Our favorite Differently-Winged MA expat may not like this: Turns out states with more guns at home have, like, more homicides. Who woulda thunk it?
Guns are used to kill two out of every three homicide victims in the United States, and new research shows that easy-access guns in the home make a difference. Homicide rates are highest in states where more households have guns, the national survey concludes.
… “The southern states and the mountain states on average have higher gun ownership levels than most states in the northeast,” [researcher] Miller said.
For instance, in Oregon, 40 percent of households have guns, and more than half of the homes in Alabama own guns. Compare that with Massachusetts, where some 12 to 13 percent of households reported firearm ownership. [my emphasis — Charley]
Gun-related deaths
In the top firearm-household states, homicide rates were more than double the rates found for states in the lowest firearm group. Overall, the top-gun states showed homicide rates that were 60 percent higher than all other states.
Most women victims of homicide are killed by guns that were already in the home, while men tend to be killed outside of the home. From past studies, Miller said, women are more likely to be killed by people they know, such as ex-boyfriends or ex-spouses.
“Overall women are more likely to die in states where there are more guns [in homes],” Miller told LiveScience.
I think that’s what David Bernstein has been trying to tell us.
Yeah Bruce, guns might be great for self-protection … but it’s much more likely the brutalizing husband or boyfriend who’s quicker to reach for the gun. And I’m proud of Massachusetts for mostly doing without.
is it also true that neighboring states’ gun laws also play a role? If so, how do we get NH to settle them guns down a bit?
<
p>
Nationally, gun legislation is a losing issue for Dems. It may be popular in the Northeast, but its political suicide elsewhere. It seems to me that the Brady Bill work really helped fuel the 1994 Democratic Congressional losses.
American life. I think it is largely due to culture, reinforced by violent entertainment which, intentionally or not, pushes the theme that the world is violent and to succeed one needs to be violent as well…a notion that is demonstrably false in civilized societies, which most of the United States is not quite.
<
p>
When I moved from Cambridge to St Petersburg, I remember looking up some comparative statistics…based on a scale where the national crime rate is 100, Cambridge was 60 or so, Saint Petersburg was over 200. And Saint Petersburg is not a high-crime city (though it has a couple of neighborhoods that are). But, that’s what you get when any fool can own a gun.
that pesky 2nd Amendment that makes people think that they, well, have the right to keep and bear arms.
…deny the federal government the power to deny (sorry for the double negative) individuals the right to keep and bear arms but…
<
p>
(i) it does not deny the federal government the power to deny individuals the right to own guns
<
p>
(ii) it is ambiguous as to whether the right is an individual one or a collective one (more than a few 2d amendment “absolutists” would prefer to read the introductory clause of the 2d amendment–the one relating to a well regulated militia–out of the amendment, altogether, but, sorry, it is there, and
<
p>
(iii) the fact is that to date the amendment only applies to the federal government; the US Supreme Court has steadfastly refused to incorporate it through the 14th amendment to apply to the states.
<
p>
Next question.
I think if Democratic Senators like Robert Byrd and Pat Moynihan had pushed forward with Healthcare Reform, that would have changed the dynamics of the 1994 election. As it was, Byrd would not let it be bundled in with the budget, to avoid a filibuster, and Moynihan kept pushing for Welfare Reform first, and so the Dems pretty much had two legislative accomplishments to trumpet: the Brady Bill and the Family Medical Leave Act. While good, those were not enough to create the sense that the government was working FOR the ordinary people.
Whoa, hang on there, Charlie. You neatly neglected to point out the article goes on to say that the “analysis controlled for factors that could affect homicide rates, including socioeconomic status, urbanization, non-lethal crimes, unemployment, and alcohol consumption.”
<
p>
The operative word is “controlled.” In other words, hey, let’s filter out the influence of police under-staffing, poverty, joblessness and boozing it up. What, no normalizing for crack cocaine trafficking or density of gang members per square mile?
<
p>
This is the equivalent of normalizing inner-city school test scores with the results from Dover Sherborn Regional High, if we only “control” for race, socioeconomic status, and household income.
<
p>
In that light, Boston wouldn’t have one of the higher gun homicide rates among US cities.
<
p>
Without actually seeing the study (available by paid subscription only via Social Science and Medicine) I’d be cautious about making any cause-and-effect pronouncements.
<
p>
Even the authors of the study warn about that in their abstract … “Although causal inference is not warranted on the basis of the present study alone, our findings suggest that the household may be an important source of firearms used to kill men, women and children in the United States.”
<
p>
Well, duh. It’s like saying if we control for everything else, the availability of gin is correlated with drunks. Actually it’s even worse than that … it’s more like sunspots predicting whether the American or National League wins the World Series: highly correlated, but not causal.
<
p>
This study not only proves nothing, it says nothing.
