Deltoid

Natalie Solent is disappointed in Lott, but is still impressed by More Guns, Less Crime. Unfortunately, the 98% figure is not an aberration. It is typical of the remarkable carelessness with facts that Lott displays and his refusal to back down even when obviously in the wrong. For more examples, see here and here.

Steve Verdon is continuing to work his way through Ayres and Donahue. He has found a misprint in one of their tables where some coefficients are missing and wonders what I would say if I found something similar in Lott and Mustard. Well, I did find something similar in Lott and Mustard (the coefficient for murder in footnote 49 is ten times what it should be) and didn’t say anything about it in my critique. I hope Steve can come up with some substantive criticisms of Ayres and Donahue.

The Washington Post has a story about Mary Rosh. Lott now claims that this review of More Guns, Less Crime was written by his 13 year-old son with some help from his wife.

“They told me they had done it. They showed it to me. I wasn’t going to tell them not to do it. Should I have?”

One of Mary Rosh’s reviews (the one of Caesar 3) reads like it was written by a child, the review of More Guns, Less Crime does not. It also seems unlikely that a 13 year-old would have loaned out his copy dozens of times.

Now, compare Rosh’s review:

This is by far the largest most comprehensive study on crime, let alone on gun control. Professor Lott examines crime rates as well as accidental gun deaths and suicides for all 3,056 counties in the United States by year for 18 years. By comparison, the previous largest study on gun control examined 170 cities within one single year 1980.

with Lott on “Uncommon Knowledge”:

The book is the largest study by far that’s ever been done on crime, let alone on guns. The largest previous study looked at 170 cities within one year, 1980. My research looks at all 3000 plus counties in the United States for crime rates, accidental gun deaths, and suicides by year for eighteen years.

It is possible that Lott’s son created the review by plagiarizing from Lott’s writings and posting it with Lott’s approval, but even if this is true, I don’t think that it differs in any meaningful way from Lott writing the review.

Mark Kleiman is disgusted by Lott’s attempts to blame his 13 year-old son for the Rosh review of More Guns, Less Crime. Kieran Healy is disgusted too, and has a nice example of an honest review of a parent’s book. Tom Spencer and Roger Ailes are also disgusted. Greg Beato, meanwhile, is merely sarcastic. skippy also comments.

Meanwhile, Glenn Reynolds notes that this was “Another story broken by a blogger”. What Reynolds does not note is that he decided not to mention this story on his blog because it wasn’t “actual news”. And he still hasn’t linked to the post where Julian Sanchez unmasked Mary. Meanwhile, Julian mentions that there will be a story about this in the February 10 issue of US News and World Report.

In the Washington Post article Lott says:

“I probably shouldn’t have done it—I know I shouldn’t have done it —but it’s hard to think of any big advantage I got except to be able to comment fictitiously,”

Well, I can think of one.

Last January, the New York Post published an opinion piece written by Lott. In that piece Lott claimed that a school shooting had been stopped by students armed with guns and that almost all the newspaper stories had failed to mention this fact, thus demonstrating that the media showed a bias against guns. Next, someone posted the Lott piece to Usenet. A long discussion ensued, with a gentleman named Ed Huntress criticising Lott for failing to mention that the students with guns had actually been police officers, and Mary Rosh stoutly defending Lott. In February, Lott’s piece was published in the Fort Worth Star-Telegram with one and only one significant change: the addition of these words:

Many stories mentioned that the heroic students had law enforcement or military backgrounds

Mary then posted this version (artfully claiming that there were a “couple” of significant differences) and suggested that the New York Post had edited out that phrase to make the piece fit in the available space. Mary also demanded that Ed Huntress call Lott and apologize since the error was the New York Post’s fault. Ed called Lott and reported back:

I talked to John Lott and learned that he hasn’t even seen the New York Post’s edited version of his editorial.

That’s just a brief summary. You should read the whole discussion to really appreciate what an enormous lie Lott told.

So, what big advantage did Lott obtain by his Mary Rosh deception? He made a major omission from his piece. Instead of having to take responsibility for his actions, he was able to blame the New York Post.

