Causation vs. Correlation

A New York Times columnist/blogger drones on with the usual nonsense about Bob Costas, the U.S. homicide rate, the Kellermann studies, etc., etc.

But hidden in the comments section was this nugget:

As a Canadian who's studied homicide stats pretty extensively, I have to tell you that I think you're barking up the wrong tree. My own home province of Newfoundland has a gun ownership rate of 45% of all households, almost exactly the same rate as the United States. Yet our homicide rate is a mere 1.4 per 100,000, compared to the US national rate of about 4.2 per 100,000.

However, looking back over the past decade, we cannot claim as low a homicide rate as North Dakota, which has been less violent than any Canadian province, and yet has one of the highest rates of gun ownership in North America, far higher than, say, Detroit (48.2 homicides per 100,000), or Washington DC (21.9 per 100,000).

The US homicide rate is about 2.5 x that of Canada. But 52% of US homicides are committed by African-Americans, who are only 12% of the US population. If you compare Canada with the other 88% of the US population, you will find that homicide rates are very close.

Likewise, comparing the 50 US states, 10 Canadian provinces, and 3 Canadian territories, the top three jurisdictions for homicide per capita are usually Louisiana, Nunavut (the Inuit territory), and the Northwest Territories, in that order.

Social problems and cultural attitudes drive homicide, and I say this not as an advocate for the gun lobby, nor as a believer in innate racial or cultural differences. I'm neither, just someone who respects what the numbers tell us.