February 2002


In “Point Blank”, Kleck analyzed NCVS data and found that while 38% of people who used any means of self-protection against robbers were injured in the encounter, only 17% (the lowest for any means for self-protection) of people who use a gun for self-defence against robbers were injured. Kleck claimed that this showed that guns were the most effective means for avoiding injury.

In their critique of “Point Blank” Alba and Messner point out the flaw in Kleck’s reasoning — the evidence from the NCVS is equally well explained if injury makes victims less likely to use guns. Kleck dismisses this alternative explanation as a “speculation”. This is a rather odd argument. The evidence shows a correlation between lack of injury and with-gun defence. Kleck’s explanation (gun defence prevents injury) and the alternative (injury prevents gun defence) are equally speculative.

Later versions of the NCVS provide a way to resolve the issue. They ask questions to find out whether the injury occured before or after the self-protective actions. Differences in injury rates before the victim took any self protection action will reflect the effect injury has on which methods of self-protection will be chosen, while differences in injury rates after the self-protection action reflect the effect self-protection methods have on injury.

In “Armed”, Kleck analyzes the new NCVS data. Over all means of self-protection against robbers, 34% were injured before they did anything and 7% after. For gun defenders, 13% were injured before they used the gun, and 8% after. The results seem quite clear: gun defence does nothing to reduce your chances of injury, but injury prevents victims from using guns for defence.

Remarkably, in his discussion of these results, while Kleck concedes that defensive gun use “does not appear to be as uniquely safe as data from earlier NCVS data suggested”, he does not take notice of the implications of the large difference in the pre-self-protection injury rates. Instead he argues that the data shows that gun defenders face less favourable circumstances because they were more likely to face multiple and gun-armed attackers, ignoring the injury data suggesting that they face more favourable circumstances.

Lowell Savage writes:

Sorry, Ron. Much as I agree with your position, I have to say that you haven’t addressed Tim’s issue: why is it that 37% of non-gun defenders were injured before they began self-defensive actions while only 13% of gun defenders were injured before they began self-defense actions?

Perhaps an anecdote from a column by Ann Coulter could illuminate a possible explanation. Ann said that she was walking alone over a bridge toward her apartment (or is it a condo? And no I don’t remember why she was doing this alone, at that time.) when she saw a man coming toward her from the area of the apartment buildings. His demeanor told her that he was up to no good and he was making no effort to hide his intentions. (In fact, I seem to recall from her column that she articulated some specific things about why she believed that he intended at least robbery.) Her column then went into how disgusted she was that the man didn’t even “have the decency” to attempt to hide his intentions because they were in DC and so he knew that she was unarmed and therefore at his mercy. She was too far from anyone to scream for help. She was too far from any place that she could run to (before he could catch her). And she was certainly no match for him in hand-to-hand combat (even assuming that he wasn’t carrying some sort of weapon). So, she faced no choice but to give him whatever he wanted. Now, if he then began hurting her anyway, she would have probably decided that fighting or running was her only option. At that point, she is in Tim’s 37% group.

Not so. The 37% group were injured before they took any action. If an incident happened as outlined above, it would count as an injury after taking action. (One of the actions listed is “cooperated with the offender”.)

As for the post-self-protection injury rates, I think the biggest missing piece of information is the deaths (I think some of the earlier reponses hinted at this problem). Here’s the question: How many of the gun-using people who were injured after beginning self-protection were people who would be dead if they hadn’t had a gun?

We don’t know. Not do we know how many end up dead because they did use a gun. Both numbers are probably small compared with the number injured.

So Tim, unless you can work up some numbers to eliminate this possibility (that a large segment of the post-self-protection-with-gun injuries represent people who would otherwise be dead), then I’m afraid that the best you can do is to try to make the case that the NCVS data isn’t complete enough for any assessments in this area.

I doubt that it is possible to collect enough to be certain about these things. Nonetheless, the best available data suggests that gun self-defence does not reduce the risk of injury.

I also note that pro-gun folks had no trouble with using less complete NCVS data to argue that guns were the safest means of self-protection. For example:

“According to U.S. Bureau of Justice Statistics data, having a gun and being able to use it in a defensive situation is the most effective means of avoiding injury (moreso even than offering no resistance) and thwarting completion of a robbery or assault.” talk.politics.guns Official Pro-Gun FAQ 1/2

and:

“The Department of Justice’s National Crime Victimization Survey reports that the probability of serious injury from an attack is 2.5 times greater for women offering no resistance than for women resisting with a gun. Men also benefit from using a gun, but the benefits are smaller: Offering no resistance is 1.4 times more likely to result in serious injury than resisting with a gun. Resistance with a gun is the safest course of action for victims to take.” John Lott “Gun Control Advocates Purvey Deadly Myths”, Wall Street Journal 11/11/98

and:

“Florida State University criminologist Gary Kleck analyzed data from the Department of Justice (1979-1985 National Crime Survey public use computer tapes). He found victims that defended themselves with a gun against a robbery or an assault, had the least chance of being injured, or of having the crime completed. Doing nothing, trying to escape, reasoning with the offender, or physical resistance (other than with a gun), all had higher probabilities of injury and crime completion.” Source