Polls = Crack

Polls. I wish I could smoke 'em. Or fuck 'em. They're like an open bar. I've been sipping from their unending fonts too much as of late. Here's a prediction of Nov. 2nd from Princeton:

Below is a meta-analysis directed at the question of who will win the Electoral College. Meta-analysis provides more objectivity and precision than looking at one or a few polls, and in the case of election prediction gives a more accurate current snapshot. The calculations are based on all available recent state polls, which are used to estimate the probability of a Bush/Kerry win, state by state. These are then used to go through all possible combinations of battleground state results. The effects of undecided voters, turnout, and polling bias are calculated using the bias analysis. Here are the full methods.

You can find the explanation here. Princeton's prediction is a meta-analysis that takes into account all the state polls plus the various "likelihood" scenarios, resulting in a probability map of how things might play out in thirty-six hours rather than a "snapshot" of this moment's likely electorate.

Posted by jason at November 1, 2024 01:51 PM