A Gun Control Head-Scratcher

Beware when editorialists deliver bromides such as “Facts are stubborn things.” Case in point is today’s brief Times-Dispatch editorial note using the homicide rate in Washington, DC to point to the folly of gun control. The fact cited by the RTD is that even though the nation’s capital introduced fairly tough gun control laws in 1975, the homicide rate subsequently increased in DC faster than the national average.

The editorial writers probably didn’t come to this observation independently–citing the Washington, DC case is a favorite trick of conservative editorialists. (Do a quick search engine search on Washington, homicide, and gun control and see what comes up.)

But let’s look at the argument on its merits. What we have here is a bivariate correlation, drawn from exactly two data points, the Washington, DC of 1975 and the Washington, DC of today. There are two problems with drawing extensive policy conclusions from this little evidence. First, correlation does not equal causation. There are many possible explanations for the increase in homicides in DC since 1975: the crack epidemic beginning in the 1980s; thirty years of concentrated poverty, social neglect, and dysfunctional schools; changes in gang activity; possibly even changes in drug shipment routes. (Needless to say, none of these alternative explanations for DC’s homicide rate are explored by the editorial.) It might in fact be the case that without the gun control laws, homicides would have risen even faster in DC than they actually did; or it might be the case that gun control laws didn’t really make much difference either way; or as the RTD wants to suggest, it might be the case that gun control laws actually contributed to homicides.

The RTD’s interpretation seems to rely on a mental image in which we think of most homicides as involving criminals preying on obviously defenseless people–as if most murders in DC consisted of professional criminals causing mayhem in the white neighborhoods of upper Northwest. If in fact most homicides are drug or gang-related, then this interpretation is far less plausible: gang members are probably not thinking to themselves “this guy who’s invading my turf couldn’t possibly have a gun, because that would be violating the local gun laws, so I’m going to shoot him and know he won’t shoot back.”

But the bigger point is, we have no way of knowing which of these three interpretations is correct in the absence of much more detailed information.

This brings us to our second problem: two data points is not enough to draw any kind of firm conclusion about the relationship between two different phenomena, and there’s no excuse for just relying on two data points when there are hundreds of other data points we might consider (such as the experience of other U.S. cities).

In fact, quantiative criminologists have considered the relationship between gun control and crime, as well as the relationship between gun ownership and crime, in some detail, drawing on many more than two data points. The basic story seems to be this:

1. According to a 1993 study of the 170 largest U.S. cities, local gun control laws generally don’t have a strong effect on local crime, though they might affect certain types of crime. It would be interesting to update the story through the 1990s, when crime rates fell, but I’ll accept this provisional finding.

2. On the other hand, local gun ownership rates are, according to a recent study by scholars from Duke and Georgetown substantially correlated with increased homicides, controlling for a number of other factors (including the overall local crime rate). This paper uses proportion of suicides by gun as its proxy for rates of gun ownership at the county and state level; an earlier study, much hyped by the anti-gun control crowd, which relied on voter exit poll data to estimate local rates of gun ownership produced opposite results. The exit poll method is flawed, because, as the Duke-Georgetown study notes, “voters are by no means a representative sample of the adult public.”

The paper’s authors (Philip Cook and Jens Ludwig) go on to argue that gun ownership rates are causally related to the homicide rate. One obvious reason why is that fights carried out with guns are far more lethal than fights carried out with knives and fists; another is that as the quantity of guns in circulation increases, it becomes easier for criminals to acquire one. Notably, they also find that their measure of gun ownership is not strongly correlated with other types of crimes, or even with homicides committed without guns. That finding actually increases confidence that is the presence of guns, not some unmeasured factor, that is decisive in the relationship between gun ownership and the homicide rate.

3. In the international context, national rates of gun ownership are also correlated with national homicide rates. The most plausible interpretation of this evidence is that there is a causal relation between the fact that the U.S. has much higher homicide rates than Europe and Japan and the fact that we have much higher rates of gun ownership, for the same sorts of reasons noted by Cook and Ludwig.

There might be a slight puzzle in understanding how finding 1 could be consistent with findings 2 and 3. If more guns are associated with more gun-related crime, why don’t local gun control laws seems to have a strong effect on reducing crime (with possibly some exceptions)?

The answer is suggested in the RTD’s editorial itself: that in places like DC, gangs and criminals can easily acquire guns from outside the immediate locality. That explanation, in turn, implies that are needed are truly national efforts to restrict the supply of guns to criminals, since isolated local efforts are likely to be ineffective.

The head-scratcher in all this is the RTD editorialist’s attempt to dissuade readers from that interpretation. “If that were the case,” the editorial states, “the national homicide rate would not be 22 times lower than Washington’s.”

I’ve been struggling all morning to try to make sense of what that statement is supposed to mean. Upon reflection, I think the claim is supposed to be this: that in the rest of America where guns are plentiful and people can defend themselves, homicide rates are low, but in DC, where no one can defend themselves, homicide rates are very high.

Note however that this claim is perfectly consistent with, rather than a contradiction, of the view that the reason the DC gun control laws don’t work is because guns come in from elsewhere. So this statement doesn’t prove what it is intended to prove.

Nor does the claim stand on its terms: as noted above, the best academic evidence suggests that as a general phenomenon widespread gun ownership encourages rather than restricts gun-related violence.

There are hardly any local issues more important than crime, and the RTD could have made a healthy contribution to public understanding of the issue by arguing that local gun control laws in themselves are unlikely to be a panacea to Richmond’s crime problem. They might even have gone on to suggest that the city needs to tackle the social roots of crime rather than rely simply on gun laws, or alerted readers to the efforts now going on in the city to do just that, such as the Richmond Youth Peace Project.

But instead of going for the scentifically defensible, constructive point, the RTD went for the ideologically sweeping claim resting on just two data points as well as some highly questionable assumptions, a claim that can’t be sustained once we look at a broader range of evidence.

As for the utility of gun control itself, there’s no question than figuring out how to reduce gun-related violence in a society that already has 250+ million guns in circulation is a massive challenge. For a sobering, even pessimistic look at the problem, check out James Jacobs’ 2002 book, Can Gun Control Work? Jacobs thinks that we can and should take some regulatory steps to make it more difficult for criminals to get guns, but that effective policing strategies and targeted social welfare policies and interventions will be far more important in reducing violent crime in the long run. That seems like a plausible conclusion, though some critics, such as this thoughtful reviewer, contend that Jacobs is too pessimistic in assessing gun control’s value as an anti-violence tool.

Published in: on August 25, 2006 at 7:10 pm  Comments (1)  

The URI to TrackBack this entry is: https://richmondtalksback.wordpress.com/2006/08/25/a-gun-control-head-scratcher/trackback/

RSS feed for comments on this post.

One CommentLeave a comment

  1. […] On the issue of gun control, the RTD, noting that New York City has reduced its crime rates while maintaining tough gun laws, has backtracked remarkably from its August editorial. The RTD now says that one can cite evidence for or against gun control’s effectiveness. This is hardly a sophisticated social science position, but it represents progress. (See our original post on gun control here.) […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: