The Gun Study Critique that Begat the Race Card

Apparently it has become necessary to explain my critique of the gun study and Greg's post on it, because Greg seems to think the argument is still about race and class. He won't explain how exactly, apparently because I didn't ask very nicely - but my argument is apparently about race and class none the less. No matter that he still doesn't seem to have the foggiest clue what I am actually arguing - he thinks he does and he thinks that the argument is racist or classist - or both.

Here is Greg's post on the study and here is the press release on the study. Unfortunately the paper itself is behind a paywall, but for those who might have access, here is the link to that. Greg actually posted the conclusion from the abstract, which is basically the same conclusion of the paper itself.

Greg starts right off by flying way past anything that either the study or the press release had to say, titling his post Carry a gun = you get shot more often. Well no, that is not at all an implication of the study or the press release - but it makes a hell of a headline. What the press release indicates, is that carrying a gun is of dubious value in the event of an assault. It mentions that you are more likely to be shot during an assault if you are carrying one. While I think the study itself is still over generalizing (something the authors themselves admit may be the case), they are limiting their generalization to urban settings.

Before I go into my critique, I will make the same caveat that I made in my comments at Greg's blog. I tend to think that the conclusions being made are likely true. I think that the study's conclusions, the press release conclusion and even Greg's conclusion in the title of his post, are quite likely true and a well constructed study that was able to gather data to support those conclusions would likely find evidence to support those assertions. I just don't think this study reasonably supports any of those assertions because of a serious flaw that attempting to adjust for in the controls does not - cannot compensate for.

The problem with this study, is that the vast majority of the data comes from shootings in areas fo Philadelphia that are prone to extremely high rates of gang activity. The study does not take into account the legality of the firearms in question, or the victim's legal right to be carrying them concealed on their person. The only adjustment made in an attempt to compensate, is that the controls for each case in the study, were taken from the surrounding area and culled for like employment - in most of these cases, unemployment and criminal record - actual crime/s undetermined. Because of this, I do not see how whether or not a given case in the study was a career criminal (i.e. a gang member) can be accounted for.

Now as far as Greg's title assertion goes, this pretty much knocks it out of the running entirely. With a very strong likelihood that a very large percentage, if not a majority of the cases in the study being involved in criminal activities, there is a confounding factor having nothing to do with guns, that makes it more likely the person is going to be assaulted in the first place. I don't think it is at all unreasonable to assume that someone involved in ongoing criminal activities is more likely than people who are not, to be victim of an assault in the first place. To be very clear again, I suspect that Greg's claim is correct. I just don't see this study, or any study that is limited entirely to large urban centers (the gang variable aside) providing any evidence that that is the case.

The assertion made in the press release is rather trickier, but not all that much so. Bottom line, their claim seems to be that carrying a gun in general - regardless of location, increases your chance of being shot during an assault. We won't go into the gang activity variable at all and rest with the obvious - the study authors did not make that claim. Presumably because they did not believe that the evidence they studied could reasonably be extrapolated to populations outside of urban centers. From the paper's conclusion; "Although successful defensive gun uses are
possible and do occur each year, the probability of success may be low for civilian gun users in urban areas." Neither that statement, or any other in the paper's conclusion makes any implications about gun use outside of urban areas.

But what about the study itself? For that matter, why would anyone who is not a criminal, have a substantively different outcome during a gun assault? What exactly makes that such a confounding variable?

First and foremost, I want to be perfectly clear that it may not make any difference at all. But there are a lot of factors that could potentially make an assault involving career criminals, different than an assault involving people who are not. Much the same way there are a lot of factors that make an assault involving a police officer different than an assault that does not involve the police. The study did not include police officers who were shot for a reason, I think it is reasonable to say that the study should not include mostly career criminals for similar reasons. Bottom line, they are not civilians in the same sense that police officers are not.

Those factors include, but are not limited to the following. First, civilians with guns (CWGs) who are not criminals and who get involved in an assault, may not be very likely to pull their gun because of a simple assault, while a career criminal may be considerably more inclined to do so. Second, a gangster is very plausibly considerably more prepared to use their gun at any given point, than your average CWG - being so prepared may actually make it less likely the gangster will be shot. Third, gangsters in larger urban centers often operate with a certain amount of discipline and training - some gangs operate on a par with tactical (paramilitary) police units. This could well provide even more of an advantage in an assault, over CWGs. And gangsters are also very likely to attack from some sort of cover and with a certain amount of stealth or surprise - such as drive by shootings.

I consider this a confounding factor that at the least raises doubts about extrapolating this study to the general population. And it would not matter if the study included a large percentage of white gangsters, or any other organized criminals. Indeed, including a large percentage of career or even petty criminals, would create confounding factors. As would including a large percentage of terrorists. It doesn't matter if the gangsters are wealthy or poor, living in the ghetto, or living in extremely wealthy neighborhoods. Being a gangster is the confounding variable, not what sort of gangster one might be or where they live.

I am more than happy to debate this - I am happy to do so here. I could well be wrong and someone might have a stellar argument that convinces me that I am. But my argument has absolutely nothing to do with race or class. Accusations that I am racist or classist, or that my argument itself is those things, is a total non-starter.

Worse, as I explained in my other post about this, such accusations are extremely counterproductive - whether they come from a privileged white person, or from someone who has suffered actual racism. All they accomplish is to desensitize people to such accusations and make them incredibly unlikely to have the slightest interest or willingness to engage in a serious discussion about racism, classism and/or other forms of bigotry.

It also has the added detriment of seriously detracting from, if not outright destroying that person's credibility when it comes to discussions about racism and other forms of bigotry. It is kind of hard to take someone seriously about racism, when they quite unjustifiably accused you of racism.