IOWA Polls Completely Off

[srpantalonas]

polling is supposed to be such a science. But just look at the DU board--we have so many people saying they are switching from one to another. I imagine folks in Iowa take it seriously enough they don't really decided until they are at the caucus meeting.

Anything can happen. Something certainly will. I wish the pundits would shut up and let it happen, because they just don't know.

[Renew Deal]

1. What the polls don't account for

is the caucusing. There isn't an indication who the second choice is.

[BeatleBoot]

5. I agree absolutely

This is a caucus - not a primary election.

Hopefully, the sampling population is representative of the caucus go-ers, otherwise its all useless data.

[iowapeacechief]

10. Iowa 2nd-place poll?

Does anybody here know of any? That's what's needed to start to predict the caucuses with accuracy.

[Bolo Boffin]

20. My guess, though it's unscientific

I looked at this poll from the Des Moines Register:

http://miva.dmregister.com/miva/cgi-bin/miva?extras/iowapoll/poll.mv+file=prez0401

And based on it, I'm going to predict Kerry grabs Iowa in a big, big way.

The reason? There are only four candidates there who will reliably score the 15% needed to keep their caucus voters: Kerry, Gephardt, Dean, and Edwards.

However, Gephardt's support is at 18%. All the others are at 20% or above. That makes me think that there will be a signifigant number of caucuses where Gephardt won't make the required 15%. He'll still have a good block of voters, and they will then have to decide on their number two.

Out of Kerry, Dean, and Edwards, which of these three is most like Gephardt? Kerry. I believe he'd be the candidate of second choice for most Gephardt voters. And that will be enough to turbo him into an Iowa win.

All speculations are the property of me, and I'm a noted flop at prophecies.

[arewethereyet]

22. I'd have to think that the newspaper's poll would account for that

its not like they are unaware of the concept.

[Name removed]

2. Deleted message

Message removed by moderator. Click here to review the message board rules.

[wtmusic]

3. The suspense is too much

for most Americans.

The loss of control goes against our grain, so even when we can't figure it out, we pretend like we can.

Kinda like religion.

[MoonRiver]

19. I've felt out of control since Selection 2000.

Iowa seems tame compared to that nightmare.

[isbister]

4. Do you think

the polls are what is influencing people to switch? I've seen a few switchers that have posted and do not remember too many that said "My candidate lost some points in the polls so I'm switching".

It seems to me that people who switch before one vote has been cast 1) May not have been 100% solid in the first place and 2) have given their decision some considerable thought.

Hey, a lot of us will be switching at some point because only one candidate can win and bush must go home for the good of the country.

[Nederland]

6. Frank Luntz

pointed out the large number of people that had switched candidates in the last few day and thought it was a bad sign for Dean.

Yes, I know who Frank Luntz is. Question is, is he correct?

[aquart]

7. Isn't the rule that ...

And I could be wrong, but that in each caucus, the group with less than 15 percent has to choose a different candidate?

And then they just talk and talk and talk and talk. Democracy being taken seriously.

But if anyone asks you about it, you just say that the Democrats have fielded such a wonderful, intelligent, capable, attractive slate of candidates that the choice is really really really tough, because they're all so damned fine.

Remember: whoever wins the candidacy, we support him. We support the hell out of him. We have a nation to save.

[LoneStarDem]

8. Yea

As far as I know, thats how it works, starting at the local level, then district, then regional, then state (I could be wrong in that grouping, though). Each time there is a 15% cutoff. But, the thing is, with four strong candidates in Iowa, it would be possible for them each to have strong support AND meet the cutoff, splitting the national delegates (and the momentum) four ways. So the question really is, how homogeneous is the support across various subsets of the state and how will that affect the eventual state representation? That and who the seconds will be for under performing candidate reps at various levels.

[SadEagle]

11. Almost

There are also some precincts which only select only one candidate (and some w/2, 3, etc.), so to get a delegate from there a candidate must eventually get to 50% of support!

[janx]

9. Luntz is a current Republican fool who is hired by MSNBC.

How or why should he be believed?

[Nederland]

13. As I said

I know who Luntz is.

Claiming that because he is a Republican he must always be wrong about everything is not a very convincing argument.

[Cuban_Liberal]

12. I think they're wrong.

I know tons of Dean supporters and lots of Kerry supporters, and I still see a fairly comfortable Dean win with a strong second-place finish by Kerry and a virtual Gephardt/Edwards tie for 3rd place.

Dean has an impressive organization and almost fanatically-loyal supporters, and most of the polls don't account for many people who are expected to vote for him--- first-time caucus goers.

I still say Dean 35-38%, Kerry 22-25% and Gephart and Edwards each in the barely +15% range, which would be great for Edwards and disastrous for Gephardt.

[Kahuna]

14. You'd be surprised at how many people wait until the last minute..

to make up their mind. In a race with 7 candidates it's not surprising that the numbers are so fluid. Also remember, the two "frontrunners," Gep and Dean are for raising taxes. And yes, reversing the bush tax cuts is a tax increase.

[Padraig18]

15. George Will just made a good observation

On ABC's "This Week" while he and George Stephanopolous were interviewing John Kerry:

WILL: "Johnson County, IA, home to the University of Iowa reports a 300% increase in new registrations over the same period 4 years ago; it's entirely possible that the pollsters have no idea (Will's emphasis) what's going on, isn't it, Senator?"

KERRY: "Yes."

[UTUSN]

17. They're Jacking Us Around with These Polls

The media whores HAVE to have excitement, a tight race, for their own idiot purposes. They will BURY the results with Shrub within hours. They want an UPSET, or an upset that NEVER HAPPENED, or whatever.

[Padraig18]

18. True.

I watched Bob Schieffer on "Face the Nation", and he also agreed that the pollsters have no clue what's going on.

[janx]

27. Of course they have no idea. You can't predict what's going

to happen in the caucuses.

I'm laughing at the television every time I hear someone try!

[spooky3]

16. I don't want them to shut up. I want the Dems to be on

the front burner. We need the message to get out.

[chascarrillo]

21. Polling is a science

If you'll look at the recent polls, they also ask your question: whether it's still possible that they'll switch their votes. The number, if I recall correctly, is around 47%.

[creativelcro]

23. all that is based on the assumption that people do what they say.

or that they know how they are gonna behave. That is a strong assumption. People have theories about themselves and about why they do this and that. But those theories are often wrong and people find themselves behaving in different ways. They rationalize that, but the point is that they could not predict their own behavior in advance.

[DjTj]

24. It's not just a right/wrong question

People may not know exactly what the chances are of changing their minds, but chances are that some people will answer more conservatively and some will answer less conservatively. These two things will generally cancel each other out and that's how we get the +- error numbers.

Polls don't always give us the error numbers, but I'd guess it would be in the +-5% range just like the rest of the poll.

The number isn't necessarily right or wrong, but a poll can get us close (within 10%) of what the real number is.

[creativelcro]

25. I disagree.... Your assumption that "things cancel out"

is convenient. But it's unproven.

[chascarrillo]

28. Margin of error: what it is

People have been throwing around MOE a lot, so I think it's prudent (and pedantiic, but I'm good at that) to explain what MOE is.

The margin of error is the spread within which the poll accurately reflects reality within two standard deviations - that is, 95%. This number is a direct byproduct of the sample size of the poll.

Here's one good resource on polls: http://www.ncpp.org/qajsa.htm#11

[chascarrillo]

26. Entirely true

That's completely true. All I was saying was that the methodology and results are correct (assuming that it's a reputable organization).