A Harvard School of Public Health Press Release describes a new study by Miller, Hemenway and Azreal:
In the first nationally representative study to examine the relationship between survey measures of household firearm ownership and state level rates of homicide, researchers at the Harvard Injury Control Research Center found that homicide rates among children, and among women and men of all ages, are higher in states where more households have guns. The study appears in the February 2007 issue of Social Science and Medicine
Earlier studies on the relationship between firearm ownership and crime had to use proxy measures for ownership because there was no national survey data available at the state level. The press release continues:
Analyses that controlled for several measures of resource deprivation, urbanization, aggravated assault, robbery, unemployment, and alcohol consumption found that states with higher rates of household firearm ownership had significantly higher homicide victimization rates for children, and for women and men. In these analyses, states within the highest quartile of firearm prevalence had firearm homicide rates 114% higher than states within the lowest quartile of firearm prevalence. Overall homicide rates were 60% higher. The association between firearm prevalence and homicide was driven by gun-related homicide rates; non-gun-related homicide rates were not significantly associated with rates of firearm ownership.
Naturally pro-gun bloggers have attacked the study. Glenn Reynolds says they were funded by the Joyce Foundation:
I’m pretty sure that these guys would call anyone who accepted grants from the NRA bought-and-paid-for. But the Joyce Foundation is every bit as biased as the NRA, and has a history of paying for scholarship that would be treated as a scandal if it were engaged in by pro-gun folks.
His “paying for scholarship” link goes to a post where he says that it isn’t unethical, so I guess he isn’t saying that it is scandal just that “these guys” would. Or something. Certainly Reynolds called it a “vicious smear” when people attacked John Lott because he was funded by the Olin Foundation.
Reynolds also links to criticism by Jeff Soyer. Unfortunately, Soyer has not read the study, so his criticism is very wide of the mark. Soyer writes:
In the current study, they claim they’ve “controlled” for factors such as unemployment, etc. I’d be interested in seeing how they accomplished that statistical dance.
No problem. They used negative binomial regression.
Once again though, this study uses the same flaw of lumping a bunch of states together, dividing all the states into just four “groups”. If you take the states with the highest gun ownership, and here’s a handy WaPo chart, you find that they must have lumped North Dakota, Wyoming, and Alaska (and Montana, Idaho, etc.) which have high suicide rates in with Alabama and Mississippi, Arkansas and Louisiana, states with high homicide rates. In my opinion, that’s the only way for him to reach his statistical goals.
They only grouped states into four groups in an additional analysis that they did as a check on their results. Their results do not depend on such a grouping. The study looks at homicide, not suicide. The word “suicide” does not even appear in the body of the paper.
But here’s the clincher. From that same chart, the “state” with the highest homicide rate in the U.S.? By a factor of almost four times the rate of ANY other state? It’s Washington DC. It’s the place where gun ownership is for all practical purposes forbidden, that is, where gun control has become a total ban on guns.
Washington DC is not a state. It’s part of a city. It is simply wrong to compare it to entire states.
As long as Miller continues to lump suicide and homicide together, his arguments are specious at best and full of crap at worst.
Again, Miller et al did not lump suicide and homicide together. This study does not consider suicide at all.
Back to Reynolds:
I find much of the public health literature on guns to be highly biased and deeply untrustworthy. It starts with an agenda, rather obviously, and then constructs “research” to confirm it. In this it resembles far too much of the politicized social science we see today, which explains in part why people are far less persuaded by social science claims than they used to be.
Unfortunately the reference Reynolds gives to support his claim about the public health literature on guns is itself highly biased and deeply untrustworthy.