“Early rising may not be a vice ... but it is certainly no virtue. The old saw about the early bird just goes to show that the worm should have stayed in bed.”
by Lazarus Long




"walls of the city" logo conceptualized by Oleg Volk and executed by Linoge. Logo is © "walls of the city".

graphics matter, year the fourth, updated

[Please disregard this post, and instead reference this updated version, which has more data.]

No, unfortunately, the FBI has not released a finalized version of their 2011 Uniform Crime Report (it is scheduled for September), so I cannot update the second half of this post series, but commenter TS brought up a very valid point on the 2009 update for the CDC side of this post: what about firearms imported into America? I actually had that conversation in email with Howard Nemerov last year, but neither of us could figure out a single, consistent source of numbers for that particular statistic, so we never really got anywhere.

However, TS thankfully pointed me towards the Firearms Commerce in the United States, Annual Statistical Update which tracks production and import numbers from 1986 all the way up to 2011, and I had exactly what I needed.

Or, well, most of what I needed. As I mentioned previously, the Shooting Industry Magazine’s U.S. Firearm Industry Report has vanished from their site, and with it went all of my firearm production data from 1981 to 1986. I have searched high and low for such information elsewhere on the ‘net, and I have contacted SIM to absolutely no avail, and I have since concluded that if I cannot adequately source the data, I cannot really present that data as factual, so I have dropped 1981, ’82, ’83, ’84, and ’85 from the "Graphics Matter" dataset. I do not really want to do this (as I have always maintained, more data is more better), but if I cannot adequately source my numbers, I am no better than the "gun control" extremists whose arguments I am destroying with these posts.

So, without further ado, the newly-updated pretty picture:


As you can see, if you compare this one against the previous version (and remember that this one only covers from 1986-forward, not 1981-forward), the slope of the "Number of Firearms" line is significantly steeper, which also means the downwards slopes of the "Firearm-Related Deaths per 100,000,000 People" and "Firearm-Related Deaths per 100,000,000 Firearms" are actually noticeable. The quicker folks amongst you are already putting together what that might mean.

I am going to dispense with the acres of text the previous posts are known for (except to say that all of the important stuff in them is applicable here as well), and get to the stuff you really want to hear about… right after this important disclaimer: you CANNOT compare the correlation coefficients in this dataset to any of the coefficients in any of the previous datasets. By lopping off those five years at the beginning, we change the playing field, and the coefficient numbers are no longer comparable (which is why I really did not want to do what I did); however, I dare say these current numbers speak for themselves.

So, with that said, the raw number of firearms in America correlates to the raw number of firearm-related fatalities in America with a Pearson coefficient of -0.63966, an arguably strong, negative correlation.

The rate of firearm ownership in America correlates to the rate of firearm-related fatalities in America with a Pearson coefficient of -0.86207, an inarguably strong, negative correlation.

At this point, it would be safe to say that the hypothesis "more guns = more deaths" cannot be true when that hypothesis is applied to American society.

If you want to check my math and see what else I omitted, the file is available here, and please, do, check it. I never would have had the opportunity to add into the importation numbers if TS had not checked my work, and I am sure there are other little details I am missing somewhere.

10 comments to graphics matter, year the fourth, updated