Graphs and Statistics

The opening graph is interesting, but it's quickly criticized in the comments for selection bias. So defenders produce other graphs -- quite a few of them -- with supporting stories.

What to believe? Well, this is one way of addressing the question.

7 comments:

  1. Yewbetcha. Even the participants of this blog have our own biases, and self-selected as we tend to be, our biases tend to convergence. It's critical to read sources of differing views and viewpoints--differing, not just opposing.

    Except, of course, those sources that disagree with my august self. Those are plainly wastes of time.

    Eric Hines

    ReplyDelete
  2. Gringo5:26 PM

    Originally posted at the wrong article:

    I have compared gun ownership rates with murder rates on a worldwide basis. Worldwide, there is a -0.1 correlation between the two. In Europe- including Russia- there is a -0.32 correlation. Russia has a low gun ownership with high murder rate. Switzerland and Serbia have high gun ownership with low murder rate.

    In the US, the negative correlation gets even stronger.
    From the Pew Report: The demographics and politics of gun-owning households.

    The new research also suggests a paradox: While blacks are significantly more likely than whites to be gun homicide victims, blacks are only about half as likely as whites to have a firearm in their home (41% vs. 19%)

    Run that correlation between blacks and whites for murder rate versus gun ownership rate, and you will get -1. Can't get any stronger than that.
    At least the Pew Report didn't say "surprisingly."

    ReplyDelete
  3. Gringo5:34 PM

    One irony about the "we must compared the US to only Western Europe" crowd when it comes to murder rates or gun death rates is that they also are glad that the proportion of Western European descendants in the US has been going steadily down.

    ReplyDelete
  4. I've always thought that Brazil was a better comparison than Western Europe: a large American nation with a great deal of ethnic diversity, a troubled history of slavery, a colonial period, and so forth. The Brazilians have very strict gun control, and a very vast murder rate involving illegal guns.

    ReplyDelete
  5. Gringo - at least one person looking at the graph made the point that the US homicide rate ought to be much higher if it correlated with legal gun ownership.

    ReplyDelete
  6. The linked Vox article was astoundingly dishonest. All those demographic comments and never once a mention that the African-American homicide rate is 8 or even ten times higher, and the Hispanic rate two or even four times higher. You will note in the supposed apples-to-apples comparisons to Western Europe, Japan, Australia, etc, those high-violence groups barely exist.

    I get it that nice people think it's racist and troublesome to even mention it - and it is true that nearly everyone in any group is nonviolent and reasonable. But 10x is just huge, and it bends all the US statistics so that comparisons are meaningless.

    ReplyDelete
  7. Gringo11:16 PM

    AVI
    The linked Vox article was astoundingly dishonest. All those demographic comments and never once a mention that the African-American homicide rate is 8 or even ten times higher.
    Nor does the Vox article mention that while blacks have a higher murder rate and murder victimization rate than whites, blacks also have lower gun ownership rates than whites.

    While I doubt that the gangbanger murdering contingent of blacks would give a truthful remark to a survey question about gun ownership, the proportion of murderers among blacks is small, so a gangbanger's untruthful remark about gun ownership would not have a big effect on the overall answer about gun ownership.

    ReplyDelete