Q&A Friday #112: What Do You Think The Effect Of Statistical Analysis Will Be On Punditry In Upcoming Elections?
Question: Oooh! What do you think the effect of statistical analysis – a la Sam Wang at PEC or that Nate guy that everybody loves – on punditry in upcoming elections? — MightyMooMan
Nate Silver’s % chance of o winning on his blog is actually the chance he’ll be called a genius if Obama wins & an idiot if he loses.
— John Hawkins (@johnhawkinsrwn) November 2, 2012
For all of the talk about Nate Silver’s model and “statistical analysis,” all he essentially did was look at the state polls and applied the Monte Carlo method to it. In other words, Nate Silver was right because the state polls correctly predicted the winners of the races. Had Gallup and Rasmussen turned out to be the ones that were right while the other pollsters were overestimating the support Democrats were drawing, Silver would have been wrong and everyone would be talking about what an idiot he is today.
As to what effect it will have, it just means that everyone will probably take the state polling data at face value in 2014. As long as it keeps working, everyone will essentially come up with the same results that Nate Silver does, without using any “statistical analysis” of consequence, just by looking at the polls. If and when the state polls are wrong again, Nate Silver and everyone else will turn out to be wrong, too.
Tomorrow will be Q&A Friday #104 at RWN. So, if there’s a subject you’ve been wanting me to tackle or
Question: “How would you answer Kristof’s “Wrong Side of History” piece in the 11/18 NYT? If Medicare, Social Security and
Question: “If the republicans take the House next year, do you think that Obama will moderate and work with republicans,