“Correlation versus causation in a single graph”

47D7zGq

Via @dgardner, @clicjf.

164 Responses

  1. Pingback: Correlation isn’t Causation: Will Eating at McDonald’s Give you Acne?
  2. Pingback: Correlation isn't Causation: Will Eating at McDonald's Give you Acne?
  3. Pingback: Correlation : Causality :: IE Market Share : Murder Rate » Notes from the Quant of Asia

  4. Jeff H.
    …”after all?”
    Were laboring under the impression that Bill Gates was a socialist?

  5. Well, it seems Monsieur Gates has found a way to REALLY help world health: give up market share, save lives. Wonder if he’s a closet socialist after all?

  6. The question here is which way does the causation run? Is it that the decrease of IE means that people are less frustrated by their browsers and therefore less likely to be driven to murder? Is it that the decrease in murder rate means there are fewer people going to IE and searching for information about the optimal location and depth of shallow graves? Is it that the murder rate had been artificially inflated because people were murdering those MS employees responsible for IE and now that most of them are dead, customer support is worse than ever and so users are moving to other browsers? Clearly additional research is required.

  7. Pingback: Decrease in Internet Explorer usage leads to lowering of murder rate | Technologyscout

  8. But that kind of graph is misleading as there’s not even a correlation let alone causation implied, since the trend isn’t one to one like the graph implies. IE use went down near 50% where murders went down only 15%. Or is that part of the joke…

  9. Pingback: Moving On With Life - CoupleDumb

  10. If only that were actually a comparison of murder rate with IE market share. The number of murders annually isn’t the same as the murder rate, which I believe is usually described in per capita terms.

  11. Wow — this is a hat trick: Not one but TWO “gee-whiz” graphs, as the great Darrell Huff defined them

  12. Billlls first law of election year rhetoric: When a pol speaks of numbers, the real story in in the percentages, and vice versa. Comparing numbers to percentages is a perfect example of this.

  13. I dunno… Windows 8 makes me wanna give it the gun sometimes, especially the new IE I have to use because Chrome and Southwest Airlines do not play well with each other.

  14. But isn’t this true of virtually every Microsoft product besides Excel?

  16. Having struggled with Microsoft’s crappy OS’s and software for 30 years, I see a strong causative relationship. :-)

  18. Pingback: Correlation vs Causation | Insomniac memos

  19. I’m going with Causation. As screwed up and demanding as IE – and Microsoft in general – has been over the years, I could well believe that thousands of users have been driven to insanity.

  21. Maybe there is a tiny bit of causation….

    If Chrome and Firefox cause Internet Porn to load up faster than Explorer did, perhaps that causes the need for the male brain to see porn to be satiated sooner, and pushes the psychology that leads to murder just beneath the surface.

  22. Pingback: “Correlation versus causation in a single graph” | tpuofuosutfud