This isn’t the only time that Lott has attempted to get out of a jam by rewriting history. Consider:

• the story that the 98% came from a never-before revealed survey after years of attributing it to other sources and the denial that he ever attributed to other sources.
• the attempts to change the story he told Lindgren and the insistence that he had not changed his story.
• changing the story about the survey so that it took one month to complete instead of three months after writing: “I am willing to bet that I don’t start mentioning this [98%] figure until the spring of 1997. If I use it before I said that I did the survey, I will say that they nailed me.”

All this suggests that Mustard’s late recollection that Lott had told him in 1997 that he had done a survey may have been caused by Lott insisting that he had definitely told him then and Mustard being less resistant than Lindgren to Lott’s history rewriting.

Atrios mentions the Washington Post article on Mary Rosh. Meanwhile, Calpundit reports that Lott has backed out of doing an interview. I guess Lott is never going to answer these questions. Julian Sanchez has an update where he observes that over in talk.politics.guns some folks, having seen Lott’s confession, three posts from Clayton Cramer and the Washington Post article have formed the only possible conclusion: there is a massive forgery campaign underway. Julian also points us to the US News article on Lott and Rosh.

Arthur Silber summarizes Lott’s appearance on the Larry Elder. They don’t seem to have gotten much past the use of the pseudonym (which in itself is perfectly OK). The problem was what he posted under the pseudonym. And be sure to scroll down to the comment section for some more good comments from Julian.

Atrios, Tom Spencer and Roger Ailes comment on the Rosh-Huntress files. Tapped and John Quiggin also have comments.

Amazon has removed Mary Rosh’s reviews. Good thing I saved a copy.

Atrios points us to Tim Noah’s article at Slate. After the Washington Times whitewash, and the US News and Washington Post completely ignoring Lott’s survey, we at last have a mainstream media article that gets to the heart of the matter.

One interesting feature that bears repeating because it is hard for it to sink in because it seems so unlikely: Lott will not admit that he attributed the 98% figure to “national surveys”. Look at what he tells Slate:

“A lot of those discussions could have been written more clearly.”

Lott is saying that this sentence does not attribute the 98% to “national surveys”:

“If national surveys are correct, 98 percent of the time that people use guns defensively, they merely have to brandish a weapon to break off an attack.”

and this sentence does not attribute the number to “Kleck’s study”:

“Kleck’s study of defensive gun uses found that ninety-eight percent of the time simply brandishing the weapon is sufficient to stop an attack.”

They just “could have been written more clearly.”

Julian Sanchez comments on the Slate article with two points, one for Lott, and one against.

1. That Lott discussed his 1997 study during a January 1999 talk that Gross attended, and since that preceded Duncan’s raising the 98% question later that year, the tape of that talk would be good evidence that there really was a survey. However, Tim Noah has misunderstood David Gross. Gross kept the tape of the talk, so we know exactly what Lott said. He presented the 98% statistic without giving a source, and Gross formed the notion that the source must have been the survey where he was asked about his own defensive gun use. After the talk he mentioned this to Lott, and even then Lott did not say that it was his survey. This part of Gross’s story is actually evidence against Lott, and is one of the reasons why I’m inclined to believe Gross—if he was making it up, you think he would have provided more support for Lott.
2. Why didn’t Lott re-enter the data from the tally sheets? Good question. He was still at Chicago in May 1999 when he wrote to Duncan claiming to have done a survey, so presumably he still had the tally sheets then.

A couple of alert readers have pointed out that while all the reviews have been removed from Mary Rosh’s page you can still read her review here.

Mary Rosh got a brief mention on CNN Crossfire.

Michelle Malkin writes an excellent article on the Lott affair. And if you think that she is one of those mysterious people out seeking revenge for Bellesiles, you should look at this 1998 article where she praises Lott.

Atrios explains why he cares about Lott. He quotes Sullywatch:

We forget now how much there was an all-out effort (kind of like a certain recent special prosecution) to throw anything they could find at Bellesiles until it stuck, and finally one thing relatively marginal to the whole thesis of the book did.

