By Jesse Marx
By Chris Parker
By Jake Rossen
By Jesse Marx
By Michelle LeBow
By Alleen Brown
By Maggie LaMaack
By CP Staff
Four years ago the local conservative think tank Center of the American Experiment (CAE) brought a lecturer from the University of Chicago named John Lott to speak at a luncheon. Lott, who is underwritten by the same John M. Olin Foundation that partially underwrites the CAE, gave a speech titled "More Guns, Less Crime: Understanding Crime and Gun Control Laws," which also happens to be the title of his 1998 book. Though counterintuitive, Lott's thesis that arming the populace would make us all safer dominated the gun debate at that time, and today as well.
In addition to appearing at conservative-think-tank-sponsored events across the country, he testified before numerous legislatures and appeared on talk shows--many of which were also funded by conservative foundations--arguing that if states would only relax concealed-weapons laws they would become safer places. This wasn't an opinion, Lott testified and lectured, it was science. (This science, he might have added, was sponsored by the philanthropic arm of the Olin Corporation, which owns Winchester, the largest ammunition maker in the U.S.)
At the time of its release, Lott's research had been heavily criticized, although not in the mainstream media. The central problem, or so it appeared at the time, was Lott's use of complex data models to infer causation from correlation. Lott claimed to have proven that when states in the U.S. let more people carry concealed weapons, citizens were actually safer; therefore, more guns caused less crime.
Critics were quick to point out that correlation does not equal causation, because there are often other unknown variables that may have figured in any supposed "effect." How to account for these unknowns? In this case, what if crime went down for other reasons?
Lott, like all social scientists of the numbers-crunching persuasion, creates complex statistical models with multiple variables to attempt to account for other potential causes. But in the matter of gun and crime statistics, it's a hopeless task. As Ted Goertzel of Rutgers University has written, "The data available for this purpose simply is not up to this task, and the studies have consistently failed." By tweaking Lott's assumptions in minuscule ways, researchers have been able to show that states that relaxed gun control laws actually became less safe. The point is the folly of using such models for complex, politically charged social science research. They are unreliable and subject to abuse. And the specific models used by Lott are riddled with logical holes.
For example, after performing all of his statistical hocus-pocus on an already thin data set, Lott came up with the astounding murder rate of -2.3 for Missoula County, Montana, for 1979 and '91. As Goertzel wryly noted, "This is odd, since a county's murder rate cannot go below zero, unless previously murdered people are brought back to life."
Now, four years later, Lott--currently a "scholar" at the American Enterprise Institute, another body heavily underwritten by the Olin Foundation--is in real academic hot water. He stands accused not just of abuse of statistics, but of refusing to admit that he made a critical mistake, and of actually fabricating research claims to cover his error. Amazingly, he also recently confirmed that he had created a fake online persona to prop up his reputation and hammer his critics.
The current controversy begins with Lott's claim in the first edition of his book that "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." [Emphasis added.] After the book's publication other gun researchers who had also used the 98 percent figure realized that they had made a mistake in the interpretation of another study. Experts then gave wildly differing figures for the effectiveness of flashing a gun, from a low of 21 percent to a high of 67 percent. Nearly all agreed the 98 percent figure had been a mistake.
Except John Lott, that is. In the second edition of his book, rather than correct the previous mistake, he attributed the 98 percent figure to a study he claimed to have done himself. When confronted with the discrepancy and asked for data from his study, Lott claimed that he had lost all the data in a hard drive crash. Despite the absence of any proof that such a study had ever been conducted, gun advocates stuck by Lott.
In addition Lott, besieged by suspicious Internet users, recently admitted to the Washington Post that his most voluminous and adamant defender in cyberspace, one Mary Rosh--who posted hundreds of messages praising Lott--was none other than Lott himself.
Lott himself had personally and publicly disavowed newsgroups and online bulletin boards; he meant to remain above the fray. Yet in the guise of Rosh, allegedly a former student of his, he alternately praised himself as "one of the young stars in the profession" and assailed his critics in the assumed voice of a small, defenseless woman: "If a woman is being attacked by a 200-pound man, is she just supposed to wait until the police arrive? I am 114 lbs. and 5'6". What should I do in that situation?" (Turn back into a man--fast!)