In 1997, economist John Lott latched onto a statistical fluke, and the gun industry have spent 15 years making that fluke into public policy. Meanwhile, the death toll from gun violence has climbed ever higher. (More)

John Lott’s 1997 Journal of Legal Studies article Crime, Deterrence, And Right-to-Carry Concealed Handguns has become tantamount to scripture for gun industry supporters. Using national data from 1977 to 1992 and complex statistical regression analysis, Dr. Lott claimed to prove there were fewer violent crimes in states and counties where a higher percentage of people carried concealed handguns. This was music to the gun industry’s ears, and in the years since so-called “right to carry” laws have swept across the nation.

Yet while violent crime has fallen overall since the early 1990s, mass shootings like Friday’s horror in Newtown, Connecticut have remained steady or even increased, depending on how you count them. Moreover, economist Steve Levitt found that the decline in violent crime could be attributed almost entirely to four factors: (1) increased numbers of police; (2) more violent criminals remaining in prison; (3) the decline of crack cocaine use after the 1980s; and, surprisingly, (4) women having greater access to legal abortion.

Dr. Levitt’s research specifically eliminated more media-friendly explanations such as the booming 1990s economy, our aging population, the Brady Bill and similar gun safety laws … and the easier concealed carry laws proposed by Dr. Lott.

He is not the only researcher to find holes in Dr. Lott’s claim. Rutgers University professor Ted Goertzel related several statistical flaws in Dr. Lott’s work, in an article for the Skeptical Inquirer, including a study that used Dr. Lott’s equations and reached a different result, simply by excluding data from Florida. A statistical analysis that works only if you use that exact data is the classic definition of “junk science.”

In 2003, Tim Lambert at ScienceBlogs examined Dr. Lott’s work and opened with this statement:

It looks as if Lott might have been caught cooking his “more guns, less crime” data.

Lambert cites an extensive discussion of coding errors in Dr. Lott’s research, errors Dr. Lott first denied and then simply refused to acknowledge as undermining his thesis. Lambert also cites the work of law professors Ian Ayres and John Donohue, who studied Dr. Lott’s data and found:

Indeed, the probabilistic underpinnings of statistical analysis suggest that running regressions for nine different crime categories to see if there is any measurable impact on crime will, by chance alone, frequently generate estimates that on their face are “statistically significant.” Therefore, it may well be the case that the scattered negative coefficients for various violent crime categories, which on their face suggest that crime decreases with passage of shall-issue laws, should be thought of as statistical artifacts. While we do not want to overstate the strength of the conclusions that can be drawn from the extremely variable results emerging from the statistical analysis, if anything, there is stronger evidence for the conclusion that these laws increase crime than there is for the conclusion that they decrease it.

Translated into plain English: Dr. Lott found a statistical fluke.

Science journalist Chris Mooney, then writing for Mother Jones, also reviewed the debate on Dr. Lott’s coding errors and conservatives’ refusal to admit them, concluding:

The right has good reason to stick by Lott: “The entire ideology of the modern gun movement has basically been built around this guy,” says Saul Cornell, an Ohio State University historian who has written widely on guns. Over the years the pro-gun intellectual agenda has had two prongs: Defending a revisionist legal understanding of the Second Amendment in constitutional law, and refuting social scientists and public-health researchers who argue that the widespread availability of guns in America plays a key role in the nation’s staggering number of homicides and suicides. Without Lott’s work, the latter argument becomes much harder to make.

This week on Fox News Sunday, Rep. Louie Gohmert (R-TX) cited Dr. Lott in claiming that the Newtown tragedy would have been stopped if teachers had been armed, saying:

The facts are every time guns have been allowed, concealed-carry has been allowed, the crime rate has gone down.

Washington Post FactChecker Glenn Kessler gave Rep. Gohmert three Pinocchios after reviewing Dr. Lott’s research and contrary findings by other scholars:

Gohmert’s statement was declarative and sweeping[.]

The actual evidence is much murkier – and in dispute. Certainly, it appears such laws have not increased the crime rate, as opponents had feared, but it is equally a stretch to say such laws are a slam-dunk reason for why crimes have decreased. Even those sympathetic to Lott’s research suggest that any decline in the crime rate from right-to-carry laws is more sporadic – as opposed to Gohmert’s claim that crime rate always goes down.

Simply, the evidence for Dr. Lott’s claim relies on using his exact data, and his exact equations, including coding errors he refuses to acknowledge. A clear and compelling scientific link between concealed carry laws and violent crime rates would not be contingent on such a singular analysis.

This is junk science … and deadly policy.