You may have noticed a new AP poll that says McCain is within 1 point of Obama. Seems like one hell of a jump, right? Well it turns out there's a huge difference in terms of how each pollster adjusts their number to reflect the "likely" outcome. Basically, some think that the actual voters this year will reflect voters four years ago -- when Republican voters were more enthused and united (the AP is one such group). So they adjust the numbers they're actually getting to fit the mold of the 2004 elections. Others think that the huge surge in Democrat enthusiasm and registration will result in a dramatic increase in Democratic voters from four years ago (all of these polls still show the race at around 8-point Obama lead).
To visit a real guru, you could do no better than Nate Silver, who's taken formidable baseball number-crunching and statistical skills to analyze the various polls (and provide some handy-dandy regression models that project the likely outcomes). Take a look at his article on the likely voter problem.
Thursday, October 23, 2008
Polling at Likely Voters
Labels:
elections,
Nate Silver,
politics,
polling
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment