Apparently I am the last person in the universe to discover this, but Nate Silver, a baseball researcher I respect quite a bit, has also taken an interest in polling and election prediction. Now, Nate made his name writing/researching for the vaunted Baseball Prospectus, a website that has helped made statistical analysis in baseball a (more) accepted thing. Next up, he has turned to polls and elections with Five Thirty Eight. It is a little hard to explain, but using some of the same statistical analysis technique he helped pioneer in baseball, he has tried to develop a system to help predict the outcome of elections (and the electoral college results) based on an aggregate of poll data. Now, I won’t divulge who they see as the clear leader in the 2008 Presidential election, but the amount of information on the website is ridiculous and worth your time.
aving taken some college classes on polling and how to interpret poll data, I have mostly come to the conclusion that they are next to meaningless. However, Nate’s 538 has given me a little more hope that real information can be gleaned from these data.