I can’t agree with this statement. If you look at the report of the Emory committee you will see that they found Bellesiles guilty of unprofessional and misleading work. This destroys the credibility of the rest of his work. And while the probate data was only a small part of his book, this was the part that was mentioned most frequently in the media.

Glenn Reynolds has a post on the Slate piece where he offers exactly the same argument in support of Lott that Sullywatch offered in support of Bellesiles. The surprising thing is that despite being privy to months of discussion about this on firearmsregprof, and presumably having actually read the Slate piece, and having been corrected on this very point before and, actually quoting a sentence of mine that mentions that he claim was in Lott’s book, Reynolds still somehow believes that Lott’s 98% claim is not in his published work. Fortunately two of his readers wrote in to correct him. However, instead of admitting his error Reynolds uses an out-of-context quote from me (without providing a link so readers could see all of what I said) to argue against them. Reynolds also pretends that the only proven charge against Lott is using a pseudonym. However, he has also been caught lying, and even if he did do some sort of survey, he is guilty of unprofessional and misleading work for falsely attributing his 98% claim to other surveys, for advancing his 98% number even though the sample size in his survey was too small to allow a meaningful estimate, and for never once giving the estimate from surveys with an adequate sample size.

Roger Ailes and Charles Murtaugh also observe that Reynolds’ Jon Ellis comparison is incorrect. Murtaugh also offers a better argument than Reynolds that the scale of Lott’s deception is less than that of Bellesiles.

Tom Spencer has comments and Julian Sanchez has an update.

If you want to see another example of Lott’s carelessness towards facts, consider this article, published a few days ago:

But, where Vernick and Hepburn said they were unable to find any attribution to the ‘20,000′ statistic, Lott said the proof is readily available in a compendium prepared twice a year by the Bureau of Alcohol, Tobacco and Firearms. He did say the gun law information takes a tremendous amount of time and energy to compile.

“About ten years,” Lott said, “somebody actually took the time to go and do this.”

Lott said final analysis of the B.A.T.F. report found approximately 20,000 gun laws to exist in the U.S. and he suspects more federal, state and local laws were added or amended by lawmakers over the past decade.

Lott claims that it took someone ten years to count the number of laws in a compendium. In fact, using the power of random sampling, you could count them in an afternoon. All you have to do is choose some pages at random, count the number of laws on each of these pages, find the average and multiply by the total number of pages in the compendium.

Even if for some reason you wanted to count them all, it would not take ten years. If you generously allow ten seconds to count each law, it would take about a week of eight hour days to count 20,000 of them.

So, Lott either made up or mis-remembered the “ten years” story, but confidently related it to the reporter as if it was true. He also didn’t have the good sense to realize that his story was obviously untrue.

Mary Rosh’s famous review has made it onto the Fallacy Files.

I should add one caution to my comments about Lott’s “ten years” claim yesterday. It is possible that the reporter misunderstood and/or misquoted Lott.

Roger Ailes quotes a new review of More Guns, Less Crime. My unbiased opinion is that these two reviews are better.

And here are a few comments I missed earlier: Adam White, William Sjostrom, John T. Kennedy, Dr Limerick and pro-gun activist Timothy Wheeler.

John Quiggin has a thoughtful post on the parallels between the Bellesiles and Lott affairs. Meanwhile, Charles Murtaugh, responding to this Tapped piece reckons that there is an important difference: there are pro-gun people like Michelle Malkin criticizing Lott, but there weren’t pro-control people criticizing Bellesiles. He’s wrong. Consider, for example, this Wall Street Journal article by Kimberley Strassel on Bellesiles:

It’s worth pointing out that not all of these professors have an obvious political agenda. Jim Lindgren, Gerald Rosenberg, Erik Monkkonen and Randolph Roth all prefaced their remarks by saying they favor gun control

Notice that Strassel is mentioning the pro-control critics of Bellesiles for exactly the same reason that Tapped mentioned the pro-gun critics of Lott: to show that the criticism isn’t politically motivated.

In response to my asking who the people were who were seeking retribution for Bellesiles affair, C. D. Tavares writes:

The stench of “sore loser” permeates your jihad, right on down to the little website shrine you have crafted for it. It may make you feel like Clayton Cramer, but you’re no Clayton Cramer.

