Daily Electoral Predictions

Robin Berjon writes " If you are both a political junkie and a statistics addict, I highly recommend Electoral Vote Predictor 2004, a site that gathers a collates polls taken in individual US states according to a well-documented method and uses that to generate a daily map and victory prediction, alongside a short and insightful analysis of the current trend. The site also includes a wealth of information for past maps, detailed tables, tools, links, the Senate elections, and much more. It also has a convenient RSS feed so you can get your daily fix."

Comments

Re:We need popular votes to count!

[Marxist Hacker 42]

Especially since if we could just make it secure, computer tabulation of votes during the month long election could give us a real time look at the election (most people don't realize this, but due to universal mail in voting my home state of Oregon can have ballots turned in, but not counted, as early as October 6th if the mail delivery is on time).

Another (Similar) Site

[Dr. Bent]

Nice link.

I also like Rasmussen Reports.

I'd say the polling methods have a basic flaw

[Skyshadow]

My guess is that the polling folks will spend the week after the election going around the talk-show circuit explaining how their polls managed to be so wrong.

Pretty much all the polls we see are of "likely voters", a group which is made up mostly (or entirely) on the people who voted in the last election. This may be a useful measure in the average election, but not this one.

The 2004 election will have a much, much higher turnout than 2000. In 2000, it didn't seem to matter a whole lot who got elected. In 2004, most everyone knows someone who has lost their job and/or knows someone in Iraq. A lot of people are still genuinely angry about Florida's lack of concern for voting rights or even following their own laws. At the same time, Bush hasn't given his more casual supporters a reason to come vote for him -- the best they've managed to do is spread a bunch of half-truth (or outright lying) reasons why they *shouldn't* vote for Kerry.

More new voters have already registered for this election than any other since 1992. That should tell you something, and logically it doesn't seem like good news for Bush if these polls are showing a close race...