Getting excited over a 2% lead in an Iowa poll is laughable

[Herman Munster]

It's a 3 way tie. Edwards, Obama, and Clinton can all win it. Who will win will depend on who has the best campaign machine to get their people to caucus on a cold night in January. Nothing more. Nothing less. It's all about turnout baby. We won't know who had the best machine until January 3rd.

[gasperc]

1. Thanks, Herman, I thought I was the only crazy one

but with that said, Edwards will win :)

[Perky]

2. It is not the 2 points.......it is the trend line that is exciting

[Herman Munster]

3. the trend is meaningless

don't you understand that? The pollsters are trying to poll all Iowa democrats when only 5% will turn out and they are making assumptions based on who will turnout when it is anyone's guess.

On the day of caucus, everyone thought Dean would run away with it. It looks like the momentum in the polls is with Obama now but it won't mean anything on caucus day if Hillary or Edwards gets more of their people to the precincts because they have better voter databases or more buses or more experienced captains that can persuade or bully their neighbors in the precinct halls.

[DemocratSinceBirth]

4. I Agree With Most Of What You Say

But the final IA polls showed Dean losing the state...

If the race is as close as portrayed the candidate with the best organiztion will win...

[Perky]

15. To some degree you are right

But the nunbers suggest Hillary's support is shallowing


Candidate . 6-Dec 7-Jan 7-Feb 7-Mar 7-Apr 7-May 7-Jun 7-Jul 7-Aug 7-Sep 7-Oct 14-Nov 29-Nov 52 Day swing 15 day Swing

Biden_____ . 2% 2% 2% 2% 6% 3% 2% 3% 1% 3% 5% 5% 8% 3% 3% .
Clark_____ . 1% 2% 1% 2% - 1% 1% - ni ni ni .
Clinton___ . 31% 35% 31% 34% 23% 31% 32% 30% 28% 30% 32% 27% 25% -7% -2% .
Dodd______ . 2% 1% 1% 1% 2% 2% 2% 1% 1% 1% 2% 3% 3% 1% 0% .
Edwards___ . 20% 18% 27% 33% 27% 25% 29% 21% 20% 19% 15% 20% 23% 8% 3% .
Gravel____ . 1% - - - - 1% 1% - - - - - - .
Kerry_____ . 2% ni ni ni ni ni ni ni ni ni ni .
Kucinich__ . 5% 2% 1% 1% 2% 4% 1% 2% 3% 1% 1% 2% 2% 1% 0% .
Obama____ . 10% 14% 23% 16% 19% 11% 13% 15% 23% 24% 22% 21% 27% 5% 6% .
Richardson . 1% 1% 1% 1% 5% 8% 5% 13% 13% 10% 7% 12% 4% -3% -8% .
Vilsak___ . 17% 12% ni ni ni ni ni ni ni ni ni .
Undecided_ . 8% 13% 14% 10% 16% 14% 14% 15% 11% 13% 16% 10% 8% -8% -2% .
Total_____ . 100% 100% 101% 100% 100% 100% 100% 100% 100% 101% 100% 100% 100% 0% 0% .


Suggesting that this means nothing is ridiculous.

If HRC was up by 6 points would your view be preciesely the same?

[wyldwolf]

5. especially when it the lead seems to change every day and in every poll

[Herman Munster]

7. anyone objective knows it's a 3 way tie

any differences are based entirely on the "likely caucus" screen each pollster uses. Will more women caucus or less? Will more younger people caucus or older? Will more new people caucus or old people that caucus before?

So really depending on each screen, you can come up with scenarios where any of the 3 win. It all depends on turnout and organization.

[Adelante]

6. Yeah, we know all that

Still, seeing Obama in the yellow on the RCP averages is kind of cool.

Even if it's by the slenderest of hairs, it's something nice for his supporters.

Really, it shouldn't hurt you so much.

Look at that sweet little smooch on the graph. :)

http://www.realclearpolitics.com/epolls/2008/president/iowa-primary.html

[Kurt_and_Hunter]

8. It's better than a 2% deficit

[dmallind]

9. Well (and this applies to ALL candidates) it's really not

Given the MOEs we've seen no movement at all lately and not all that much for quite some time. You can't trend anything if the numbers are in the MOE, and 2% down is the same as 2% up. If it isn't, then the whole idea of probability samples and statistical inference means diddly squat from the start.

