Yep. Plus, as I noted this morning, this isn't a primary, it's a caucus. Two totally different things.

Plus, no one seems to point out that it has a FOUR POINT statistical margin of error. That's not much of a sample.

Indeed, I think that it is 4.5

Then again, how accurate are the OTHER polls? Until we have some people really making irreversable decisions in voting booths, I wouldn't be too surprised to find any of the real candidates (i.e. not Kucinich or Sharpton) the actual preferred candidate.

Of course, the poll corresponds to trends that have been making themselves clear over the past couple of weeks. Dean and Gepardt down, and Kerry and Edwards up.

So it does say something different -- though it doesn't necessarily mean something different.

Sounds like we need to run those statistics again.

When Dean was at 3%, Deaniacs didn't believe in polls. When he was up by double-digits, they figured the primaries were a formaility. Now that Dean is shrinking (rather than imploding) they are going crazy. Democrats are getting a real or second look at these guys and Dean is not doing well.

First, a poll with 4.5% margin of error is still a poll (as compared to "not much of a poll"); it simply has about 550 respondents.

The bigger problem as identified above is that it is a caucus, which means public debate and relatively public voting. This creates a huge tendency for "informational cascades" (a.k.a., the mob effect). Hence, the Iowa caucus is not a matter of preferences, but of institutional organization -- who can get the troops out en masse to influence other caucus participants.