You think the availability of gin has nothing to do with drunkeness? You think the availability of guns and the likelihood of their use in a fight is comparable to “sunspots predicting whether the American or National league wins the world series”?? I guess smoking doesn’t cause lung cancer, either, since that is also just “highly correlated.”
To prove that gun control is good social policy (and I believe it is), one really does have to prove which direction the causal arrow points. Yes, there is a high correlation between smoking and lung cancer but no one seriously believes, for example, that a gene predisposing one to lung cancer makes one eager for the joys of Camels. We know where that causal arrow points. With gun control, there are a lot of variables to control for before correlation suggests causality.
<
p>
I take your point as being that the availability of guns has to increase the chance of their being used in crime. On one level, that is obviously true. But is the effect tiny or significant? If it is tiny, the social costs of regulation don’t argue for it.
<
p>
We use science, after all, to tell us what is true but counter-intuitive. If no truths were counter-intuitive, we could rely on our intuition. We wouldn’t need science.
between causality and correlation.
<
p>
Having a bottle of gin in many homes does not explain drunkenness. Alcoholism has many roots. Gin explains nothing.
<
p>
Even the researchers in the Harvard study warn readers not to infer any causality between guns-in-homes and homicide rates.
<
p>
Why are you twisting correlation into causality?
The study says there’s more homicide in the more gun states, mainly in Southern states and mountain states. Maybe the causation is warm weather or thin air.
with N=1.
from Bruce, who points to this post, which sorts through the data used in the study (scroll down to the “update”) and claims that it calls Miller’s conclusions into question.
<
p>
I haven’t got the time to wade through it. Anyone?
Of the 10 safest states (those with the lowest homicide rates), five were graded D’s and F’s for gun control by the Brady Campaign. This incudes Maine, statistically the safest state, which received a D-.
Only three of the top ten, Massachusetts, Rhode Island & Hawaii, received B or better.
So it isn’t all about more gun control. There are other, more significant factors at work here.
<
p>
US homicide rates in the year 1900 were an estimated 1 per 100,000. This,during a time when practically anyone could buy a gun. No restrictions, no background checks.
Even allowing for advances in data gathering since then, there can be little doubt that the current rate of 5.6 murders per 100,000 cannot be explained away by blaming it on the availibility of firearms. But rather it is the willingness to use them, especially among youth, that is the real problem – one that many of todays leaders seem unable or unwilling to deal with in any meaningful or realistic manner.
<
p>
Tougher penalties for crimminal behaviour, domestic violence, and addressing the asocial content in pop-culture entertainment are the ways to reduce gun crime.
There were over two thousand two hundred murders in New York City in 1990. By 2002 that figure had been reduced by 75% – because Mayor Giuliani took a tough stand on all crimes at all levels, from petty theft right on up to violent assault.
He didn’t take the easy road of anti-gun rhetoric and gun manufacturer lawsuits to do his heavy lifting. He rolled up his sleeves, and made New York City the wrong place to be a bad person.
In other gun news – Mittrick’s Latest Flipety-Flopety. Yepper, he was for gun control afore he was agin’t. Now Willard claims to be an NRA member (did he even let the ink on the ID card dry before the proclamation?).
This (not the Yahoo article, a press release from Harvard relating to the paper on which the Article is apparently based) is being discussed over at Tim Lambert’s blog Deltoid, URL http://scienceblogs…. (Tim Lambert is a computer scientist who appears to specialize in statistics.) Lambert provides a link to the paper on which the press release is based, but apparently you have to pay to get to the paper.
<
p>
I haven’t read the paper, but according to Lambert, the press release indicates that
<
p>
<
p>
I would interpret that as indicating that
<
p>
(i) the states in the highest quartile of firearm prevalence had over twice as many firearm homicides than those states in the lowest quartile of firearm prevalence, but that
<
p>
(ii) the states with the highest quartile of firearm prevalence had only a little more than half-again as many total homicides (including firearm homicides) than those states in the lowest quartile of firearm prevalence.
<
p>
Of course, as one commenter noted above, correlation does not indicate causality, but there seems to be a rather significant relationship between firearm prevalence and homicides committed with firearms. Query why that is.
<
p>
The same commenter apparently complained about the fact that the study was “controlled” for various factors. The fact is that, in order to determine whether there is a statistically-significant relationship between two variables, call them variable A and variable B, one needs to filter out (i.e., control for) the possible influence of other potential variables C, D, E, F, etc. There are several standard statistical methodologies for doing so, and that is what the researchers did. If one wants to consider the influence of another variable, such as variable C on variable A, then the researcher would control for variables B, D, E, F, etc. And, if one wants to combine two variable, that’s possible in the same way.
<
p>
If someone wanted to find out the influence of other variables suggested by the commenter, one could certainly do so. Fund a study. Although one might seriously have to ask how a statistician is supposed to control for a variable entitled “police under-staffing” when the appropriate staffing levels are within the eye of the beholder.