirical Analysis of Lojack," Quarterly Journal of Economics (1998) could not be found by others (e.g., John Lott, "Does a helping Hand Put Others at Risk?" Economic Inquiry, April 2000, p. 257).

This claim is a pure fabrication. There is no evidence for it. The cited paper does not say this. --TimLambert 11:13, 18 March 2006 (UTC)
 * Judging by the abstracts (Levitt paper, Lott paper), you're right. I believe your version is largely superior - some other editors seem too keen to dismiss Levitt's work with a broader brush than warranted. However, the list of papers in your version isn't suitable; the list should be cut down to just references, and the useful descriptions of the papers merged into a couple of paragraphs in the main text. Also, I like the longer version of the Economist quote ("moral turpitude") - it adds flavour. Rd232 talk 16:49, 18 March 2006 (UTC)


 * Dear Anonymous, I agree with rd232 that this page is not a discussion of each paper that was written (many academics have written 100 papers and books) and we had already list many papers, but an attempt to give to readers a general description of what is important and noteworthy about his work. You are right that we can not jeopardize the truth here, and that is why you should not be allowed to remove even the smallest critical notes of the work of Levitt. On the abortion work, Levitt was nailed by the economists at the fed of Boston. If people try to redo the estimates the way that Levitt said they must be done, and they cannot obtain his results, that seems important for talking about Levitt's work. Is the Economist magazine wrong?  Is the Wall Street Journal wrong?  Others have had similar problems with other papers.  You can't just wish these significant problems to go away.  As to TimLambert's claims about "fabrication," the Lott paper is known to many in law enforcement.  It indicates on page 257: "I followed Ayres and Levitt's paper, which identifies when Lojack was adopted so that I could control for both a dummy variable for the presence of the law and a time trend for the nymber of years that the law was in effect.  Although both variables implied that autotheft fell when Lojack was adopted, neither coefficient was statistically significant.  Unfortunately, neither Ayres and Levitt nor Lojack were willing to share data on the number of Lojack devices sold."  TimLambert must not have looked at the paper before he commented with his extremely strong bald claim of fabrication or he did not think that others would check it. In any case, it shows poor investigation abilities. RichFarmer 13:41 18 March 2006
 * Sorry, Rich, but cliches about glasshouses and stones spring to mind. The footnote you quote says it all: Unfortunately, neither Ayres and Levitt nor Lojack were willing to share data on the number of Lojack devices sold. Time is quite likely to be a bad proxy for the impact of lojack, and hence Lott not finding it statistically significant isn't surprising. I'm reverting to TimLambert's version as a better basis for future revision. Rd232 talk 23:20, 18 March 2006 (UTC)


 * Rd232, so why do you cut out or rewrite all the discussions of mistakes made by Levitt? What is wrong with the point made by The Economist or the WSJ?  What is wrong with the mistake in the American Economic Review (and dispite claims to the contrary it was a programming mistake too)?  By the way, if you look at the Ayres and Levitt paper you will see that they do there work two ways.  They use time as a proxy and then they do it with the number of lojack devices.  Lott was using time because that is the way that Ayres and Levitt said they did it and he could not reproduce their results.  This is no different then the fed of Boston people or the McCrary point.  The only common point that you all make is that any mistake by Levitt must not be discussed. In the first two cases, Levitt came back with a changed model.  In the third, as far as I can tell he never responded.  RichFarmer 19:18 18 March 2006
 * TimLambert has covered the Ayres+Levitt / Lott issue below. It is ludicrous to say that we are trying to prevent all discussion of Levitt errors - merely not to exaggerate their significance, or the importance of one or two papers (albeit one of them very controversial) in Levitt's body of work. Rd232 talk 21:26, 19 March 2006 (UTC)


 * Ayres and Levitt use number of years since Lojack was introduced to the market as well as Lojack sales as a proxy. This is not what Lott used -- he used a dummy variable and a trend variable.  His model for crime was also different, since he was not trying to replicate Ayres and Levitt but test to see if including a measure of Lojack changes the results in his paper on minority police hiring.  There is no support for your claim in Lott's paper.  And in any case there have been so many errors found in Lott's other work, it would be wrong to rely on him as a source in any case. --TimLambert 08:48, 19 March 2006 (UTC)


 * You do not explain why you rewrote the section on the errors McCrary found. Explain why these computational errors McCrary found are not the same as those found by the fed of Boston people.  On the other case, you go from stating that the Economic Inquiry paper did not do any of this to stating that Lott is not to be trusted.  So what one is it?  You are not consistent.  Do you want to cut The Economist magazine quote because The Economist magazine is not to be trusted? RichFarmer 12:56 19 March 2006


 * I've reported you for breaching WP:3RR after a polite reminder. I also strongly object to your reversion of quite a lot of work I did this afternoon. (I've now added back the Economist quote which got lost along the way.) I don't know what TimLambert meant by the unreliability of Lott, but it's clear that Lott's paper doesn't (isn't trying to) invalidate Ayres and Levitt. As for your question above about the similarity about the McCrary / Fed Boston errors, it would help if you could be more precise. And please discuss here rather than revert. Rd232 talk 21:22, 19 March 2006 (UTC)

Fair Use Image
Is it proper to use the fair-use book cover image in the author's article? It shouldn't be!--69.208.166.69 05:22, 13 August 2006 (UTC) Why should one care if Levitt is Jewish/historically jewish? His bio is related to is research and none of his academic research addresses a 'jewish' concern.