Wednesday, November 25, 2009

Flimflam inoculation:

Many of the podcasts I listen to spend time going over some of the common errors in logic that confuse people into believing something silly. I'd love to put together a talk some day on how to detect these logical fallacies, and go on tour spreading the skeptical gospel.

Here is one installment.

Correlation is not Causation

This has a fancy name: "post hoc ergo propter hoc", meaning literally "after this, therefore because of this". OK, I've had my Latin fix for the day.

Consider this common example. Evidence suggests that children with larger feet are better readers. Shocking, isn't it?! Well, not really, when you consider that kids in grade 1 tend to have smaller feet than kids in grade 5. But it's human nature to interpret the statement as meaning either
  1. shoe size somehow affects how well kids read, or
  2. reading ability affects how quickly feet grow.
The fact is that neither is true. The two things just happen to be influenced by a common factor, age! The only thing the data tells us is that reading and shoe size are correlated... it doesn't tell us which caused which. In this case, neither.

Another infuriating example: When someone says, "I went to the chiropractor, and my back felt better within a week", it does not prove that the chiropractor caused their back to get better.

No comments:

Post a Comment