Friday, October 17, 2008

Electoral vote predictions

The Princeton Election Consortium provides a meta-analysis of electoral vote predictions. This takes state polling data and applies some heavy-duty math/statistics and comes up with a daily data point reflecting the likely electoral outcome. As of today, this graph looks like this:

The gray band represents the 95% confidence level, and you'll note that the whole gray band has resided above the 270 mark for the lion's share of the race since Obama sewed up the Democratic nomination.

How accurate is their methodology? In 2004, they predicted that Bush would win with 279 electoral vote to Kerry's 259:


The actual result was 286-251: You'll notice that the 95% confidence band (in gray) indicated it could go either way.

5 comments:

Michael said...

http://hillbuzz.wordpress.com/2008/10/21/battleground-2008-poll-shows-statistical-tie-with-5-undecided/

Michael said...

Tuesday, October 21, 2008
A Three Point Play, Plus Pennsylvania And McCain Wins
Update: Latest poll for Nevada has it tied. Just got an email from Our Country Deserves Better taking some credit for the uptick because of a grassroots effort and ad buy. If you want to support them - click in here and take a look. They are moving on to Michigan - might want to make that PA folks.

Combining an RCP page and this item h/t Instapundit - makes McCain's current most logical play fairly clear. It means, first, he would have to win five states currently polling with him behind under three points: Florida -2, Ohio -2.8, Nevada -2.3, North Carolina -1.5, and Missouri -2.7.

If McCain can pull those out and then win Pennsylvania -11 (but said to be closer) - he wins the presidency with 273 EVs. He doesn't need Virginia, Colorado, New Mexico or anything else. Pretty simple, huh? On paper, anyway.

It's a steep climb, but sure beats looking at those national polls that have him down from 6 - 10 points.

Michael said...

What I think is happening, is this - the polls' headquarters were based deep in liberal territory, where the assumption was that Obama's candidacy would actually create a groundswell of pro-democrat voters unseen in the country since 1932. That McCain is more experienced with the key issues than Obama was ignored, that the historical significance of the debates shows that the effects appear several weeks later was also ignored. That the economy could be as reasonably blamed on the democrat-controled Congress as on the republican President was never considered. That character would be a salient factor in the decisions of voters was rejected out of hand.

http://wizbangblog.com/content/2008/10/21/gallup-and-new-coke.php

Stephanie said...

Don't need Nevada. Don't need Ohio or Florida either. It's all about Colorado, New Mexico and Virginia. Don't need any of the currently-undecided states.

Stephanie said...

But yes, national polls are ridiculous nonsense at this point and no one who knows anything is paying any attention to them. That's why we're instead looking at EV meta-analysis. And yes, McCain's big play is to get Pennsylvania. If he can do that, Obama's got to win someplace more difficult to make up for it.