pollery wonkery

[mmcirvin]

For what it's worth, the state polls seem to agree that the US presidential race is teetering on an incredible knife-edge, with Kerry possibly slightly ahead unless you are the Washington Post, and if the election were held today and followed the polls, the whole thing might well rest 2000-style on the usual residue of random luck and dirty tricks, only this time in Florida and Ohio. Not that the results ever track the polls quite that closely (Democratic optimists say that undecideds traditionally break late toward the challenger in an election like this one, but it remains to be seen how typical this election is). It's another nail-biter, folks, and only time will tell how big.

But Bush's big public-goodwill surge, that began while he had a spending-cap advantage over Kerry and continued through his remarkably large convention bounce, seems to be over. I honestly didn't expect that.

Comments

[mmcirvin.livejournal.com]

...Though, like everything, it looks inevitable in hindsight. My wild-assed guesses in the summer were that nobody would get a significant convention bounce, Bush would pull ahead during the month-long McCain-Feingold gap through saturation advertising, then Kerry might come back.

What happened instead was that Kerry, who had already been leading, got a small bounce that lasted maybe a week into the McCain-Feingold gap; Bush pulled ahead in the gap through saturation advertising, then parlayed that into a rather large convention bounce (large horse-race-wise, if modest approval-wise) that dissipated following the debates.

[arsonnick.livejournal.com]

I wonder how many of these electoral predictors are taking into account the possibility of Amendment 36 passing in Colorado, which would, undoubtedly split our electoral votes (knocks on wood).

[arsonnick.livejournal.com]

And aren't there any other states doing this as well?

[ronebofh.livejournal.com]

Nebraska and Maine already do it, but i'm not sure how it works. They didn't split their EVs in 2000.

[mmcirvin.livejournal.com]

In Nebraska and Maine it works by congressional district (http://uselectionatlas.org/INFORMATION/INFORMATION/methods.php), with two at-large electors chosen by winner-take-all and one per district. In practice, their votes have never actually been split.

I think the Colorado proposal's a little different, to split the votes proportionally by popular fraction.

I believe that the spacerad.com page is currently treating Colorado as winner-take-all but has some sort of disclaimer about the vote-splitting referendum.