Saturday, July 30, 2005

Guns and Trigger Locks: What's the evidence?

This week the senate passed legislation to limit the liability of gun manufacturers for criminal acts committed with their products. An amendment mandated trigger locks for handguns. The medical literature suggests that it's very difficult to convince people to use trigger locks or otherwise store their guns more safely, even with economic incentives. Thus mandated provision of trigger locks with guns may be the only way to ensure their widespread use.

The problem is that there is no evidence whether trigger locks are effective or not. Nada. Zip. But if you hesitate to recommend trigger locks on the basis of this lack of evidence, we should remember the context that gun ownership itself has no evidence to suggest it confers protection from crime. The New England Journal published a controversial and much cited case-control study by Kellermann suggesting that gun ownership increases the risk of homocide in a household 2.7 times, after adjusting for various possible confounding factors. This study has had significant substantive criticisms, the most salient of which I have quoted below with refutations:

1. "99.8 percent of the protective uses of guns do not involve homicides," says Paul Blackman of the NRA. Defensive gun uses include waving the weapon, firing warning shots, wounding the intruder, etc.

It is simply untrue that researchers cannot measure the nonfatal protective benefits of firearms, or that Kellermann's survey failed to detect such a benefit. If firearms deter, scare away or wound intruders, then the murder victimization rate of gun owners should be lower than non-gun owners. The absence of a gun in the home would have been recognized as a murder risk, rather than the presence of a gun.

Kellermann's case-control method was ideally suited to detect such benefits, if they existed. Of course, Kellermann's survey found quite the opposite -- a risk 2.7 times greater.

2. Guns do not emit magic rays that control people's minds, or magnetize murderers to the doorstep.

This strawman argument is based on a false stereotype. Over 76 percent of the homicides were committed by a relative or acquaintance of the victim, and only 3.6 percent were verified as strangers breaking in. Furthermore, arguments and romantic triangles comprised half the homicides. But the most important point here is that a gun in the home only raised the risk of gun homicide -- not homicide by any other means. The most straightforward explanation is that greater gun availability transformed a normal family fight into something much more deadly.

3. People threatened by violence bought guns to defend themselves, hence the correlation between gun ownership and murder.

This is possible, but the number would only be very small, for the following reasons. The study already controlled for domestic violence, so the only way this could happen is if the murderer threatened the life of the victim before things escalated into violence. The victim would then have to buy a gun, which would fail to protect.

Several things make this unlikely. First, we would expect a history of violence to precede any threats or attempts on a person's life, which is, after all, the ultimate form of violence. Second, the study showed that gun ownership resulted in an increased risk in gun homicides only, not any other type of homicides. Why would the murderer restrict himself to a gun, and then only if the victim had a gun? Third, this makes a poor case for gun deterrence, since the correlation is only possible when the gun fails to protect. Again, the researchers found no protective benefits of gun ownership.

6. "These people were highly susceptible to homicide," says Paul Blackman of the NRA. "We know that because they were killed."

If there is an Illogic Hall of Shame, this remark deserves to be emblazoned above its front entrance. By this reasoning, we should not put seat belts in cars, because people killed in car crashes were susceptible to those accidents anyway.

What Blackman is doing here is evoking a general risk for murder, while ignoring its specific risk multipliers. You may, in general, have an antagonistic person in your life given to flashes of murderous temper. But there are specific factors that may increase the risk of murder. Does he drink? Use drugs? Commit crime? Own a gun? Increasing any of these behaviors increases the risk. But it makes no sense to increase the risk multiplier, let someone get murdered, and then argue that the multiplier was not at fault, since the victim was obviously susceptible to murder anyway.

This argument also ignores one of the study's findings, that a gun in the home increased the risk of gun homicide only, and not any other method of homicide.

9. The researchers did not include in their analysis those cases where the home-owner shot a non-resident intruder.

These cases were rare, but even so, this objection is irrelevant. The protective benefits of a gun would have still shown up in the different victimization rates of gun-owning and gun-less households. (See point 1.)

10. This study was conducted by medical doctors who were out of their league; this is an issue best left for criminologists.

Epidemiologists are highly experienced at using the case-control method to determine risk factors. This is how cigarette smoking was linked to lung cancer, for example. The statistical method is the same no matter what the risk factor, be it cigarettes, a virus, a missing vitamin or a gun. A good analogy is that of an astronomer using optics technology to make a breakthrough in optometry.

Hat tip to Steve Kangas

In the setting of a widespread intervention (i.e. guns) with such significant adverse effects, I would argue that the evidence burden for trigger locks doesn't need to be that high.


Post a Comment

<< Home