Saturday, January 2, 2016

Correlation if not Causation


  1. I see your error. You assume that every country in the world has uniform data collection and reporting on gun ownership and homicides. They do not.

    So, you have a very nice couple of graphs based on really good data for the USA, really good estimates for Western Europe, South Korea and Japan and really sloppy wild guesses, based on thin air, for the rest of the world. You can make the colors anything you like.

    1. Well yes we also have fun things like Western Europe doesn't add a death to the homicide statistics until they get a conviction, as opposed to our standard of if T-Dawg is lying bled out in the gutter and five sumdude's didn't see nuthin we still make the reasonable assumption that he didn't just trip and fall on a pile of bullets.

      However even given the above disclaimers you'll notice that on the world wide scale more guns per-capita seems to correlate to less gun homicide. It would just be sloppy thinking to ONLY use that to prove causation.