Data Freaks

I link to FiveThirtyEight.com all the time, but I thought I would give a special shout out today. This site is essential if you are at all interested in the data that goes into campaign decision making. Their model is based on probabilistic outcomes generated from running multiple election scenarios, using aggregates of national and state polling data as inputs. They explain their methodologies in great detail at the site, and they are extremely transparent about how and when they tweak weightings, etc.

The key insight that drives a lot of the writing on FiveThirtyEight is that elections are not won by planning to win the most electoral votes; they are won by maximizing the likelihood of getting 270 electoral votes. For this reason, they focus not just on "swing states," but rather tipping point states ... whereas swing states are the states that could "go either way" on November 4, the "tipping point" states are "among the closest states –- taken alone or in combination –- that would give the losing candidate at least 270 electoral votes if transferred to him from the winner’s column, with no wasted electoral votes."

In any event, the site is an orgy of pretty data, so be sure to check it frequently. Also, whenever someone comes to you, freaking out, like, "I just saw the Camden County Community College poll, and they have McCain up by 38 points in New Jersey," go to FiveThirtyEight and listen to the man in the picture below:


(Hat tip to Ta-Nehisi on the picture. He's awesome.)