After the San Bernardino shooting and the intense debate over guns that has followed, some conservative commentators are pushing back on the evidence that more guns mean more gun deaths, citing their own data and charts. But they're making a grave statistical mistake.
Over at National Review, for example, Robert VerBruggen wrote that more guns really don't lead to more homicides. He pointed to his own chart, which simply plots gun ownership against homicides at the state level:
"There is actually no simple correlation between states' homicide rates and their gun-ownership rates or gun laws," VerBruggen wrote. "This has been shown numerous times, by different people, using different data sets."
But VerBruggen's chart and the analyses he cites make no effort to control for factors besides guns — poverty, urbanization, other crime rates, and so on.
Usefully, quick-and-dirty scatter plots like VerBruggen's aren't actually necessary here. This issue has been studied carefully, at length, and with substantially more statistical firepower than he brings to bear. And the results are clear.
"Within the United States, a wide array of empirical evidence indicates that more guns in a community leads to more homicide," David Hemenway, director of the Harvard Injury Control Research Center, wrote in Private Guns, Public Health.
When you control for other factors, more guns really do mean more gun homicides
It's a basic rule of any empirical research: If you want to evaluate how much a single factor impacts something else, you should do your very best to control for all other variables to ensure that the single factor is the only thing being analyzed. So with studies on gun ownership and gun violence, researchers go through great efforts to control for all sorts of variables — economic outcomes, alcohol consumption, rates of urbanization, other crime rates, and so on — to make sure the results look, as much as they possibly can, only at gun ownership and its effects.
This is why, for example, Vox's charts look at the correlation between gun ownership and gun violence in developed countries: It helps weed out the many, many social and economic factors involved if you compare the US with, for instance, Honduras — a nation mired by poverty and weak government institutions.
So what happens when researchers do this? Here's one chart, from a 2007 study by Harvard School of Public Health researchers, showing the correlation between statewide firearm homicide victimization rates and household gun ownership after controlling for robbery rates:
The study went on to look at other variables, including urbanization, other types of crime, and poverty. Time and time again, researchers found a strong association between firearm prevalence and homicides after controlling for these factors. And the increase in overall homicides was driven by an increase in gun-related homicides — homicides that didn't involve guns didn't significantly increase as gun ownership did. In other words, more guns meant more homicides, particularly gun homicides.
A more recent study from 2013, led by a Boston University School of Public Health researcher, reached similar conclusions: After controlling for multiple variables, the study found that each percentage point increase in gun ownership correlated with a roughly 0.9 percent rise in the firearm homicide rate.
This holds up around the world. As Vox's Zack Beauchamp explained, a breakthrough analysis in the 1990s by UC Berkeley's Franklin Zimring and Gordon Hawkins found that the US does not, contrary to the old conventional wisdom, have more crime in general than other Western industrial nations. Instead, the US appears to have more lethal violence — and that's driven in large part by the prevalence of guns.
"A series of specific comparisons of the death rates from property crime and assault in New York City and London show how enormous differences in death risk can be explained even while general patterns are similar," Zimring and Hawkins wrote. "A preference for crimes of personal force and the willingness and ability to use guns in robbery make similar levels of property crime 54 times as deadly in New York City as in London."
This is, in many ways, intuitive: The prevalence of guns can cause petty arguments and conflicts to escalate into deadly encounters. People of every country get into arguments and fights with friends, family, and peers. But in the US, it's much more likely that someone will get angry at an argument, pull out a gun, and kill someone.
These three studies aren't the only ones to reach similar conclusions. Multiple reviews of the research, including the Harvard Injury Control Research Center's aggregation of the evidence, have consistently found a correlation between gun ownership and gun deaths — including homicides, suicides, and accidental shootings — after controlling for other factors.
The National Review's VerBruggen neglects this research, instead drawing conclusions from a much more simplified analysis.
But all VerBruggen's chart shows is that multiple issues contribute to gun violence — demographics and poverty, for instance. But no one is arguing that guns are the only factor contributing to violence. They are, instead, just one of many contributors — and a very common denominator in much of the violence we see in the US.
Going after guns won't solve every problem, but it would help
With that said, it's probably true that this aspect of the gun control debate is not emphasized enough: Guns are a factor, not the only factor. The fact that you need to control for all these other variables is indicative of this — if guns were the single dominant issue when it came to gun violence, major statistical controls wouldn't be necessary.
This is why so many lawmakers, including President Barack Obama in recent remarks, have gone out of their way to say that gun control would make gun violence more rare, but not totally end it. And in fact, no single policy could stop all gun crime — there will always be a black market for guns, and humans have been killing each other since they began to exist. (More comprehensive solutions would, for instance, seriously consider solutions to poverty, excessive alcohol consumption, and the other variables that gun studies control for because they're known contributors to crime.)
But the tremendous number of guns in circulation is one of the issues that makes America — and some states in particular — unique in the world, and explains why the US seems to be so far ahead of its peers when it comes to lethal violence.
So it might not stop every incident like the San Bernardino shooting, but the research is fairly unequivocal in demonstrating that reducing access to guns — and reducing the number of guns — would reduce gun violence in America. And it might even bring the US closer to its developed peers.