Reply #56: Yeah, keep beating that "denialism" drum; you might even convince yourself

[Euromutt]

<...> in research circles, it is the rate of gun ownership (i.e. % of households and/or % of individuals owning guns) which is the key statistic, rather than total number of guns in circulation. And for good reason: if a household with 10 guns buys 10 more, it doesn't increase gun availability very much. But if 10 new households buy a gun, that's a bunch of new people that now have access to a gun, and this is where you get the increase in homicide and suicide rates.

That argument would hold water if there were a discernible correlation across countries between rates of gun ownership and homicide and suicide rates: there isn't.

I note, moreover, that you blithely skate over my point about the increase in states with "shall issue" legislation for concealed carry permits. It's certainly been the position of gun control advocacy organizations like the Brady Campaign and the Violence Policy Center that more liberal carry laws result in more people having a firearm available to them in public, thereby (per the "more guns => more death" hypothesis) causing a higher risk of homicides at least.

As an aside, you seem to be assigning this sort of informal analysis a very precise amount of weight: too little to draw a positive conclusion, and yet enough to disprove scientific research supporting the opposite conclusion.

What's odd about that? It is unfortunate but true that it is easier to point out flaws in research than to perform it oneself. That's a large part of why science--particularly social science--produces far fewer positive conclusions (i.e. identifying what is true) than it does negative ones (i.e. identifying and subsequently discarding what is false). At least, if we're talking about conclusions that withstand further scrutiny. The inability to discern what is true does not preclude the ability to identify what is false.

But back to reality, where nobody serious thinks that, taken alone, the coincident drops in gun ownership and homicide constitute much evidence for anything, because there are other major demographic forces at work.

No argument from me on that point. Of course, that raises a problem for the "more guns => more death" hypothesis, in that (as I keep pointing out, and you have not seen fit to dispute) the evidence for this hypothesis consists entirely of findings from the very kind of studies that are ill-equipped to control for other variables. This is precisely why their findings need to be replicated by more rigorous types of studies; no amount of futzing with the numbers in an effort to control for variables can do the job as effectively as eliminating a variable by eliminating it from your control group beforehand, which is only possible in prospective studies. That is, the kind of study which those public health researchers who focus on firearms-related issues have yet to perform (and always will have yet to perform if past performance is any indication).

Because if there are other demographic and socio-economic forces at work that influence levels of violent crime (including homicide) and suicide, then why should we assume that it will be more cost-effective to tighten restrictions on private gun ownership than to address some of those other forces? How would we determine that increased gun control is even having a positive effect at all? The very fact that the gun control lobby has resorted to phrases like "if saves just one life, it'll be worth it" is an admission that they at least don't think it is possible to tell (much as it would be in their interest for it to be possible).

I might add that I don't dispute that there are demonstrable correlations between violent crime and firearms; where I have doubts is that the number of firearms, or the number of people who own them, is a significant causal factor in violent crime. When people prone to violent and/or criminal behavior feel they need firearms, they will acquire them and use them, and you'll see a resultant increase in firearm violence, much like parts of western Europe have been seeing over the past 15 or so years. But the "more guns => more death" hypothesis is founded on some extremely questionable premises, not least that an overwhelming percentage of homicides is committed by "regular" individuals with no notable prior history of violent behavior who happened to have a firearm available to them "in a fit of rage." To put it bluntly, this is a fabrication, a myth, albeit a particularly persistent one.

The solution is to take a rigorous, statistical, scientific approach to studying the issue. <...> Yet you dismiss them as "crappy studies", and no doubt you will continue to do so.

Indeed. I cannot improve upon the words of Ted Goertzel, professor of sociology at Rutgers, when he stated that:

There are, in fact, no important findings in sociology or criminology that cannot be communicated to journalists and policy makers who lack graduate degrees in econometrics.

Or, for that matter, public health policy. As I noted above, I found Lott's conclusions hard to accept, precisely because he had to crunch such a massive volume of data to produce the findings on which he based his conclusion, and I'm not inclined to defer to statistical one-upmanship from the other side either, especially when it produces results that strain credulity. Take, again, the 2009 study by Branas et al.: from a study population of shooting victims, in which those carrying a firearm about or close to their persons when they were shot are outnumbered by those not carrying almost 16 to 1, they manage to produce a conclusion that those carrying were 4.5 times as likely to be shot as those who weren't, it is perfectly reasonable to react with incredulity. Again, compare this to Richard Doll's findings that, among lung cancer patients, smokers outnumbered non-smokers 9 to 1.

But the fundamental reason I dismiss the public health research overall as "crappy" is because in study after study, the researchers establish a correlation but do not establish that a causal relationship exists or, even granting that it does, which way it runs (i.e. they do not account for the possibility that individuals who consider themselves at high risk of criminal assault, particularly due to their involvement in illicit activity, will acquire a firearm as a result) and then write their conclusions as if they had. As I've noted in posts responding to yours, Branas et al. acknowledged that they "did not account for the potential of reverse causation between gun possession and gun assault" (e.g. being a drug dealer is what makes you likely to get shot by competitors, and drug dealers carry guns because of that), but they buried that admission in the article itself; you sure as hell wouldn't gather it from reading the conclusions or abstract, let alone the (non-peer-reviewed) press release.

Similarly, with Kellermann's 1998 study "Injuries and deaths due to firearms in the home" (Journal of Trauma 45:263-267), you had to dig deep to discover that of assaultive shootings studied, 67.3% were known to have been committed with a firearm known not to have been kept in the household in which the shooting occurred. When you title your study "Injuries and deaths due to firearms in the home," and you then include instances of shootings committed using firearms brought in from outside, that's just rank dishonesty, both intellectually and otherwise.