Comment Re:That's nice (Score 1) 717
I bothered to read it and there are some cringeworthy abuses of statistics in there. I'm not particularly on either side of the gun control debate at the moment, either, so I didn't go in looking to pick a fight with either side. Just to point out some of the failings, though:
They use Greenland as one of their prominent countries with a surprisingly high murder rate compared to the US. Seriously? Your statistic is muders per 100,000 people and you focus on a country with a population of around 50,000? These numbers are far too sensitive to noise to take any real meaning away from. And the other countries cited weren't terribly surprising to see. They're all countries that still have some pretty extreme cases of poverty. Desperation leads people to do terrible things. We're one of the world's wealthiest countries. I would hope to see us compared more favorably to the other wealthy countries of the world.
Then we get into that whole "well, it's not so bad if we don't count black people" bit. Let's ignore the potentially offending bit of that and focus on the numbers. Oh! So if we remove our underclass murder rates, then things are closer to Europe. Well, yeah. Because we removed all of the underclass murders from our numbers, but not theirs. This isn't a reasonable comparison to make at all. If you're going to futz with the numbers, you at least need to futz with ALL of them so they're on equal ground.
Finally, we get to my favorite. The correlation between gun ownership and murder rates by country. There's a pretty chart of the data! The author's eyeballs conclude that it looks like higher gun ownership is related to lower murder rate. Now, I look at the same chart and mostly come away thinking there's not much correlation either way, but if I had to choose a side, I'd side with the author and say things are a slight lean in his favor. Thankfully, the author has brought statistics to the table this time and goes on to say that there's a statistically significant -0.23 correlation coefficient between murder rate and gun ownership.
...beg pardon?
This is such a small correlation coefficient as to be essentially meaningless. Checkthis page on correlation coefficients and look at the visualization of what different coefficients look like near the top. Correlation of -0.23 is basically nothing.
Again, I'm not an advocate for a lot of this increased gun control legislation. A lot of it has sounded pretty terribly written and thought out. I am an advocate for the proper use of statistics to back up arguments, though, and this article doesn't do it for me. This isn't the first time I've been to American Thinker, and honestly, it's becoming a game for me to go through and see where the authors misuse numbers. That's why no one wants to take your article seriously. You have sources, but the existence of sources does not make their content credible.
They use Greenland as one of their prominent countries with a surprisingly high murder rate compared to the US. Seriously? Your statistic is muders per 100,000 people and you focus on a country with a population of around 50,000? These numbers are far too sensitive to noise to take any real meaning away from. And the other countries cited weren't terribly surprising to see. They're all countries that still have some pretty extreme cases of poverty. Desperation leads people to do terrible things. We're one of the world's wealthiest countries. I would hope to see us compared more favorably to the other wealthy countries of the world.
Then we get into that whole "well, it's not so bad if we don't count black people" bit. Let's ignore the potentially offending bit of that and focus on the numbers. Oh! So if we remove our underclass murder rates, then things are closer to Europe. Well, yeah. Because we removed all of the underclass murders from our numbers, but not theirs. This isn't a reasonable comparison to make at all. If you're going to futz with the numbers, you at least need to futz with ALL of them so they're on equal ground.
Finally, we get to my favorite. The correlation between gun ownership and murder rates by country. There's a pretty chart of the data! The author's eyeballs conclude that it looks like higher gun ownership is related to lower murder rate. Now, I look at the same chart and mostly come away thinking there's not much correlation either way, but if I had to choose a side, I'd side with the author and say things are a slight lean in his favor. Thankfully, the author has brought statistics to the table this time and goes on to say that there's a statistically significant -0.23 correlation coefficient between murder rate and gun ownership.
This is such a small correlation coefficient as to be essentially meaningless. Checkthis page on correlation coefficients and look at the visualization of what different coefficients look like near the top. Correlation of -0.23 is basically nothing.
Again, I'm not an advocate for a lot of this increased gun control legislation. A lot of it has sounded pretty terribly written and thought out. I am an advocate for the proper use of statistics to back up arguments, though, and this article doesn't do it for me. This isn't the first time I've been to American Thinker, and honestly, it's becoming a game for me to go through and see where the authors misuse numbers. That's why no one wants to take your article seriously. You have sources, but the existence of sources does not make their content credible.