The logical contortions that antis have to undergo to hold on to their dogma never cease to amaze me. Let’s look at the latest “study” the antis are trumpeting. It looks at changes in the firearm-related suicide rate in Connecticut after they passed a permit to purchase law and in Missouri after they repealed their PTP law . . .
Emerging from the Johns Hopkins Bloomberg School of Public Health ,the lead author of the study is Cassandra (and isn’t that one helluva given name for an anti-gun fabulist researcher!) Crifasi PhD, MPH, “an injury epidemiologist focused on policies, procedures, and practices that improve safety and prevent injury.” The headline breathlessly proclaims Suicide-By-Firearm Rates Shift in Two States After Changes in State Gun Laws and it starts off with:
A new study examining changes in gun policy in two states finds that handgun purchaser licensing requirements influence suicide rates.
Okay, so what method did the authors use to reach their conclusion?
We used synthetic control modeling as the primary method to estimate policy effects.
Whatinthehellis “synthetic control modeling” you ask? Excellent question; I had to do some digging on this one myself. Basically what Cassie and her peeps did was to use a dozen or so “factors previously associated with suicide rates” to look for states which were demographically similar to their test states (CT and MO). According to an economist friend of mine (PhD from the U of MN’s Carlson School of Management) this technique is extremely useful when you already know what you want your study to prove, because cherry picking which states to use as controls is somewhat subjective. You just do data runs with different control groups until you get the results you want and then declare that your “synthetic control” is substantially similar to your experimental group.
Not, of course, that I am in any way shape or form intimating that Cassie and/or her crew engaged in any sort of academic fraud or chicanery. Perish the thought.
Anyway, on to their findings:
Researchers estimate that Connecticut’s 1995 law requiring individuals to obtain a permit or license to purchase a handgun after passing a background check was associated with a 15.4 percent reduction in firearm suicide rates
But . . . when we skip down a few paragraphs we find that Cassie “cautions the findings do not indicate a clear causal relationship.” In fact, if we look at the graphs (see below):
Solid line is Real Connecticut
we discover that passing a permit to purchase law made non-firearm suicides increase in both real and synthetic Connecticut. Hmmm, substitution effect anyone? Perhaps if we could look at data from a larger sample, one that has a significantly larger population than Connecticut or Missouri (or even Connecticut and Missouri) and one doesn’t involve systematic lies I mean “synthetic controls”, someplace where say in 1977 they passed truly draconian gun control, someplace like Canada perhaps? As Dave Kopel points out in his The Failure of Canadian Gun Control:
Suicides involving firearms fell noticeably after 1978, reversing the previous trend. The overall suicide rate, however, did not drop …
And there’s the rub; antis around the world jumped up and down yelling and screaming about how successful these laws were because of the drop in firearm related suicides while utterly ignoring the overall suicide rate. In fact, Cassie even admits it more fully when she states:
“When we examined whether there were changes in suicides committed by other means following the changes in the laws, there was some evidence that Connecticut experienced lower than expected rates of suicides by means other than firearms,” she says. “This suggests that factors other than handgun purchaser licensing may have contributed to the decline in suicides.”
In that case you might want to change the title of your “study” then, Cassie.
So having completely destroyed their own credibility in Connecticut what does Cassie’s posse have to say about Missouri?
… while Missouri’s repeal of its handgun purchaser licensing law in 2007 was associated with a 16.1 percent increase in firearm suicide rates …
There was no significant change in suicide by other means following Missouri’s repeal of the law.
Now I freely admit that I am not a PhD, or MHP or even BS; I am more of an OFWG. I do not work for any of the myriad bought-and- Everytown paid-for Bloomberg sock puppets *AHEM* coalitions, campaigns, or centers who strive to ban guns across the board improve the Safety of Our Children; heck, I don’t even work for the Second Amendment Foundation any more (but that’s another blog entry). I am just a guy with a computer and access to the interwebz who has managed to track down some information sources over the years.
One of my favorite sources is the CDC’s WISQARS™ (Web-based Injury Statistics Query and Reporting System site or, as they explain on their front page, “Your Source for US Injury Data”). Despite the CDC’s institutional anti-gun bias, they are scrupulous about the accuracy of this data providing an excellent resource for researchers of all stripes.
This brings us back to Missouri; in 1998 the CDC significantly changed the way deaths were coded, so I pulled MO’s data from 1999 to 2013 (the most recent year for which the data are available). Then, without processing, synthesizing, accurizing or even folding, spindling and/or mutilating said data I put together this table:
Please note that in 2007 the rate of Gun Related Suicides (hereinafter to be referred to as GRSs) dropped for a year before starting to rise. Also please note that non-GRSs continued the generally upward trend that existed before the law was passed. Finally, look at the upward trending overall suicide rate for the whole country and note that MO’s overall rate shows a similar trend.
But now I have some data that will surely turn that frown upside-down! It turns out that when Missouri repealed their permit-to-purchase law we had a significant drop in the number of . . .
That’s right, from 2007 – 2013 the rate of accidental drownings dropped thirty-four point zero-two percent!
For the umpteenth time, correlation is not causation.