Sunday, May 11, 2014

An important website for you.

I'm not normally one to pass along every good website I run into; in fact those websites that are AMAZING because fact number 23 will BLOW YOUR MIND are pretty annoying, in my opinion.   So I guess that means in the modern web environment I'm doomed to a low Klout score.  Oh well.

But this is too good not to pass on.  The lesson imparted by a long, fun slog through this website is among the most important a person can learn, whether one is deciding what to eat, which scientists to believe, what political stance makes the most sense, how to save and invest your money, or anything else.

Sourced from but not endorsed by http://www.tylervigen.com pursuant to a Creative Commons license.




Correlation does not equal causation.  Someone can show you two things which are very closely correlated and make you believe that one has something to do with the other.   Most often those things being related make sense in an intuitive sense.  Almost as often, the person presenting the correlation honestly believes that the relationship makes sense and provides evidence that one causes the other.  Occasionally there really is a cause to the parallel, whether that cause is one of the presented data sets or some other cause not shown.

When you see something like this, particularly in graph form, put your skeptic hat on.  That's especially important when the putative cause/effect relationship makes sense on that intuitive level.  Some of the people presenting that data have fallen for their intuition, others are counting on you to make that error.   The per capita cheese consumption in the US is actually more correlated with golf course revenue than with per capita consumption of mozzarella cheese.  One relationship makes no sense; the other makes so much sense that you aren't likely to question it.  But even that might still be misleading or even incorrect.

I promise you that, with the possible exception of Browns fans, there is no causative relationship between how much people spend going to spectator sports and suicides by hanging.  The per capita cheese consumption in the US is actually more correlated with golf course revenue than with per capita consumption of mozzarella cheese. 

Go have some fun on that website and find things which are closely associated but which have nothing to do with each other. 

No comments:

Post a Comment

For reasons I don't understand, third-party cookies must be enabled in order to comment.