Think like a statistician

This is probably the coolest example of statistical thinking I've read about:

http://www.johndcook.com/blog/2008/01/21/selection-bias-and-bombers/

Selection bias is a huge problem in statistics and comes up all the time with clients. I've had several projects tank because we determined the data was non-random. Wald's solution to the non-random data problem is just brilliant.

h.t. MR