US polls.

[zotz]

Slashdot has just linked to http://www.electoral-vote.com/, which tracks US opinion polls as they come out and estimates the electoral college makeup that would result from that electoral result. It's very interesting to scroll back a week and watch Bush's conference bounce, which may now have played itself out if the latest poll is to be trusted (a big If there, of course).

The coloured map showing the collective opinion of each state is also interesting, although I suspect it won't tell actual USAnians anything they don't already know.

If you are a Democrat, you can stop crying in your beer; if you are a Republican, carefully try to pour the champagne back into the bottle. It is still very close.

Comments

[nils.livejournal.com]

The post-conference polls aren't in yet (though they should be later today). Will be interesting to see how the conference bounce affects the various states...

But it's certainly interesting that, despite national polls showing the candidates running more or less even, Kerry has actually been ahead in electoral votes for the last three months (up until about a week ago)...

[ciphergoth.livejournal.com]

The map is interesting but annoying. It would be more useful if it were one of those "distorted" maps that scale each state so its area is proportional to the number of EVs it returns. It would be more useful still to do away with the map and use a swingometer...

[thepaintedone.livejournal.com]

Hmm he headline count at the top makes it very close

Kerry 264 Bush 222

, but that aggregates the 'solid', 'weak' and 'barely' results together. If you ignore all the non-solid results its

Kerry 99 Bush 162

which is much more distinct. That would tend to suggest that Kerry has to do all the work, as his vote is much more marginal than Bush's. i.e a swing state which goes to Bush is a bonus, a swing state which goes to Kerry is his core vote.

The basic problem I see is that Kerry's two main virutes seem to be that he

a) Is a Democrat candidate.
b) Isn't Bush.

He seems to have little going for him on his own.

[ciphergoth.livejournal.com]

Those are pretty strong virtues though, especially that latter...

[zotz]

Your point about Kerry's support being softer is definitely true, but the Wall Street Journal's own article about the most recent poll points out that it implies a heavier win for Kerry than the headline you note - 307-231. So it's still very unclear, and not really in line with the other recent polls which showed an overall 10+ point lead for Bush. Time will tell, of course.

[nils.livejournal.com]

If you include California (which will almost certainly vote for Kerry, but where his current 8% lead doesn't quite qualify as 'solid Kerry'), it becomes Kerry 154 - Bush 162. If you go back a few days, the 'strong' states were Kerry 109 - Bush 142, which is probably closer to reality. Go back a bit further (to the last poll where Kerry was 10% ahead in California), and Kerry was ahead in strong states.

Or, if you ignore the 'barely' states but count solid and weak states, it's Kerry 211 - Bush 193.

Some of the states currently showing as solid in the most recent polls aren't really all that solid. Swing states like Ohio is currently showing as solid Bush, but was leaning towards Kerry a couple of weeks ago. On the other hand, New Mexico is currently showing as solid Kerry, but was tied a couple of weeks ago.

Basically, it's gonna be a close election no matter how you look at it... ;-)

[nils.livejournal.com]

the other recent polls which showed an overall 10+ point lead for Bush

Another pollster, Rasmussen, has criticized the recent Time and Newsweek polls as being biased towards the Republicans...

[zotz]

Apparently it's a bad weekend for doing polls because so many people are away from home.

[deliberateblank.livejournal.com]

I'm surprised to see CA is solid Kerry when they let Arnie in not so long ago. Are they that sick of him already?

[nils.livejournal.com]

States don't always vote consistently in different types of elections. For example, many of the southern states that are solidly Bush (and solidly Republican in presidential elections) have Democratic governors and Democratic Congressmen. New York, on the other hand, tends to vote Democratic in presidential elections, but Republican in gubernatorial elections.

There are a couple of other reasons why Arnie's governorship doesn't necessairly mean much for Bush - he ran on a fairly moderate platform (which is more than you can say for Bush), and the Democratic governor he replaced was pretty unpopular (well, he'd have to be for there to be a recall election in the first place). Oh, and he's married to a Kennedy. ;-)

[ciphergoth.livejournal.com]

Bush *did* run on a moderate platform. "A uniter, not a divider" and all that. He just went back on it all to a spectacular degree as soon as he got elected.

[ciphergoth.livejournal.com]

One remark intrigues me and I'd like to know more.

the sample contained a far greater percentage of Republicans than there really are

I'm guessing that that means the following: in addition to asking who they plan to vote for as President, they ask whether they are either a registered Democrat or registered Republican. The exact numbers of each are publically available, so this can be used to detect a systematic bias in the polling method. It would be *very* useful to know these figures for every poll when assessing what the poll means. Or to plot each poll as a point on a graph, where the one axis is "estimated lead of candidate" and the other is "registered voter measured bias" - would you get something approximating a straight line?

[nils.livejournal.com]

That's a fair point. I was talking about this year's election, though - Bush can hardly claim to be a moderate after four years of anything but moderate policies...