Internet Explorer causation

13 Comments

• Filip — January 25, 2013 at 3:36 am

  No Tumblr, but close enough: http://filipspagnoli.wordpress.com/?s=%22correlation%22%2B%22causation%22
• David Higgins — January 25, 2013 at 4:08 am

  http://imgs.xkcd.com/comics/correlation.png
• Jasja — January 25, 2013 at 4:57 am

  Nice. But you should really label your y-axes!
• John Nelson — January 25, 2013 at 6:22 am

  Look, an immediate impact, no lag rate! :)
• Kartik Trivedi — January 25, 2013 at 7:31 am

  I think there is an implicit theoretical underpinning – Lower internet explorer market share corresponds to freer internet, more active internet which then without any lag (as noted in comment above) corresponds to more informed public which means more public vigilance hence lower crime. :)
• 5up Mushroom — January 25, 2013 at 8:05 am

  I redid this yesterday to better show correlation. Came up with an R-squared of .965 There must be something to it ;)

  http://5upmushroom.tumblr.com/post/41380940072/i-changed-this-chart-around-a-bit-to-better-show
• Javier — January 26, 2013 at 2:00 pm

  There is no way to assume causality from this evidence (or lack of it)
  • Jasja — January 29, 2013 at 1:12 pm

    Javier, i think it is a joke/message. Much better than the old stork/birth rate graph they showed me in school back in da day.
    • Al — February 4, 2013 at 2:46 am

      The funny thing is, there are many serious criminologists suggesting that recent (last few years) unexpected declines in crimes including graffiti, vandalism and gang involvement (so presumably also, including gang-related murder) are caused in part by an abundance of technology-driven better things for young people to do – including smartphones and games consoles: which of course use web browsers other than IE, lowering IE’s market share.

      So there could actually be a real causal link here… :-)
• Meic Goodyear — January 28, 2013 at 2:23 am

  Whoever produced this ought to have a basic principles graphics revision course – a column chart not beginning at zero!? Tut tut!
  Better anyway if you are showing a correlation to do both series as lines or scatter plots.

  If it were causal, which way?
• Chris W — January 28, 2013 at 9:04 am

  1. This is freakin’ hilarious!
  2. Wow.. some people just take life WAY to serious.
• Kyle Alm — February 21, 2013 at 8:34 am

  That is just awesome. It’s probably some whacked out designer killing IE 6 users.
• Shaun — February 23, 2013 at 1:16 pm

  “Is there a meaningless-correlations tumblr yet?”

  Yes: http://stats-from-correlated.tumblr.com