NRA: Lies, damn lies - and statistics

Like many others, I was horrrified by the Sandy Hook shootings, but I was rather taken aback by the NRA's statement on the affair, blaming (in essence) mental health and computer games for the shootings, and not the US's very high gun ownership rates - the Grauniad has the full speech, over here, but here is the passage on mental health:

The truth is that our society is populated by an unknown number of genuine monsters — people so deranged, so evil, so possessed by voices and driven by demons that no sane person can possibly ever comprehend them. They walk among us every day. And does anybody really believe that the next Adam Lanza isn't planning his attack on a school he's already identified at this very moment?

How many more copycats are waiting in the wings for their moment of fame — from a national media machine that rewards them with the wall-to-wall attention and sense of identity that they crave — while provoking others to try to make their mark?

And here is the one on Computer Games:

And here's another dirty little truth that the media try their best to conceal: There exists in this country a callous, corrupt and corrupting shadow industry that sells, and sows, violence against its own people. Through vicious, violent video games with names like Bulletstorm, Grand Theft Auto, Mortal Kombat and Splatterhouse. And here's one: it's called Kindergarten Killers. It's been online for 10 years. How come my research department could find it and all of yours either couldn't or didn't want anyone to know you had found it?

Then there's the blood-soaked slasher films like "American Psycho" and "Natural Born Killers" that are aired like propaganda loops on "Splatterdays" and every day, and a thousand music videos that portray life as a joke and murder as a way of life. And then they have the nerve to call it "entertainment."

So, do these claims hold up? Is the US uniquely in the grip of homicidal monsters and video-gamers? Well, I had a look at the UN OCD data and the Guardian's data vs the 9 OECD countries with the next highest gun ownership rates. Being Western, democratic OECD countries they are those most likely to be similarly afflicted with said monsters and gamers, in the table above. My assumption being that, given not too dissimilar lifestyles and access to videogames, the populations would have a not too dissimlar level of mental illnesses (in all their guises) and vile video game playing.

The results are quite interesting (see chart above). Firstly, US Homicide rates overall are about 5 times higher than the average of the other 9, and US gun ownership rates as a % of population are about 2 1/2 times higher than the "high gun owning OECD" average. (this is a by country average, by the way, if I adjusted it for populations it would be lower, as France and Germany - the most populous countries - are less violent than the smaller ones up there).

But now look at the middle columns - US Non-gun homicides are only about 2 1/2 times higher than the other countries rates (1.98 vs 0.72) but the 4th column is very interesting - the rate of gun ownership as a predictor of non-gun homicides is near enough the same in the US and the 9 OECD countries, 45 to 47 thousand guns per non-gun homicide. In other words, the % of guns in the population as a predictor of non-gun related homicide is near-as-damnit exactly the same as in the 9 OECD countries on average - in fact, Americans are less likely to kill each others in non-firearm ways than Swedes, Finns and those very violent Canadians for every extra gun they buy. the problem is , they buy a lot more than those Swedes, Finns and Canadians.

From that you can conclude that US society being more afflicted with "Monsters and Video Gamers" than the other 9 OECD countries is unlikely, in fact its those Polar Region countries that need to keep an eye on their videogamers...the problem probably lies elsewhere.

So, look at the last 2 columns - gun related homicides - the US rate of 2.97 homicides per 100,000 is 10x that of the other OECD countries' 0.3. And the last column sums it up - you need 30 thousand US guns per gun homicide in the US, and 115 thousand guns per gun homicide in the other OECD countries - so rather than having roughly the same ratio of guns needed per non gun homicide, the guns neded per gun homicide is 1/4 in the USA, ie you are 4 times more likely to get a gun homicide per gun sold in the US than the other OECD countries.

So in the US, yes, people do kill people - but its not that much of an outlier vs "rest of OECD". But in the US, per head of population, guns definitely do kill people a hell of a lot more than elsewhere. Which leaves me to conclude (contrary to NRA assertions) that People kill people, sre - but Guns kill people much better!

Now these statistics, to be sure, only show correlaton - but the totally different rates for gun and non gun homicides, and the similarity to the OECD for non gun homicides per % guns owning population, argues rather strongly for a hypothesis that there is in fact some form of causation, i.e that ready access to guns leads people to ready access to "gun homicide" solutions to their problems.

In other words, if gun ownership was brought down by a factor of 2.5 to look more like OECD levels, gun homicides would very probably show a far steeper decline, by a factor up to 4x.