Well, this battle in Iowa is truly going down to the wire.
Today Zogby has a poll out with Dean 25%, Gephardt 23%, Kerry 14%, Edwards 13%. And Zogby says Edwards was picking up steam through the three days of interviews.
Those numbers are somewhat countered by two other polls out in the last two days which have the Dean/Gephardt spread at 7% (LA Times) and 5% (Quad-City Times/KWQC). But clearly it's pretty tight.
At various points over the three-plus years I've run this site I've commented on the quality of Zogby's polls (here's one of the first things I ever wrote about him, an article in Slate from almost four years ago). Zogby's polls do often pick up on trends well in advance of their appearance in other polls. But almost as frequently he seems to pick up on trends that turn out not to exist. So to the extent that you can personify methodology I see him as ingenious but erratic.
That's I suppose just a roundabout way of saying that I don't know quite what to make of his very tight margin between the two top contenders.
More generally, however, the closeness of these margins and the broader dynamics of this race, I think, tell us that we don't know what's going to happen on Caucus eve.
The Caucus system works by a sort of snowbound version of the uncertainty principle. The particulars are so detailed that I'm sure to get some part of it wrong. But basically how the caucuses work is that everyone shows up and they divide into groups based on candidate preference. But if your candidate has less than 15% of the attendees then your guy (or gal) is out.
(Presumably, at the beginning of the evening someone gets out an envelope, counts who's there and does some quick math to determine how many people get you over 15%. I'd be ruled out for that job.)
Once your candidate is out you have to pick another.
Now, the numbers we're seeing are statewide. And the demographic gap between, say, the Gephardt and Dean voters is great enough that in particular caucus locations the spread is apt to be very different. However, you don't have to look too long at the numbers to see that there are some candidates with not insubstantial support that are going to get knocked out on the first round at many locations.
To put it succinctly, in many caucuses, the issue is going to be less whether Gephardt and Dean are separated by 2% or 6% as who the Kerry and Edwards supporters go to on the second round.
Before you flame me with your emails, I'm not trying to prejudge Kerry's and Edwards' chances. But you can see the level of uncertainty that plays into the calculus.
On the face of it, it would seem this volatility would work against Dean since the race has been polarized between Dean as the outsider vs. anti-Dean Washington candidates. If you're for Kerry, who has made opposition to Dean an increasing focus of his campaign, do you switch to Dean or to Gephardt, if Kerry falls under 15% if your caucus? The latter seems more logical to me. But so many factors must play in to this that I'm not sure 'logic', especially from a thousand miles away, gets you very close to the truth.