XDA

Here is a bit of coverage from the Nation about a study I find incomprehensibly stupid. Unfortunately, the story contains no link to the study itself, which is pretty damning right there.

So let's look at the story. The headline is: Actually, Guns Do Kill People. No, actually they don't, but people do use guns to kill people. And that happens all over the world at varying rates in each country. These differences in rates of gun murders have very little to do with gun control legislation. These differences have everything to do with the culture of the country.

Here is the sub-headline: The research is now clear: Right-to-carry laws increase the rate of violent crime.

Based on that statement, one would expect the research which the story is talking about actually to show increases in violent crimes in states that have a shall-issue, concealed carry permit system. Almost all of America has that system, 42 states. Only 8, all solidly blue on the political color chart, let law enforcement deny qualified citizens the right to carry guns concealed about their person.

By way of brief history. In the 80s it was just the opposite; almost all the states allowed law enforcement officials to deny applications for concealed carry and as the sweeping change started, the gun haters all predicted that increasing the number of concealed carry permits would cause increased violent crime, especially assaults and murders with guns. It would be the wild west again and the streets would run with blood. But that never happened. The early 90s were the high tide of gun murders in America and despite the proliferation of shall-issue permit systems and a lot more guns in private hands over the next 20 plus years, the murder rate plummeted by over 42%. Here are the figures from the FBI for 2010 and 2015.

So I was eager to see the research that found the opposite of what the crime statistics clearly showed, that is, an actual increase in murders and assaults (most using guns) after the establishment of shall-issue permit systems. I have to admit, having read the stats for a few decades, I was very skeptical. But I was ready to learn.

Unfortunately, the research is not about actual figures but about what might have been. Look at what the story in the Nation describes;

But with an updated paper by legal scholars John Donohue and Abhay Aneja and economist Kyle Weber, there’s a new consensus: Right-to-carry laws actually increase the rate of violent crime. Ten years after a state passes a right-to-carry law, violent crime—which includes murder, manslaughter, forcible rape, robbery, and aggravated assault—will be 13 to 15 percent higher than if the state had done nothing.

Wait. Since when is a single, unpublished paper a "new consensus"? It get's worse. The study calculated what the future rates of violent crime were supposed to be in each state? How? The story doesn't actually say, but it does tell us that there is a lot more data now. Yeah, data showing a massive decrease in violent crime since 1993. Here is the beginning of the Nation's explanation:

Starting in the early 1990s, violent crime plummeted across the United States. That reduction has masked the effects of right-to-carry laws, but the states that implemented them showed a smaller decrease in violent crime than the ones that didn’t.

Hold on there, kitty cat. Masked? How about right-to-carry laws were a primary cause of the plummet in violent crime?

OK, let's see what the rates of shall-issue states were against may-issue states. It will have to be a small sample, cause I just don't have to time to do more than two states. Let's compare Wyoming to Delaware. Wyoming passed shall-issue legislation in 1994; Delaware never did. The study says the incubation period from this change, or lack thereof, is 10 years. So, let's look at Delaware's murder rate and aggravated assault rate from 1990 to 1994 and from 2005 to 2009 and compare that to the Wyoming rates for both over the same time period. Here are the stats for Delaware. Here are the stats for Wyoming.

OK, from 1990 to 1994 Delaware had 170 murders (most will be from gunshot wounds but there was no break-down by weapon). That averages out to 34 per year and 5.12 per 100,000 people (I used the 1990 population number) After 10 years of not giving out concealed carry permits to every qualified citizen who wanted one, for the period 2005 to 2009, Delaware had 214 murders or 42.8 per year and 4.93 per 100,000. That's an actual rise in murders but per capita, it's a fall of .29 murders per 100,000 or a 6% fall in that rate. OK, at least it's going down.

From 1990 to 1994 Wyoming had 80 murders (less than half Delaware's even though the population difference between Wyoming and Delaware then was less than 1/3rd). That averages out to 16 per year and 3.45 per 100,000 people (still used the 1990 population number) After 10 years of giving out concealed carry permits to every qualified citizen who wanted one, for the period 2005 to 2009, Wyoming had 62 murders or 12.4 per year on average and 2.28 per 100,000. That's an actual drop in murders and, per capita, it's a fall of 1.17 per 100,000 or a 33.9% fall in that rate.

But wait, I thought the study the Nation is swooning over showed that Wyoming, compared to the closest in population blue state not handing out concealed carry permits like candy, should have had a 13 to 15% rise in murder rate after 10 years, or something like that, not a more than 1/3rd drop in the murder rate after 10 years of shall-issue carry permits? Hmm. Maybe that's an outlier. Let's do aggravated assault.

From 1990 to 1994 Delaware had 11,400 agg assaults (again most will involve guns). That averages out to 2280 per year and 342.34 per 100,000 people. After 10 years of not giving out concealed carry permits to every qualified citizen who wanted one, for the period 2005 to 2009, Delaware had 18,642 agg assaults or 3,728 per year and 421.28 per 100,000. That's an actual, substantial rise in the number of agg assaults and a per capita rise of 23%. Ouch. Most other states were having a general, substantial lowering of most violent crimes. Hmmm.

On the other hand, from 1990 to 1994 Wyoming had 5,737 agg assaults (again nearly half as many as Delaware). That averages out to 1147 per year and 253.29 per 100,000 people. After 10 years of giving out concealed carry permits to every qualified citizen who wanted one, for the period 2005 to 2009, Wyoming had 4927 agg assaults or 985.4 per year and 181.13 per 100,000. That's an actual, substantial drop in the number of agg assaults and a per capita drop in the rate of 28%. Delaware's went up 23%.

Perhaps Delaware should start handing out concealed carry permits like candy to the willing, qualified citizens. It's the only safe thing to do. But my calculations are not taking into account what these crime rates would have been in Wyoming had there been no change in the law regarding concealed carry permits.

But I have to think that these actual, historical numbers make the "study" and the story on it pretty much complete bullshit.

How did the "study" get it so completely wrong? Here's an interesting admission:

The authors of this new paper have taken advantage of cutting-edge statistical techniques. They constructed synthetic control groups for states and used what’s known as a LASSO analysis to pick the best variables for comparison.

Synthetic control groups? Why not just look at the actual rates? But the question stands and remains unanswered. How do you know what the rate of murders and assault in a state would have been if the state hadn't gone to shall-issue permits?

Is LASSO analysis something akin to crystal ball gazing?

I presume because the actual, historical crime rate statistics didn't give the gun haters any data with which to support their efforts to take away law abiding peoples' civil rights, they had to go with this mumbo-jumbo regarding estimates of crime rates in an unreal place with an alternative history. Normal people might well think "what might have been" is pretty much unknowable in an alternative universe, LASSO analysis notwithstanding.

Really stupid study that shows pretty much nothing at all.

Labels: Gun Control: Really Stupid Study