How Zogby final polling compared in Iowa to actual caucus results...

[flpoljunkie]

Zogby final polling on 1/19 in Iowa

Kerry 25%
Dean 22%
Edwards 21%
Gephardt 18%

Iowa Caucus Results

Kerry 38%
Edwards 32%
Dean 18%
Gephardt 11%

Draw your own conclusions.

[jenk]

1. The Des Moines register nailed it

I'd like to find some local NH newspaper polls.

[Nicholas_J]

18. Unfortunately the closest is the Boston Globe

Some By KRC of Newton Mass.

And WMUR, which is dome by thr University of New Hampshire that has:

Kerry at 37, Dean at 22.

http://www.unh.edu/ipssr/survey-center/track2004.html

Poll is in PDF format.

[WI_DEM]

2. I guess any poll that has Dean closing the gap

is discredited. I'm sure you can do it with the SUSA poll too.

[lancdem]

5. The polls don't matter

The voters will decide tomorrow.

[Nicholas_J]

19. Insomnia last night

Listening to late news with people from New Hampsire Public Radio, and they stated that while polls are actually not very important in Iowa, they have extreme influence in New Hampshire.

[kstewart33]

6. Not at all

I'm a Kerry supporter, but I and others recognize that Zogby has made some great calls. It's just that he has also really missed the boat at times, especially in state-level polling. For example, here in Colorado in the 2002 election, Zogby had Dem Strickland comfortably winning the Senate race by a large margin. He lost to Allard by 4 points. It was a REAL bummer. I lost some faith in Zogby after that. But he still merits a hard look.

So who knows?

[Nicholas_J]

20. Zogby has a reputation for calling one election right

That was the 2000 presidential election. His calls are notoriously bad, as just before Christmas, he stated that he could not forsee any possible way that Kerry's campaign could come back and declared that Dean would win the Iowa Caucuses. His local polls have been rather off the money. Ten out of 12 polls give Kerry big double digit leads over Dean, and actually reveal far more about their methodology. like ARG adn University of New Hampsire tell you how many independents were included in their polls. ANd these polls have Kerry much farthest in the lead, and they also give a better picture of where the much valued independent voter is actually doing.

Zogby has only been around since 1999, and the regional pollsters in that are either in New England or near it have far better records, like ARG, KRC, Suffolk, and so on.

[ACPS65]

3. Stop making sense

.

[TLM]

4. Caucuses are hard to poll...

Because one can't really account for the switch overs that take place when DC insider candidates throw their support to each other to shut out the grassroots.

Primaries are much easier to poll.

[lancdem]

9. Although N.H. polls have notoriously been known to be wrong

in the past.

[West Coast Democrat]

7. Dean Results Were Within the Margin of Error

However, the 14% of Undecideds were a huge factor.

[Cooley Hurd]

8. Caucuses are notoriously difficult to poll...

...especially for an outside pollster (non-Iowan) like Zogby. Primaries are quite different and pollsters like Zogby are more in their element...

I'll bet they're very accurate in the NH primaries...

[maxr4clark]

10. Zogby admits that he picks his demographic to correspond

to his opinion about what is going on on the ground.

Statistics don't lie, but statisticians lie about (or more often misinterpret) how they collected the statistics.

[NewJerseyDem]

11. You can't poll caucuses

It isn't just that they are hard to poll, they are impossible!!! Because of viability and distribution of delegates they can't be polled. What is important is the comparison to the entrance polls which are very accurate which was something like 28% for Kerry, 22% Edwards, 22% Dean and 16% Gephardt. If my memory is correct then the Zogby poll was pretty good.

[HFishbine]

12. Quite simply

Edwards was a lot of people's second choice.

[Snivi Yllom]

13. apples and oranges

It does not work like that.

[creativelcro]

14. To draw conclusions you need to remember that the MOE was +-4.5%

Which means that Gephardt could have been at 22 and Kerry at 21, and it would still be compatible with this poll. This poll only showed that they could not make any predictions.

[Nicholas_J]

15. Yes

The local polls dome by regional pollster far more closely indicated the actual point spread between the candidates.

If the same holds true than the differnce between Dena and Kerry in the New Hampsire Zogby polls should turn into an even larger lead for Kerry in New Hampshire than he had over Dean in Iowa.

[Nederland]

16. My Conclusions

I'll explain Zogby's numbers, starting by adding a little something you forgot:

Zogby final polling on 1/19 in Iowa

Kerry 25%
Dean 22%
Edwards 21%
Gephardt 18%
Other/Undecided 14%

So what happened? Simple: Undecided voters broke for Edwards and Kerry--just like they had been doing all week long. The bottom line is this, when the undecided voter count is in double digits, the margin of error is a meaningless number because on election day, there are no undecideds. Undecideds are exactly that, undecided. They can break anyway they want to.

I guess the biggest problem is that people don't understand how to view polls. People range from thinking that polls are perfect to (the more common DU opinion) believing that polls are useless. The fact that polls aren't perfect is obvious to most people around here. It would be a mistake, however, to go from this assertion to claiming that polls are meaningless. They are not. They are really good at spotting trends in opinion and telling you who is coming up fast and who is dropping like a rock. They also are very good at telling you what the broad picture is going to look like. Did Zogby get it exacly right? No--but you didn't see Kucinich coming in at 20% either. So long as you don't expect too much, polls are one of the most useful tools politicians have to deal with.

[NewHampster]

17. Seems you forget the big Switch

Kucinich and Gephardt voters moving to the Edwards and Kerry camps.

Not accounted for in polls kind sir.

Howard Dean - Karl Rove's Worst Nightmare

[WiseMen]

21. The Direction in NH is just the opposite of Iowa. In NH Dean is SURGING.

Iowa comparison is irrellevant esp since Iowa was a caucus.

[SadEagle]

22. You're comparing incomparable numbers.

The poll is the first-choice support number. The results are delegate counts after re-allocation post viability. They are not directly related, nor they should be. If you looked at Zogby more carefully, however, you would have noticed that it put Edwards and Kerry doing very well in the second choice category. So it in fact did do a pretty good job predicting what would happen, if one took some time and thought out the effects of the actual caucusing.
Also, the poll was done before the Kucinich-Edwards deal, which probably also gave Edwards a couple of points.