Defensive Gun Use: Gary Kleck Misfires Again

150209_kleck_guns_ap[1]

This is a response to Gary Kleck’s article in Politico.  This response is attached as an addendum there:

When Gary Kleck can’t defend, he attacks. Instead of offering new insight, Kleck instead baselessly speculates on our motives, suggesting we “hope that total gun prohibition will one day be politically achievable.” To be clear: prohibition is not something we have ever suggested in any of our writing, all of which can be found at armedwithreason.com. Not that it should matter, but neither of us are merely “investment counselors” either, as Kleck suggests. In fact, DeFilippis spent most of last year helping design and analyze surveys much larger than Kleck’s.

The Florida State professor even goes so far as to describe Dr. David Hemenway, director of Harvard’s Injury Control Research Center and author of more than 130 articles and five books in Economics and Public Health (a total that includes two decisive rebuttals to Kleck and several surveys), as “a man named David Hemenway… who is also untrained in survey methods.”

Rather than confront the significant, multidisciplinary research showing that the false-positive problem is ubiquitous when measuring rare events, Kleck pretends the problem is negligible, and links us to a 1998 “rebuttal” where he references surveys that have nothing to do with rare events. As Dr. Hemenway has extensively detailed, suggesting that false negatives could somehow outweigh false positives is indulging in fantasy.

Kleck also ignores the fact that his results repeatedly fail tests of external validity. In our original article, we mention that Kleck’s data would require, impossibly, that gun owners use their gun in self-defense in more than 100 percent of burglaries. Kleck’s data also suggests that every year hundreds of thousands of criminals are shot by law-abiding citizens. But where are the hospital records to validate this claim?  Kleck insists, with no medical knowledge and without citing a single study, that the vast majority of these criminals never seek hospital treatment, a claim scoffed at by medical professionals.

Kleck concludes his article by saying we “have not offered any new criticisms” and, like Dr. Hemenway before us, do “not once cite the one thing that could legitimately cast doubt on our estimates—better empirical evidence.” However, had he read the second page of our column, he would have seen that the entire point of our article was to highlight new empirical evidence debunking Kleck’s claims.

Here are the facts Kleck missed: According to his own survey more than 50 percent of respondents claim to have reported their defensive gun use to the police. This means we should find at least half of his 2.5 million annual Defensive Gun Uses (DGUs) in police reports alone. Instead, the most comprehensive nonpartisan effort to catalog police and media reports on DGUs by The Gun Violence Archive was barely able to find 1,600 in 2014.  Where are the remaining 99.94 percent of Kleck’s supposed DGUs hiding?

It would be disappointing to see any professor relegated to using falsehoods and ad hominem attacks in a desperate attempt to preserve the tattered remains of his thoroughly repudiated research. Yet, such tactics are particularly deplorable when they are used in service of a gun-worshipping culture that regularly generates tragedy on a massive scale.

Related posts

  • CNS

    Once again you try to tout the “The Gun Violence Archive” as having a perfect 100% recording rate of all defensive gun uses in the nation, when the organization itself has never made any such claim. All studies on this issue have shown defensive gun use to be in the tens of thousands at a minimum, including ones from the Bureau of Justice and the National Crime Victimization survey.

    It’s also amusing you accuse him of “ad hominem attacks”, but then go on to denounce him as being “in service of a gun-worshipping culture that regularly generates tragedy on a massive scale”. Once again the Pot calls the Kettle black.

    • >>Once again you try to tout the “The Gun Violence Archive” as having a perfect 100% recording rate of all defensive gun uses in the nation, when the organization itself has never made any such claim.

      We have communicated with the Gun Violence Archive. They make every effort to catalog every available defensive gun use that can be found in police and media reports. Given that survey respondents of Kleck’s VERY OWN survey said that they report their DGU to the police in 50% of cases, we should be able to find at least half of 2.5 million in police reports. Instead, we find 99.94% missing. Where are all these missing DGUs?

      >>All studies on this issue have shown defensive gun use to be in the tens of thousands at a minimum, including ones from the Bureau of Justice and the National Crime Victimization survey.

      You have made this claim before, we have responded to it, and you have refused to improve upon your position. I can only assume you’re not amenable to learning. Let me be very clear: the “studies” that have shown defensive gun use to be in the “tens of thousands” use the VERY SURVEYS we spend 2 pages methodologically critiquing. Our systematic indict of the use of surveys in this instance has simply not been responded to by anyone, let alone Kleck. The irreconcilable weaknesses with surveys attempting to estimate rare events is taught in Introductory Texts on Epidemology and Statistics, so it is baffling that this concept it difficult for some to get.

      Let me give you an analogy: We write an article about how we should stop using tarot cards to predict the 2016 election. You bring up the fact that a hundred tarot card readings all say that the same specific candidate will win. Cool. Interesting trivia. That does not respond to our overarching indict of the use of tarot cards as a predictive mechanism in the first place. In order to even begin to respond to our argument, you need to have a justification for the use of surveys to estimate rare events. Perhaps point to some surveys that have successfully estimated rare events in the past (we could find no such examples).

      >>It’s also amusing you accuse him of “ad hominem attacks”, but then go on to denounce him as being “in service of a gun-worshipping culture that regularly generates tragedy on a massive scale”.

      This is not an ad hominem…Kleck’s work has literally been used by the NRA to justify a hyper-aggressive pro-gun response to any national tragedy. The de facto solution proffered by the NRA in response to any form of violence is more guns, and the research of Kleck and Lott are largely responsible for this militarization.

      This is different, for example, from Kleck calling Dr. Hemenway “a man named David Hemenway… who is also untrained in survey methods.” and calling Devin and I “investment counselors.”

      • Gramm McDaniel

        “We have communicated with the Gun Violence Archive. They make every effort to catalog every available defensive gun use that can be found in police and media reports.”

        Did you ask them the conditions under which they acquire this information? Do they call just a random police department or do they read about a DGU on/in a website, newspaper, news program, etc. and call that respective PD?

      • BG Brix

        Perhaps “ad hominem” should/could have been replaced by “hyperbole”? “Gun-worshipping”? That is either ad hominem (name calling) or hyperbole (you really think there is an alter for idolators prostrating to a firearm?).

        And “generates tragedy on a massive scale” is certainly hyperbole. Per the 2013 FBI Uniform Crime Report, Table 20, 0.00266% (12,253) of the U.S. population died of homicide. That is statistically insignificant, but to be empathetic, these homicides remain tragedies to those directly affected, but not the public at large, until the media exploits the suffering of those affected and publicizes the perpetrator (any press is good press in the perpetrator’s perception, so why give them any?).

        We live in the safest period of all human history, and globally, life continues to become safer, so why do some strive to find ways to further restrict law abiding citizens vice address underlying issues of poverty, lack of education, mental health problems, etc.? Those with nothing to lose (albeit in their own perception) will not conform to social norms nor legal repercussions, so let’s assist them in becoming or remaining viable members of society rather than choose a mass punishment or restriction on those who are already productive and law abiding members of same society.

  • They’re missing because they’re fantasy reports claimed by individuals looking to use hysteria to finance the gun industry.

  • Gramm McDaniel

    “author of more than 130 articles and five books in Economics and Public Health.”

    I can write 130 articles and five books in Economics and Public Health too. If I do that, can I conduct biased studies funded by the NRA without receiving critique too?

  • Gramm McDaniel

    BTW, why did you not post this up on Politico? Gary will probably be more likely to see your article if you post it there.

  • Kyle

    What does being a public health expert have to do with writing about the subject of guns in society? Guns are not a public health issue no matter how many times the gun control-oriented try to claim it. No more than rape, murder, robbery, assault, etc…are public health issues either. Public health deals with things like sanitation, disease, air and water quality, etc…it has nothing to do with criminal behaviors by people. People deciding to maim or kill other people, whether with guns or not, has nothing to do with any aspect of public health. Those are the purview of the field of criminology.

  • Kyle

    “Yet, such tactics are particularly deplorable when they are used in service of a gun-worshipping culture that regularly generates tragedy on a massive scale.”

    We do not have a “gun-worshipping” culture, we have a culture in which the possession of arms is seen by many as a fundamental, natural human right to be protected. I believe in free speech rights very much too, but I have little interest in becoming a journalist. And I believe very much in LGBTQ rights, but I am not an LGBTQ person.

    A very large amount of the gun violence in the country happens in the inner cities due to gangs.

  • TVB

    Gary Kleck one of those Junk scientists that shill for the gun industry. Remember cigarettes also were not bad for lyour health a few years ago according to many scientists

  • Quiet Professional

    “Kleck’s data also suggests that every year hundreds of thousands of criminals are shot by law-abiding citizens. But where are the hospital records to validate this claim?”

    Your attempted definitional sleight of hand with respect to what constitutes a defensive gun use is — pardon the pun — as ham-handed as it is unscientific.

    A gun can be used defensively without ever being fired. That is why tens of thousands of DGUs annually — on the low side — is an entirely plausible estimate.

    The fact that armed citizens aren’t shooting and killing bad guys by the tens of thousands is a testament to their remarkable restraint, notwithstanding your efforts to slur ALL law abiding gun owners by tagging every criminal use or misuse of a firearm as having been committed by a “good guy with a gun” or a “2A supporter”.

    Of course, if armed citizens were shooting and killing bad guys by the tens of thousands, you’d be trumpeting that fact as a further argument against gun ownership in general. “Trigger happy vigilantes.” “Yosemite Sams.” “Rambos.” “Wanna be Dirty Harrys.” Etc.

  • Quiet Professional

    Moreover, there are now 12 million — and growing — gun carry permit holders nationwide. If only a scant 1 of every 1,000 of those armed citizens annually had occasion to use his or her gun defensively, that’s 12,000 DGUs annually among that particular subset of the general population alone. Not at all implausible — especially given the odds (much greater than 1 out of 1000) of any one of those armed citizens encountering a criminal threat.

  • Pingback: More Holes in the Defensive Gun Use Myth: New Study Finds DGU is Ineffective and Rare | Armed With Reason()

  • Pingback: UCC Student who carried on campus explains why he didn't get involved - Page 2()

  • Pingback: Debunking “More Guns, Less Crime” 3.0 | Armed With Reason()

  • Pingback: A Thanksgiving-Edition Cliff Notes Guide to the Gun Debate | Armed With Reason()

  • Pingback: Guess What? With Your Help There’s A Chance That CDC-Funded Gun Research Might See The Light Of Day. | mikethegunguy()

  • Pingback: Guess What? With Your Help There’s A Chance That CDC-Funded Gun Research Might See The Light Of Day. | Mister Journalism:()

  • Pingback: Marco Rubio’s claim that no recent mass shootings would have been prevented by gun laws – Lexington Discusses Guns()

  • Pingback: Ted Cruz Is Wrong: Owning Guns Doesn't Make Women Safer - Alex Poucher()

  • Pingback: BUSTED: Bloomberg’s “The Trace” Intentionally Misrepresents Defensive Gun Uses – CrossFire Defense()

  • Pingback: BUSTED: Bloomberg’s “The Trace” Intentionally Misrepresents Defensive Gun Uses | GOP News()

  • Clark Coleman

    “Kleck’s data also suggests that every year hundreds of thousands of criminals are shot by law-abiding citizens. But where are the hospital records to validate this claim? Kleck insists, with no medical knowledge and without citing a single study, that the vast majority of these criminals never seek hospital treatment, a claim scoffed at by medical professionals.”

    Actually, Kleck’s seminal 1995 paper claimed that survey respondents were likely to “report with favor” their own marksmanship. e.g. someone fires at a criminal who runs away, and the citizen reports to the survey taker that he wounded the criminal. Kleck debunks this aspect of the survey results in numerous ways (e.g. the private citizens would be much better marksmen than both police and criminals, based on wounding data available elsewhere). He only mentions the possibility of criminals not seeking treatment as a lesser factor, and it would certainly be a likely factor in the case of minor wounds. For major wounds, the criminal would be forced to seek treatment.

    Kleck also cautions against scaling up of these numbers for other reasons. The number of claims of wounding is only 17 in the 1995 paper. Kleck explains in great detail why he thinks the sample of more than 200 defensive gun users is accurate, but he would never try to scale up the 17 reports of wounding criminals as you have done.

23 Comments