If the pollsters screw up NH the way they did Iowa, is their credibility..

[kentuck]

shot? Pollsters have been a negative influence on our political system for years now. They mold public opinions with their poll numbers. For example, before he war, they had numbers showing a vast number supporting Bush on the war. How can that not affect public policy? They are scum.

We can only hope they are off by another mile in NH, so we can say they have no credibility. That would be one of the best things to come out of the primary. People need to put less faith in the poll numbers they see on the boob tube.

[THUNDER HANDS]

1. well if they mold public opinion

Why are they wrong so often?

[kentuck]

4. They are a lot like Vegas....

they are usually within the numbers...

[cthrumatrix]

2. I think people are getting smarter about polls and the media.

[Abigail147]

32. We wish.

People are amazingly self-absorbed, however, times are hard. Maybe this time.....

[creativelcro]

3. The huge variance of the polls yesterday already shows

that you can have them say whatever you want. And you can see certain sectors of the media picking up Zogby because they want to create last minute suspense; other sectors pick up polls showing Kerry with a 15 point lead, because they may have some other agendas...

[HereSince1628]

5. No, polls are like religion

folks may not entirely believe, but week after week, year after year the people return to check out the message.

[lcordero]

6. I wouldn't say that they mold public opinion

It's more like horribly distorting public opinion.

[Democrats unite]

7. If you really want to talk about polls...

You need to go back & look at the 2000 election & polls.

[Nederland]

8. Did they really screw up?

Zogby correctly placed the top four candidates within the margin of error. Can't ask for better than that.

http://www.zogby.com/search/ReadNews.dbm?ID=785

[kentuck]

9. Wasn't Zogby the exception?

Weren't most of the others totally off?

[Renew Deal]

12. Zogby was the only one correct about the popular vote in 2000.

He was also right in 96

[Nederland]

16. Des Moines Register was good too

John Kerry 26%
John Edwards 23%
Howard Dean 20%
Dick Gephardt 18%

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

[spooky3]

15. This is not correct. The candidates' actual percents were OUTSIDE

the MOE, except Dean's was barely inside. (See Will Pitt's thread on polls for details).

[Nederland]

21. See post #20 (nt)

[creativelcro]

17. Look at the numbers. What you say is incorrect

Zogby:
MOE was +-4.5%. Kerry 25 Dean 22 Edwards 21 Gephardt 18.

The results:
Kerry: 38%
Edwards: 32%
Dean: 18%
Gephardt: 11%

The only one that was consistent was Dean (lower end of the prediction). The order Zogby had means nothing because with that MOE it is consistent with all possible orders). The main thing I give Zogby is that he got that Gephardt was the last one. But even that one was a weak preditcion.

[Nederland]

20. Need to account for undecideds

Its not fair to compare poll prediction numbers with actual results because actual results don't have an undecided category. Besides, we aren't using polls to predict the exact percentage of the vote, we are using it to predict who will win. Ultimately, that's all that matters. If you are willing to say that Zogby doesn't correctly predict percentages but correctly predicts who the winner will be you'll get no argument from me.

[creativelcro]

25. Zogby did not predict much, when you look at the numbers.

The MOE was +- 4.5%. All numbers where within 9%, which means that all orders were compatible with the Zogby poll... Sorry, I was under the false impression that Zogby predicted something, before I actually rechecked the numbers...

[Nederland]

26. He predicted a Kerry Win

and Kerry won. 'Nuf said.

[creativelcro]

29. He did not predict a Kerry win. Not based on his poll data.

All candidates were within the margin of error of the poll. So, any order was possible. Pretty basic stuff, not sure why you keep arguing against the obvious.

[Nederland]

30. Last Try

I suppose you are correct in asserting that Zogby did not predict that Kerry would win. I suppose the more accurate way of stating it would be to say that Zogby thought that there was a high probability that would a Kerry win, and a low probability that someone else would win. To most people however, I think when a pollster puts a guy at the top of a list, they interpret that to mean that the pollster is predicting a win by that person.

But perhaps I'm denying the obvious. :eyes:

[mot78]

34. He also predicted Dean would come in second

[redqueen]

10. They screw up all the time and their credibility never suffers.

Go figure.

[creativelcro]

27. One problem is the pathetic math curriculum in US high schools.

[mistertrickster]

38. No kidding, just because the pollsters happened to get the top three

in the right order in no way shows that they were "right." God, I wish we could get basic statistics into the curriculum for every student. People are so ignorant about the subject.

[lastknowngood]

11. not as long as the sheeple keep quoting them

n/t

[xultar]

13. Puhleezzzzzzzzzzzzzzzz!

We say that now and then as soon as these races are over we'll start posting poll results and daily tracking polls.

[Ficus]

14. the polls in IA

were pretty much on the last week.

[JoblessRecovery]

19. That is my recollection too

They called it Kerry/Edwards/Dean in that order, and that's how it came out. Although I don't think their numbers reflected Dean would be as far behind as he was.

[Snivi Yllom]

18. well Zogby was right

and Zogby has it Kerry 28% Dean 25% and tightening....

[Feanorcurufinwe]

22. I hate to break it to you, but the pollsters predicted Iowa correctly.

What are you talking about?

[mistertrickster]

23. They predicted the winners, but not the percentages . . .

Kerry and Edwards got a much higher percentage than expected.

[NewYorkerfromMass]

24. Chalk that up to the caucus format and dynamic

Kerry and Edwards support ultimately snowballed and buried all comers.
In NH the percentages should be closer by virtue of all votes being cast once as final.

[creativelcro]

28. Look at the numbers. Then use them to back up your points

Otherwise they'll convince nobody.

[Feanorcurufinwe]

31. Yeah.

You are expecting them to get the percentages right? lol You are asking for too much. They're just polls.

[mistertrickster]

33. I expected the percentages to be within the margin of error.

Des Moines Register poll / Actual

Kerry 26 percent / 38 percent
Edwards 23 percent / 32 percent
Dean 20 percent / 18 percent

Only Dean was within the margin of error.

According to the poll, Dean and Edwards and Edwards and Kerry were running statistical dead heats. Yet in the final tally just some 12 hours later, they differed greatly. What if they had been "off" 12 points in the other direction, and Kerry had come in third with 14? They would have been just as WRONG or should I say, just as RIGHT.

[Feanorcurufinwe]

36. Well, now you know better.

[Nederland]

39. They were

If you take into account that undecideds can all break for the same candidate, or in this case, for two candidates.

[mistertrickster]

40. So your saying that the polls can't predict the final outcomes because

of all the undecided voters, which is what I'm saying, so we're in agreement on that.

The pollsters give a margin of error in which their results are supposed to fall statistically. If the results come in outside the MOE, the poll is essentially in error.

If they can't measure the undecideds, then they aren't measuring anything really worth knowing because their predictive value is meaningless.

My problem with polls is they tout these things like they really reflect people's opinions and they don't. But they do become self-fufilling prophecies when news media (like the Des Moines Register) scream a big headline about how the "momentum" is shifting away or for a candidate.

Take a statistics class . . . the scales will fall from your eyes. I took two, and a research methodologies class with a leading academic researcher in his field.

[Monte Carlo]

35. Not to antagonize, but has their credibility ever been sterling?

I mean, there are lies, damned lies, and statistics. I think we're going to end up right where we started; we like polls when we like the results, and we don't when we don't.

[kentuck]

37. It just hasn't been questioned....

But it should be a big deal if they are totally off track. Because, there is too much dependency on polls in this country and we should be ready to question their credibility if they screw up. Just my opinion.