We’ve got one last bit of evidence in the Lott affair, so here’s one last wrap-up.
After looking back over the evidence compiled so far, I noted that there was some ambiguity on a key bit of information from one of Lott’s former colleagues, an economics professor named David Mustard who, as a graduate student, had collaborated with Lott on a paper [pdf]. James Lindgren’s report contained the following:
John told me that he had conducted a survey in 1997. I did not participate in the survey–it was after our concealed carry paper had been published (Jan 1997) and was after I was on the job market and while I was finishing my dissertation and then moving to Georgia (Aug 1997).
As Lindgren noted, “when Lott told him about it is a little unclear from Mustard’s email,” since it’s not clear whether the prepositional phrase “in 1997” modifies “told” or “conducted.” Email queries to Mustard from both Lindgren and myself elicited replies that were supportive of Lott, but which remained ambiguous about precisely when he had first heard about the study.
Though he did not send me any unambiguous assertion that he remembered being told of it in 1997, he did have a longish conversation with Lindgren earlier today, the details of which he granted Lindgren permission to pass on to me. They are as follows. Mustard is “absolutely certain” that Lott talked about doing spinoffs of the article they wrote together in late 1996. Among these were a survey of the general public on defensive gun uses, the book More Guns, Less Crime, and a study on multiple victims shootings that later became an article [pdf] written with William Landes. He is “fairly confident” that he later was told by Lott, sometime in 1997, that the general survey had been carried out. He has no recollection of when he first heard that the survey data had been lost in the crash, but as he wasn’t involved, he points out that there’s no reason he would have heard about it, though Lott did tell him in the summer of 1997 that unspecified data had been lost in a computer crash.
One final bit of evidence. David Gross, the gun-activist who came forward as a respondent in the 1997 survey, said there was a question on that survey – one Lott hadn’t mentioned — concerning what the subject knew about his or her state’s concealed carry laws. Mustard recalled that this was a point on which their paper was criticized early on: concealed carry laws would only deter if it was widely known just what those laws were. So it’s plausible that such a question would’ve been included on Lott’s survey, even if he didn’t remember asking it.
Now, that is not as clear cut a confirmation as Lott, in his Rosh drag, attempted to make it sound. Indeed, there’s something vaguely shady about implying that someone has unambiguously made a strong assertion about what happened, when you must know, from the horse’s mouth, that the person in question is actually able to offer only a more qualified statement. Still, this is, I think, good enough. To conclude that there was no study, we must be willing to say that Lott, Mustard, and Gross are all lying about their recollections. (These are the only three who provide direct evidence of a survey; of course, there are plenty of others who provide pretty well established details which are consistent with Lott’s account, but not conclusive.) Three people’s testimony seems to me enough to outweigh whatever concerns or doubts we may have about that of any one of them taken alone.
So, what do I think we should conclude? First, that Lott is probably telling the truth, and did in fact conduct a survey. In light of the new evidence, I think people who have publicly concluded that Lott lied should at least make some new statement and, if their judgment comports with mine on this, also be willing to give Lott the benefit of the doubt. It seems to me that we should withdraw accusations of fraud on this point. Second, that confirmation of the survey doesn’t change the fact that given the sample sizes involved, as several bloggers schooled in statistics have noted, the claim about 98% of defensive gun uses involving mere brandishing was irresponsible to the point of being deceptive. Especially if you are not going to cite the data you’re using as a basis for the claim, it is misleading in the extreme to offer a figure that any statistician would consider as meaningless as the report that 100% of D.C. residents are male based on my walking out the door and seeing a group of 5 guys. As my economist friend observed, it’s not clear that presenting a result that you know to be statistically useless as though it were reliable is really any better than just making up a number. So unless there’s some other data out there somewhere that does use a statistically meaningful sample size, Lott should withdraw the “98% brandishing” claim — or at the very least stop repeating it or arguing that his new study “shows” a similar result.
Finally, a quick meta-note on the ‘sphere coverage of this thing, which has been pretty extensive. First, folks who’ve been gleefully posting the aspects of this that looked bad for Lott should probably mention, at least in passing, this new confirmation once Lindgren has had a chance to update his report. Second, though, is that I’ve been a little unsettled to watch the pattern of coverage on this from blogs on the right. Now, obviously, I didn’t expect any particular people to pick this story up. If you’re not interested in it, hey, cool. But there were a few folks who followed it when there were pieces of evidence bolstering Lott’s story, and suddenly forgot it existed when less flattering information came out. And that also strikes me as unfair if you’re expecting people to form judgments based on the information you provide. Ok; may I never write another word on this whole sordid business again in my life.