[Kurt_and_Hunter]

16. The margin of error is not a range of equal probability. It is always better to be ahead than behind...

If a candidate polls 30% with a margin of error of 5%, 30% is more likely than 27.5% or 32.5%, and considerably more likely than 25% or 35%.

The margin of error means that in that case there is a 95% chance the true total falls in a range defined by the MoE, and centered on 30% as the likeliest single result.

Imagine a little bell curve centered on the headline number, with a certain part of the curve above a 95% level of certainty.

And results just outside the margin of error are only a bit more unlikely than results just within the margin of error. It's all a smooth curve. The polling industry standard of 95% accuaracy is just a number somebody settled on... they could have settled on 94% certainty or 96% certainty as the standard and it wouldn't change the curve one bit... just where you chose to draw a line through the curve.

(and, as with all things, if I'm wrong I'm wrong)

[dmallind]

17. I don't think you are wrong, but...

...you are I think at least implying that polling is more precise than it is. You would be spot on if you were dealing with an objective, discrete measurement. But opinion polling is subjective and very fungible. We can't treat it like a mathematically precise measure.

[Kurt_and_Hunter]

18. Yes, sampling is very imprecise

I agree that polling is imprecise because it is VERY hard to sample people randomly, though the same techniques are wonderfully precise when applied to predicting the number or red marbles in a well-mixed jar containing exactly 10,000 marbles of four different colors.

The biggest wild card in Iowa polling is that nobody has the same model of likely caucus-goers, so it's all guess-work in a GIGO way.

We certainly agree on the key point that we don't know who leads in Iowa.

[dmallind]

21. See that's the difference. Marbles ARE either blue or red

but apart from a minority of rock solid partisans, most people aren't necessarily Clinton or Obama voters for sure. They may change their minds. They may not vote at all. That's why it works perfectly on quincunx boards, very close to that with repeatable mechaincal processes, but not at all perfectly with subjective opinions. Now sure would I rather be 2 up than 2 down, as a subjective boost? Absolutely! But I wouldn't bet so much as a left finger let alone a right arm that it wasn't really the other way round even at the time of the poll, let alone in the election.

Again we don't disagree much at all - just an interesting way of looking at poll data IMO.

[1corona4u]

10. Hey now...

Joe moved up 3 points in the last 2 months in Iowa...and I am excited, LOL...

[DrFunkenstein]

11. Not True. Looking at the 2004 Polls, I Knew Dean's Lead Was Softening

While Kerry was solidifying his support.

Clinton recently brought on an extra 100 people who had to be trained, compared to the other campaigns who have been wooing the notoriously fickle Iowans, who literally demand the attention. Clinton's support was never that solid with so many refusing to get pinned down, yet she maintained high negatives with people who consider her unacceptable.

The national gloss is coming off of Clinton as they pay more attention, and Iowans are like canaries in a coal mine because they pay attention a little earlier. Furthermore, little of Clinton's support comes from past Caucus goers. I'd say the trend line is pretty exciting, just as Clinton supporters find it a little unnerving.

[Jack Bone]

12. I wonder if there's another DK-Edwards deal in the works?

I wonder if Dennis swung his support to John again, like he did in '04. If that might just be enough to push John on to victory..Just wondering?

[Adelante]

13. Doubtful

Dennis has been banging on Edwards pretty hard this time. Last time they were buds.

[Rydz777]

14. Right. All polls of the caucus system are virtuallty worthless. nt

[suston96]

19. The trend line....

....is very important. But Hillary lost the entire effect of the debate on the 15th of November because of the lack of polling during the Thanksgiving recess taken by pollsters.

Trends are important, but one must remember that trends can go both ways.

[AtomicKitten]

20. It just seems that way when when your opponent is behind.

wink, wink

[asdjrocky]

22. Yes indeed, it is a three way tie.

Which is absolutely amazing when you consider how much has been spent by Obama and Clinton.

Personally, I like a candidate who can get the most bang for his bucks.

Obama 3 mil+ Clinton 1.4 mil+ Edwards 0.6 mil

Kind of seems unfair don't it? Imagine, spending all that money, and only being tied with a guy that's spent less than half. That is a President I want.

[BeyondGeography]

23. Obama has the best machine in Iowa

Thanks for the reassurance.

[Forkboy]

24. Oh, NOW they're laughable.

And you're predictable. :rofl: