This has been huge on the blogosphere: the Gates foundation may have blown a lot of money after it fell for a statistical fallacy. Here’s Megan McArdle riffing on Tyler Cowen who drew from Howard Wainer’s recent book*.
What the Gates Foundation did was steer a lot of money towards small schools because a bigger fraction of them performed far above average.
The fallacy is that because the schools’ student bodies are smaller, the variance of their scores will be larger. This almost guarantees that many of them will be far above the mean.
You can tell if this issue has come up by seeing if there are also many of them that are far below the mean, and Tyler has a plot of that. Of course, the education nomenklatura love to point out the huge number of clinkers among small schools to justify agglomerating into bigger fiefdoms.
*Back in the stone age, I enjoyed another one of Howard Wainer’s books so much for classroom examples that I photocopied the whole thing so that I could more quickly run off a page or two for students when the need arose.
The "Tyler Cowen" link takes you to the right place but the post is written by Alex Tabarrok.
Posted by: Alan | October 23, 2012 at 03:29 PM
Oops. Good catch. I stand corrected.
FWIW: I wrote this post quite a while ago, but sometimes drafts get lost. So, an oldie but a goodie.
Posted by: David Tufte | October 23, 2012 at 05:00 PM