Monday, April 19, 2010

Correlation Does Not Equal Causation

There was no dial tone on my business phone this morning.

Harrumph!

Clearly, the workmen swarming around outside, installing my lovely new air conditioning unit (brrrr!), had knocked out my telephone service. It was pretty obvious.

A couple weeks ago, I spent an entire afternoon trying to get my computer mouse to work. I had disabled the critter when I took my laptop out of town, and clearly I had not properly re-installed it when I returned. Easy diagnosis, right?

Wrong. In fact, epic fail.

In both cases, I committed the dreaded cum hoc ergo propter hoc logical fallacy, also known as false cause.

On closer inspection (and thankfully before I accused the A/C guys) the volume on my headset was turned down, making it impossible to hear the dial tone - which was working just fine, thank you very much.

My better half took one look at my complex mouse wranglings and said: "Your mouse is shot. Go get a new one." I did. And it worked great.

Correlation does not equal causation. Correlation does not equal causation. Having been raised on magical thinking, it's hard to get this useful principle through my head. But I am trying.

My favorite related story dates back to the early '90s, after a swarm of termites emerged from my Batchelder tile fireplace in Monrovia one sweltering afternoon. I promptly duct-taped the fireplace screen shut and called three exterminators for bids.

The first two guys gave me quotes for tens of thousands we did not have. The third, a woman, took one look at the situation, ripped off the duct tape and peered into the hearth. "Here's your problem," she said, poking at the half-burned logs left over from the previous winter. "This wood's full of termites!"

"You mean ... we don't have to tent the house!?" I asked.

"Naw. These old houses are made with hard, dry redwood. Termites hate it." She brought in some garbage bags, evacuated the buggy lumber and drove off.

I wanted to kiss her, I really did.

So: Correlation does not equal causation. Or: Always look into the fireplace before you hire an exterminator.

6 comments:

  1. Sooo true, my dear. In fact, my statistics instructor in grad school used to swear there was no causation, like, ever! Or "Ev-ah" as the Valley girl would say it. He came into class the first day with a sweatshirt that read: "I manipulated the data." His point: All data is manipulated. There are too many variables left unexamined to say the one you are examining caused it.

    ReplyDelete
  2. I like syllogisms when they're paper-trained.

    ReplyDelete
  3. Karen, your blog gets better every day. And that's saying something because it started out great.

    ReplyDelete
  4. Thanks for the feedback, folks. And Petrea, you're a lovely friend. And such good taste! ;-)

    ReplyDelete