The Washington Post article exposed an organized anti-Lott campaign by its very existence. Even given the paper’s well-known bias, a pissant “story” like this doesn’t make the Post unless it’s being pumped harder than Pamela Lee Anderson.

Don’t consider this a private communication—feel free to repost it anywhere you wish.

Umm, if I’m so “sore” about Bellesiles and out for retribution for him why did I post this over at the History News Network back in August?

I think the reason why the Lindgren critique of Bellesiles is so effective is that he confines himself to just those sort of points where we can say with certainty that Bellesiles is incorrect.

Incidentally, Glenn Reynolds has made a copy of Lindgren’s piece available at http://www.instapundit.com/archives/003080.php#003080

And if there is an organized anti-Lott campaign, wouldn’t they have contacted me by now?

Another link I missed: TBOGG criticizes Glenn Reynolds’ attempts to downplay the Lott affair.

kuro5hin has a story on Lott/Rosh. There is even an on-line poll. At the time of writing the results were:

John Lott is..
.. a fraud.	50%
.. a good researcher who made some mistakes.	5%
.. victim of a vast left-wing conspiracy.	13%
.. transgendered.	30%

Of course, the results from an on-line poll like this have no more scientific validity than Lott’s 98% statistic.

Blogroots also mentions the unmasking of Rosh. The High Road (a pro-gun message board) has a discussion on Lott. I think this comment is interesting:

My problem with Lott now is his “the dog ate my research” story. I laughed when Bellesiles told that story because I didn’t believe it for an instant. It’s not so funny now, for the same reason. Incorrectly citing Gary Kleck is another problem. These aren’t questions about his tactics or about Mary. It’s about his data and how careful he is with the facts.

Lott says he’s trying to replicate the lost experiment. Even if it confirms the lost one, I’ll be more skeptical this time, and I’m someone who has repeatedly cited Lott’s work. I want to believe him, but I just can’t. Entire research projects don’t disappear without a trace in a computer crash, even if you’re stupid enough to fail to back up such a large amount of important work.

Kevin Drum observes that Mary Rosh has become a TV star.

Don Watkins thinks Lott should be ashamed of himself. Steve also is not impressed.

Otis Dudley Duncan and Tim Lambert

The discussions on so many blogs on the Lott case have been invaluable not only in turning up new evidence but also in illustrating the wide variety of reactions to the evidence. One thing that strikes a professional researcher is that many who have commented do not look at such cases in the way that researchers are trained to do. What this case is about is the professional work of a a social scientist, and the question is whether that work meets the ethical standards of scientific inquiry and reporting. And commentators are not at liberty to define for their own purposes just what the relevant ethical precepts are. A great deal of thought has been given to the codification of the main ethical principles, violation of which constitutes scientific misconduct. We think that the quality of comments would be enhanced if the writers took a little time to look at some of the formulations that have been proposed. The American Society of Criminology has a page that gives links to several of them.

(more…)

Otis Dudley Duncan makes an important point: whether or not Lott actually did some sort of survey, Lott is guilty of scientific misconduct. Duncan has uncovered the case of a Dr Duan, who was suspended for two years for publishing a study for which the supporting data had been lost.

Lott’s wife has posted (I have confirmed via email that it was really her) to the comment section of this Electrolite post:

When the screen name is used, it always, automatically registers as MaRyRoSh, NOT as MaryRosh or Mary_Rosh, which I am sure must have suggested to some that this was some amalgam and not a Mary who happened to have the highly unusual name of Rosh.

The screen name was originally used by the boys for messages within the scout troup, for ordering old coins on the internet, and for posting some book and game reviews.

At some point later, each son got his own screen name and MaRyRoSh was rarely used by them anymore. As there would occasionally be some e-mail coming in and it did not cost anything to keep the screen name, I never bothered to delete it.

So when my husband later ended up using MaRyRoSh (which I was not even aware of, as I did not check that mailbox), he did not have to use much imagination to take up the fictional character of Mary who is a student of his (which of course our sons are). So much for the tranvestite or other pop-psycological spin in the media.

Of course, she’s mistaken about the name used being MaRyRoSh. The Amazon review was by maryrosh and the AOL email was MaryRosh@aol.com. As for the psychological side of it, Mary’s talk of being raped and wearing heels, were, at least to some extent, a ploy to win an argument, but who can tell if there weren’t psychological overtones as well?

The Chronicle of Higher Education has an article (subscribers only) on Lott, Rosh and the 98%. If you have been following the affair, the only new piece of information is Lott’s explanation for this Rosh post where Rosh attempts to find out who reviewed one of Lott’s papers:

Mr. Lott says that those questions were “purely rhetorical” and that he was simply trying to taunt “Alpha Male” into confessing that he is not an academic and had never actually reviewed any of Mr. Lott’s papers.

And Rosh’s review seems to be completely gone from Amazon.com now.

Lott’s wife, Gertrud Fremling, has responded to a question I put to her about the similarity between Rosh’s Amazon review and Lott’s writings.

Obviously ” …this is the review:” is a false statement by you. You should have said that “… this is part of the review:” Am I supposed to believe that this was a mere error? And I notice that you have selectively quoted from my website, too, without indicating that these were selected parts. If this is how you choose to deal with me, then I have no reason to continue any debate.

I will answer the general query about the book review once and for all: Especially seeing the left-out part makes it very clear that it is not my husband who wrote it. I even notice the errors of not making a space after the period a couple of times - obviously not an experienced typist wrote this material. As to the part that supposedly is so similar, I do not think that it really is. These are very basic facts he mentions. You must also realize that when writing the book review, my son must have looked at some written material from the book and elsewhere. After all, you don’t just pull the figure 3,056 right out of your head. Plus, he must have overheard his dad on the phone a huge number of times when being interviewed by the media.

By using the phrase “selectively quoted” she seem to be implying that yesterday I somehow took her comments out of context. I did not. I left out the part that contained information (about the derivation of the name Mary Rosh) that should already have been familiar to someone following the affair. Nor is it normal practice when quoting from a document to indicate that you have not quoted the entire thing. I also included a link to her comments so any interested person could read the whole thing.

She suggests that the similarity between Rosh’s review and Lott’s writing might be because both are expressing the same facts. However, the Rosh review expresses them using the same phrases that Lott uses. It is not just the facts, but the wording. And if Lott’s son copied the sentences from something Lott wrote, then we have a review at least partly written by Lott and posted with his approval. This does not differ from a review written by Lott in any important way.

The Washington Post has a short item on Lott and his survey.

In response to this story in the Washington Post, Lott has apparently orchestrated a letter writing campaign. Eugene Volokh has posted the four letters. Julian Sanchez points out that all four letters make the same incorrect claim: that the Washington Post is questioning whether Lott had a disk crash. In fact, the article is questioning whether Lott lost his survey in the crash.

In his letter, John Whitley also makes a couple more errors:

I am not an expert on the Dr. Bellesiles case, but my understanding is that there was little or no contemporaneous corroborating evidence of the flood that he reported as destroying his records and that his results were not reproducible by other scholars in the field.

No-one disputes that there was a flood at Emory. What people question is whether the flood destroyed Bellesiles’ records. No-one disputes that Lott had a disk crash. What people question is whether the crash destroyed Lott’s survey data.

Dr. Lott himself has now reproduced the survey and released the names of all people who worked on it and the phone records from the calls. The results are largely in line with his previous results and no one has questioned the integrity of the new survey.

The sample size in Lott’s new survey is much too small for any meaningful measure of the brandishing percentage. It is ridiculous to claim that it is a replication. Whitley also seems unaware of the nine published surveys give numbers ranging from 21% to 67% as to how often defenders shoot, which most definitely do not reproduce the 2% shooting Lott claims.

Archpundit has a few choice comments about Lott. Postwatch makes the same error as Whitley: Lott’s new survey does not bolster his 98% claim—the sample size is just too small.

