This place matters

This place matters

Thursday, July 22, 2010

Lack of pirates leads to global warming (details at 11)

I figure I've probably brought up correlation and causation before, and I will again - so I should define 'em, I suppose. If you already know this, sorry.
The Random House dictionary tells me a correlation is "a mutual relation of two or more things, parts, etc." In statistics, it's "the degree to which two or more attributes or measurements on the same group of elements show tendency to vary together." Comes from... well, comes from co and relate. I love science words. 
You'd think correlation would be a cousin of the word corollary, that thing in geometry that Sr. Mary Bitchmeyer made you prove even though some dead dude already proved it. Not so much, oddly enough. According to the Online Etymology Dictionary, corollary comes from the Latin corollarium for tip or gratuity, which supposedly comes from corolla, another way of saying corona, or crown. This seems absolutely absurd to me, so I'm going to say that correlation and corollary are cousins and the business with the crown is just something the Internets made up to make me look foolish.
But I digress, as usual.
OK, so correlation is different from causation, and this is where I was going when I got sidetracked by crowns. Causation, I don't need Random House to tell me, is the act of making something happen. Which would be a silly word to use at a party, but an important one in statistics.
Because, you see, just as the Internets are conspiring to trick me into believing crazy lies about word origins, the TV news and Readers Digest are conspiring to trick you into thinking correlation and causation are the same thing.
For instance, if Fox News were to claim that asking too many questions causes war, and then explains that the TV show Mythbusters premiered in 2003, just weeks before the Gulf War began, you'd... well, you'd be watching Fox News, and that's your first problem. That's post hoc ergo propter hoc logic, which is only one flavor of this sort of fallacy, but my typing fingers are tired, we'll get there later.
You know how they say cutting paper with fabric scissors will make the fabric scissors dull? My friend James insists that that's correlation, not causation as well. He claims that people who use fabric scissors to cut paper are more likely to, say, also use fabric scissors to cut wire. It makes some sense, but I'd like to see him try and test that on my Ginghers. 

No comments: