I think the race is genuinely this tight, with any of the three current leading candidates -- and perhaps also even Bill Richardson (or, heck, maybe even Al Gore, if he decided to get in late in the game) -- being in a position to turn a little bit of late momentum in the state into a victory in the Democratic caucuses.

Late momentum?  Do you own a calendar?  It's August not December 2007. This is a wide open race between HRC, Edwards, Obama and Richardson. In 2004, almost half of the Iowa caucus voters didn't select their candidate until less than a month prior to the election.  

And no candidate has ever gone wire to wire in first place in the polls through election day as the winner in a contested Democratic race in Iowa. John Edwards isn't going to break that streak in 2008.

That poll was taken before the last debate.

Still I think it is totally significant, and I plan to further riff on what Jon is writing here. At least incorporate it.

The numbers here are totally interesting to me.
Basically, two sets.

Generically speaking, the first set is a two way between Edwards and Clinton

The second, between Obama, and Richardson.
IMHO this is due to Richardson (who gained 3 points) and Clinton (!) gaining ground
while Obama losing it, and Edwards looking like he's holding on.

Totally right on with the analysis. Definitely a tighter race. Its almost like Obama and Richardson will slug it out on the progressive level, and Edwards (no slouch) and Clinton will slug it out for the populist vote.

Better still, Edwards can roll up his sleeves and really get HRC on the go.  I like it.

Man. .. We'd better take all the sharp implements out of the room when these four face off again.

   Hillary is doing so much better in states where no one is paying any attention! The mainstream media and Karl rove are having a much greater effect wiht least interested voters and least interested states. In Iowa just like the focus groups said after the debate when people pay attention they switch to Obama.

It looks like this race is now indeed very close in Iowa. It's going to a battle until the very end, which was always expected. Right now it looks like Edwards is losing the most, as far as his past support goes, however it will be interesting to see if he can pick up more liberal voters from Hillary in the future, which could be bigger in this particular caucus, considering there will be both Republican and Democratic caucuses this year.

At this rate, it's going to be totally unpredictable. Who is going to fade at the very end?  No telling, but whomever does, and slips below the 15% threshold across Iowa precincts, then look to where their supports head for viability on the night of the caucus. But right now, Edwards was right when he called this a dog-fight. Richardson is clearly getting close to being a top-tier candidate.

Can anyone run away with it in Iowa?  Very doubtful.

Can someone fade?  More likely, and that will be the key. Heck, if this is how it is going into the caucus night, it might be Joe Biden's 5% column that decides (if they go block across the state) the winner.

In SV's polls: from 29% to 26 to 23.

Based on a sample of 600, the decline from 29% in May to 23% in August is a statistically significant change.

I hate to say that, because I'd like to see Edwards win, but what's true is true.

Ya know....less than 50% of the eligible voters vote in the General Election....as for a caucus?
Here in Colorado we have a caucus....
In my caucus area in Denver... there were 2,800 registered Dems in 2004....36 showed up and voted...

Polls in caucus states are worthless unless they are limited to those that attended the caucus in 2004...

It is much harder, inherantly, to poll for a caucus than for an election or primary.  Caucuses are different.  It is harder to predict who will vote in a caucus vs, a primary or election.  There is less voting history with caucuses.  There is no absentee voting with caucuses.  Timing is inflexible.  Blah, blah...polls are fun, they tell us something esp. about trends but with caucuses in particular don't put too much stock in 'em.

No way!!!

Jerome's been telling me that Hillary is inevitable.

I think this poll indicates that the Zogby poll over the same time period (showing Clinton with a clear lead) was an outlier. Zogby's polls are always so off-the-wall. I don't understand why people ever take his numbers seriously. It will be very interesting to see if Obama will get any bump from the last debate. If he doesn't, it may be cause for concern.

Clinton and Richardson are in good shape in Iowa, Edwards is trying to hold on to any kind of lead even a tiny one, and Obama has an uphill fight to compete with the other three.

that the various polls we see ARE consistent with the overall average plus statistical noise.  The race is that close.  That's why we variously see Edwards or Clinton leading, and occasionally even Obama.  

I think everybody agrees that Iowa is a tie right now.  That means the race will come down to organization.  Edwards and Obama, from most media accounts, are winning that battle right now.  Edwards has been there for years building a strong base.  And Obama has twice as many offices as any other candidate, and is outpacing the others in number of volunteers on a regular basis.  Edwards and Obama also well outpace Hillary (I'm just guessing - but am near certain it is true) in the number of Iowans who have donated to their campaigns.

When it comes elections, there's only one poll that counts...

Richardson is catching up to Hillary...

none of this matters until late December, early January...