Another letter has been sent to the Washington Post, and by an amazing coincidence makes the same error about the Post article as the previous ones:

A column appearing in the Post yesterday (Feb. 11, “Fabricated Fan and Many Doubts”) implies that economist John Lott made up the claim that a computer malfunction destroyed data from his research on gun control. At the time Lott was engaged in this research, we were colleagues at the University of Chicago Law School. I clearly recall John relating the computer data-loss incident to me then—many years before the current controversy about his work arose. Just so you know, I’m not relating this information to you because I support Lott’s position on guns (I don’t). I’m relating it to you because I think journalists—even the ones you employ to write political gossip columns like this one—should live up to their professional obligation to check out the facts before they make claims harmful to an individual’s reputation.
Dan M. Kahan, Professor of Law, Yale Law School

Meanwhile, Glenn Reynolds thinks that the letters “have some probative value”. This is not true, since no-one was questioning whether Lott had a disk crash. The only thing the letters are evidence for is the extent Lott will go to in order to make it appear as if the argument is over whether he had a disk crash.

Glen Whitman argues that we should not dismiss Lott’s work on concealed carry laws just because of his misconduct about the survey. This is true as far as it goes. Others have verified that if you use his models and his data, you get the numbers he reports. The problem is that many critics, most recently Ayres and Donahue have presented good evidence that Lott’s models are wrong, and when better models are used Lott’s results vanish or are reversed. Now Lott disputes their findings, but his behaviour over his 98% claim demonstrates that whatever the truth of the matter is he will never admit to being wrong.

John Quiggin notes some parallels between John Lott and Trent Lott cases.

Postwatch complains that I misunderstood his statement about Lott bolstering his 98% claim. Apparently he just meant that Lott is attempting to bolster his claim. If that is all he is saying then we have another parallel with Bellesiles, since Bellesiles is also attempting to bolster his claims. You can visit his web site of probate material here.

Roger Ailes makes an good point—now that Lott’s mystery survey is getting wide publicity, why haven’t any of the students who allegedly did the interviews come forward?

Archpundit points to a Valentine poem in The St. Louis Post-Dispatch. Kieran Healy has a Valentine poem too.

Ron Grossman has a story about Lott and his survey in the Chicago Tribune. I have three comments on the story:

1. This is the first newspaper article that mentions that Lott is looking for the students who conducted his survey and it’s in a Chicago newspaper. You would expect that some of the students who conducted the survey would still be in Chicago and hence likely to see the article. Also, with two million readers, you would expect about twenty of them to have participated in the survey. If nobody comes forward after this article, then that is strong evidence against Lott.
2. Otis Dudley Duncan first tried to make this matter public in 1999. It seems to have been just as hard to get the press interested in Lott’s survey as it took to get them interested in the Bellesiles affair.
3. Kleck is rightly sceptical of Lott’s claim that he has replicated his 1997 survey. Kleck estimates that it would have taken 500 evenings to conduct the new survey instead of the 10 that Lott claimed. The factor of 50 discrepancy is explained by three things. First, Lott had about ten interviewers instead of “a couple” as Kleck was told. That is a factor of 5. Second, Lott did not re-create his 1997 survey—the sample size was halved. That is a factor of 2. Third, Lott did not attempt to conduct serious research in defensive gun uses. Instead he asked just enough questions so that he could get a brandishing number. That is a factor of 5 difference in the amount of time that the survey took. Multiplying these three factors together accounts for the discrepancy. Lott did do a survey last year, it’s just that it wasn’t a serious research effort—it’s purpose was to have something to point to and claim to have replicated the mysterious 1997 survey.

Ken Parish demonstrates the effectiveness of the Lott-orchestrated letter writing campaign. Ken believes that the evidence in the letters make it significantly more likely that Lott did some sort of study. However no-one was questioning whether Lott had a disk crash—the letters do not tell us anything we didn’t already know.

Lott’s survey gets a brief summary over at the History News Network’s History Grapevine.

A couple of weeks ago I described how Lott used his Mary Rosh sock puppet to blame the New York Post for the fact that an article he wrote omitted to mention the fact that the students who captured a school shooter were police officers.

