Can we agree on a reputable poll or polls?

[NYCliberal]

Lots of time wasted in this forum citing any poll that shows Kerry up, only to trash that same poll if it shows him down or losing ground.

Historically, what have been the most accurate polls, and waht do they say now?

[seabeyond]

1. i dont do any of the polls, closer to election more off they are

they arent including youth, new registration or cell phone users, and seems to exclude minority. so no polls for me. none of them feel right. the ones for or against kerry. let them all go

[sonicx]

2. use common sense. if a poll has either candidate leading OH or FL by

more than...i dunno..5%, take it with a grain of salt.

don't sweat polls. they sucked in 2000.

[kaitykaity]

3. Rasmussen, Zogby, and Harris

are some of the best around.

The real stinkers are the Gallup, RealClearPolitics,
and that electoralvote.com thing (they use whatever poll
is handy and do not discriminate).

[DemocratSinceBirth]

6. Rasmussen Had Bush Up By Nine In Their Final Poll In 00

nt

[kaitykaity]

13. So?

Is it evidence of a permanent bias, or is it a
mistake of methodology since corrected?

Whatever.

:shrug:

[NDFan]

4. Historically

Polls have been wrong.

In 1996, they were off, on average, nearly 5 points. In 2000.... well, you know what happened there.

No poll, when you think about it, is "reliable", especially in a close election, as 2000 proved.

They only get it right when a candidate wins by a landslide.

[DemocratSinceBirth]

7. 1996 Was A Qualified Landslide And The Polls Got The Right Winner...

When a race is as close as this polls lose their predictive powers...

[NDFan]

9. I agree

Because 1996 was a 9 point landslide, the fact that the polls were off a few points didn't matter because it wasn't a close race.

"A few points" in this race means the presidency.

[DemocratSinceBirth]

11. If You Look At The Polls Over The Last Fifty Or So Years

they do a nice job but in close races like 68, 76, and 80 the Margin Of Error is the fly in the ointment...In those years most pollsters were within the moe but some picked the wrong winner...

[Stuckinthebush]

17. Historically

Polls have been rather close.

Remember, polling is not an exact science. People seem to want polls to predict the actual breakdown of the vote. They can never do that because there will always be error in polling - hence, the MOE.

Polling is an amazing tool. If you see a poll that shows a tie or Kerry down by 2 or 3, then this is good news. Usually the methodology of polls is inadequate to get at the Democratic vote. Therefore, the Dem vote is usually underrepresented.

Don't expect polls to be right on the money. They can't be. Use them the way they are intended to be used...as a measure of the general attitude of voters at a single point in time based on a sample that is assumed to be representative of the population (but that can never be representative). In summary, a poll is a general snapshot that usually underrepresents Democrats.

Don't hate polls, hate the use of polls by the media.

[Mistahkleeen]

5. Everyday the results are different

The polls serve no other purpose, other than to confuse rookie voters.

[newyawker99]

22. Hi Mistahkleeen!!

Welcome to DU!! :toast:

[billyoc]

8. No. n/t

[Renew Deal]

10. Election day is the most accurate.

5 days to go.

[Orangepeel]

12. even well done polls sometimes give you screwy results

A margin of error of +/- 3 means that 95% of the time, the sample results will be within 3% points of the population.

That means that 1 out of 20 times, even the most well done poll (which most of these are probably not) can give you off-the-wall results because of random error. Sometimes, you just happen to call a bunch of crazy people in one poll.

Everybody is polling so often, that a lot of the polls are going to be screwy, even if they were done perfectly.

[archineas]

14. let's not forget

that niether candidate is polling outside the margin of error. the poll results aren't even statistically significant until you have consistent polling outside the MOE.

the only conclusion that i would come up with, as a mathematician and if this data were my own, is that there is no sufficient information to offer a prediction at this stage, simply by the polling numbers.

i would, however, LOVE to see the internal polling done by both campaigns!

j

[Marxdem]

15. Zogby has been the most accurate

None of the polls were as close as Zogby in 2000. This election is going to be close, theres no hiding that.

[DemocratSinceBirth]

16. He Was The Least Accurate In 2002

nt

[Marxdem]

19. Big difference

Presidential elections are easier to poll then non-presidential elections. Many people polled don't show up to vote in non presidential elections.

[cubsfan forever]

18. GOTV n/t

Professor 2

[Killarney]

20. I think Zogby is as good as you'll get, but what matters

much more than polls is turnout and we don't yet know that. :)

[Marxdem]

21. Exactly

Thats what I have been saying right along.