Tuesday, February 12, 2013

Dean Chambers: Statisticians' Worst Nightmare

I have to give the man credit. I thought that after his predictions were revealed to be the laughing stock of the last elections, we wouldn't hear from him again. Instead, Dean Chambers continues to make statisticians weep with frustration as he gives statistics and math in general a bad name. Only after finding this article on Dispatches from the Culture Wars did I realize that Mr. Chambers didn't disappear from shame like any decent person would after completely ruining so many peoples' understanding of mathematics.
Ed Brayton said:

Gosh, it seems like you tried to make this exact same argument last year and ended up with enough egg on your face to make an omelette the size of Toledo.

If anything, Mr. Brayton is letting Mr. Chambers off easy. Essentially he didn't like the way that the numbers were turning out for his choice of president. Instead of accepting the numbers, as near as I can tell he decided to arbitrarily pick polls that he thought reflected reality and claim that the others were skewed. He would take data from a poll that he liked and their internal numbers to modify a different poll. The results was that nothing he produced was even remotely accurate.
Mr. Chambers seemed to try to use mathematical slight of hand instead of real statistics to force reality into his liking. Fortunately for us reality doesn't work that way. So Mr. Chambers, go on unskewing your polls. Reality isn't going to care any more about your numbers now that it did last year.

No comments:

Post a Comment