A couple of months later, however, in Lott published his article again, without mentioning that the students were police officers. So we have proof not only that the Post was not to blame for the omission, but that the omission was quite deliberate and not accidental.

Atrios points us to a Mary Rosh posting at freerepublic.com where she urges folks to rig the download counters at the Social Science Research Network by downloading a Lott paper as frequently as possible.

Some commentators have been saying that the 98% is the only dodgy thing Lott has done. Actually there are plenty more. For another example, look at how he fudged a graph in More Guns, Less Crime.

General Christian has a few, slightly sarcastic, comments about John and Mary.

Over at the History News Network, Clayton Cramer has a post where he comments on the parallels or lack of same between the Lott and Bellesiles affairs:

1. There were legitimate questions raised about the 1997 survey, most notably by Tim Lambert. Proving that the survey didn’t take place is impossible, unlike Bellesiles’ problems, which often involved documents that were easy to find.

It’s also easy to find out what “national surveys” and Gary Kleck’s study say. Lott’s claims here are wrong and he won’t even admit to making them.

2. Lott has managed to scrape together enough evidence now (including replicating the 1997 survey results with a new, better document survey) that I think even Mr. Lambert recognizes that the 1997 survey probably did happen. Mr. Lambert’s criticism now is that the survey size is so small that the 98% number is meaningless, and I generally agree with him on this.

Lott’s new survey does not replicate the results of the alleged 1997 survey. The sample size is too small for it to be evidence for anything. Since no students have come forward after the Chicago Tribune article, it looks increasingly likely that there was no 1997 survey.

3. The parallel between Lott and Bellesiles is strongest in that Lott lied when he posted as Mary Rosh, and claimed to be a former student. Pseudonyms are fine on the Internet, and part of a proud American tradition; lying about who you are, is not.

I actually don’t have much problem with Rosh claiming to be a former student of Lott’s. It’s the other lies he told, like this one.

4. The 98% number is ONE SENTENCE out of Lott’s book. If Bellesiles’ problems had been one sentence, or one paragraph, nothing would have happened to him. I wouldn’t have pursued the matter, and neither would anyone else.

If the 98% claim had just been one sentence I don’t think that anyone would have pursued the matter either. The problem is that Lott repeated the claim over 50 times.

Rob Levine has a very critical article on Lott at citypages.com.

Here is the story behind the removal of Rosh’s review of More Guns, Less Crime from Amazon.

No new developments in the Lott case today. This is bad news for Lott. The Chicago Tribune story was published a week ago. If Lott had a dozen students doing interviews for a month in 1997, you would expect them to have told their friends about what they were doing. If each interviewer told three different friends about it, we have a pool of about 50 people who could tell us about the survey or put us in contact with someone who could. I would guess that roughly one third of these, about 16 people, would still be in the Chicago region. The Chicago Tribune claims to reach 46% of college graduates in the Chicago region, so you would expect 46% of 16 or about seven of these people to have seen the article. None of them have come forward. The numbers here are only approximate so this is not dispositive, but this points to Lott not having done a survey.

Mike Magnum writes a funny article about Mary Rosh.

John Quiggin concludes that “with a high degree of confidence that no survey was ever undertaken”, but suggests that declining ethical standards at the American Enterprise Institute mean that Lott will keep his job for now.

Atrios and Tom Spencer link to the Rob Levine article I mentioned a few days ago. Tom also writes: “Many are saying that it’s probably only a matter of time before [Lott] loses his position at the American Enterprise Institute.”

Kevin Drum observes that even an 11-year-old kid seems to understand the problem of having too small a sample size in a poll. The Sunday paper here printed a supplement giving property auction clearance rates by suburb. Except, of course, for those suburbs where there had been less than 20 auctions, where it printed “nsr” (not statistically reliable). An 11-year-old gets it, the Sunday paper gets it, why doesn’t Lott?

Archpundit and Tom Spencer comment on Lott’s fudged figure.

Costa Tsiokas discusses Lott’s integrity.

The ad that Lott said was placed in the December University of Chicago Magazine has appeared in the February 2003 issue: (The delay was apparently caused by a mix up.)

Attempting to reach two people who worked on and helped coordinate others on a survey given by John Lott while they were undergraduates at the University during the beginning of 1997. Some questions about verifying the survey need to be answered. Please call 202.862.XXXX.

(I’ve replaced part of the number with Xs—if you are one of the two students that Lott seeks, email me and I will send you the number.)

compiled by Otis Dudley Duncan and Tim Lambert
revised 26 Feb 2003 by Tim Lambert

This pages documents direct and indirect references Lott has made to a survey he claims to have carried out in 1997. Further analysis is here, and the latest update is here.

The information of over 2 million defensive uses and 98 percent is based upon survey evidence that I have put together involving a large nationwide telephone survey conducted over a three month period during 1997.
Letter from John Lott to Otis Dudley Duncan, dated May 13, 1999.

There are 15 national surveys that have been conducted by academics as well as polling organizations like the Los Angeles Times and Gallup, and their average estimate indicates that people use guns defensively well over two million times each year. My own survey put the defensive uses at about 2.1 million in 1997. Both numbers are about five times greater than the 440,000 crimes that the FBI reports were committed with guns in 1997.
John Lott, May 25, 1999, Letter to the Editor, The Wall Street Journal A27

(more…)

BuzzFlash has an interesting story which details some more examples of apparent dishonesty by Lott.

I was able to check one of them myself: Mary Rosh’s defence of Lott’s statement that the “the worst thing people can expect from dioxin is a bad rash”. Rosh argues that this isn’t Lott’s claim, but that of Michael Fumento, whose book Lott was reviewing. However, if you read Lott’s review, it is quite clear that he makes the claim his own. And if you read Fumento’s book, you will also see that Lott exaggerates Fumento’s position. Fumento argues (convincingly, in my opinion) that the dangers of dioxin have been grossly overstated, that while it might possibly be carcinogenic, the evidence for this is weak. But he is not saying that is safe to put it on your cornflakes.

Henry Farrell observes that while Joyce Lee Malcolm’s Reason article about Bellesiles describes several cases of misconduct other than Bellesiles, it does not mention Lott. However, this article is from the print edition, so it is likely to have been written before the Lott affair became public.

Dick Dahl has a story about the new book Evaluating Gun Policy. Meanwhile, Handgun-Free America are going to protest against Lott at the AEI Annual Dinner.

The Publisher’s Weekly review of Lott’s new book is up at Amazon.com. I liked this bit:

Even measures to keep guns away from children, like “gun-free school zones” and “safe storage” laws that require guns to be locked away, are misguided because children need guns for self-defense (he cites news reports of kids as young as 11 gunning down criminals).

I came across this quote from Lott in a May 28, 1999 article by Jeff Jacoby:

“There are 15 national surveys that have been conducted by academics as well as polling organizations like the Los Angeles Times and Gallup, and their average estimate indicates that people use guns defensively well over two million times each year. My own survey put the defensive uses at about 2.1 million in 1997.”

I thought that this could potentially help Lott. If the source for the quote was something that Lott wrote before Duncan raised questions about the 98% statistic in April, then that would be good evidence that Lott did not invent his survey to provide a source for his 98% claim.

Jeff Jacoby promptly answered my query—the quote comes from a May 25 letter to the editor in the Wall Street Journal. This means that it was written shortly after the May 13 letter and the May 21 phone call to Duncan where Lott first claimed to have conducted a survey. Instead of helping Lott, this just makes things more suspicious. The survey was supposedly conducted in early 1997. Why did the first two mentions of it occur within a few days of each other in May 1999? Then, on June 23 1999, in response to a Usenet posting of mine, Lott sent me an email that stated:

The 2.1 million defensive uses is the average of 15 private national survies on this topic. It is also the same as the estimate that I got from a survey that I did in early 1997.

I was rather surprised to get this email. We had exchanged a couple of emails in 1998, but had not corresponded since then. It looks as if he mentioned the survey in his letter to the paper and email to me just to provide backing for what he told Duncan.

Glenn Reynolds links to a Richmond Times-Dispatch editorial that is very critical of Lott.