Thursday, 12 April 2012

Weirdest use of statistics for the week

Sam Bowman at The Pin Factory Blog gives this example of a dodgy looking (mis)use of statistics:
The overall impression the article gives is how shallow Nudging is as an idea. One example:
A study into the teaching of technical drawing in French schools found that if the subject was called “geometry” boys did better, but if it was called “drawing” girls did equally well or better. Teachers are now being trained to use the appropriate term.
Well, if you say so. I suppose this might be true, but it could just be another of those “correlation equals causation” fallacies that bedevil the social sciences. Statistically, all manner of associations may appear in data surveys, but they don’t necessarily mean that one thing causes the other.
I have to that I have trouble believing that just changing the name of the course really changes the success of girls relative to boys.

No comments: