Baron von tarv wrote:
Taking out them the graph is as close to exponential that any statistician would expect a real life data graph to be.
Lol. Only because the trendline was drawn as an exponential graph function. The thing here is that we can't say for sure what a "typical" slope should be. If we take a straight 1-1 ratio (every doubling of guns, doubles the rate of firearm deaths), we end up with about half the nations plotted (which is a pretty small sample already) above the line and about half below. We plot an exponential line, and about 2/3rds of the nations are above the line with about 1/3rd below.
What does that say? No way to know for sure. It means that some nations will have a higher slope than others. That's really it. And interestingly enough, more nations have a higher slope than the US when calculated as an exponential than as a linear trend. What does that mean? Again. That's hard to say except that for some number of factors some nations will have a greater ratio of increased firearms deaths as the percentage of people in the nation own firearms as others.
We can run around in circles trying to find out why (and honestly, it's not a bad thing to study), but to simply assume it's the guns themselves that cause this is silly. If that graph shows us one thing it's that there really isn't a universal correlation between the two.
And simplistic as it sounds, the expression that it's not guns that kill people, but people who kill people would seem to be born out by that data...
Edited, Oct 27th 2008 5:13pm by gbaji