STOP LISTENING TO NATIONAL POLLS-THEY DON'T MATTER!!

[ncteechur]

We constantly get ourselves in a panic about what the national polls say.
They don't amount to a hill of beans.

The election isn't won nationally. (We saw that in 2000)

They don't poll representative samples. They constantly poll more repubs.

Stop worrying about the national polls!!!!

Kerry-Edwards all the way!!!

[secular_warrior]

1. The polls matter for a few reasons

like guaging momentum and overall public sentiment about the incumbent and the country's direction, but I agree that it makes no sense to get caught up in any one poll, especially with how wacky they've been this election season.

[JusticeForAll]

10. State polls may matter, but national ones do not...

What type of momentum can you truly gauge from a national poll that is more specific and defining than a state poll? I really need someone to explain that immense leap of faith from the hypothetical to the causal that is implied over and over here.

It seems like a lot of DUers are falling for the stupid CNN-SHIT they have been spewing for the past several weeks, because CNN cannot find an electoral map of state polls that SUPPORTS their hero, the treasonous lying bastard *.

So therefore CNN says oh LOOOK! National poll says Bush ahead!! that's MOMENTUM, you sheep!

Hogwash, it says you polled more from red fucking states than blue ones, or it says you polled more Repubs than Dems, or it says you ignored the fact that Kerry-supporting cellphone users were ignored, or it says that your likely voters did not include all of the Democratic/independent newly registered who WILL vote.

Quit buying into the hype. Follow the details.

[jdsmith]

2. No, but uninformed voters' reactions to polls are of utmost importance

Some people don't decide how to vote till the bandwagon starts rolling and they jump on. No, we should never allow national polls to affect how we feel--we know how valuable they are (not) as barometers of the political weather. But we HAVE to "listen to national polls" (and, I'm afraid, watch as much of CNN as we can stomach) if we want to anticipate where those uninformed voters might be likely to go.

[Stuckinthebush]

3. Actually, the polls do matter

They are very accurate and quite amazing when you look at the science involved.

The problem is in the interpretation of polls. People see a national poll with Kerry down by 2 or 3 and they begin to panic. They forget all polls have error built into them and that error is quantified using the MOE.

The polls give us a pretty good glimpse into the mind of the electorate at a certain point in time. If we didn't have them, we'd be campaigning in the dark.

Don't attack the polls, attack the use of polls.

[secular_warrior]

8. "Don't attack the polls, attack the use of polls."

I think if they are wildly off this election, people will have a right to attack the polls. People of all political persuasions should insist on some sort of consistency and accuracy in light of how they can influence an election - it can't just be about how sound the scientific method is on paper.

I agree though, the use of the polls are the bigger problem.

[Stuckinthebush]

12. Polling is a wonderful tool

If a poll is off by 5 or 6 points, it is still a good tool.

I am not convinced that any of the major pollsters are intentionally skewing results to help out Bush. I think Gallup, for instance, has bought into their methodology of weighting the results with unrealistic turnout numbers, but I don't think they are intentionally trying to distort the national picture in order to help Bush. It would be a risky move to try it. Actual results will be wildly off of any manufactured poll result, so the pollster loses in the long run because people begin to question the validity of their methodology. Once that happens, a pollster is toast. The almighty dollar trumps any pollster's political affiliation.

[secular_warrior]

15. I agree. I don't think they're intentionally skewed to help Bush.

They'd be aiming for accuracy first and foremost in order to be able to command top dollar for their services. But like we discussed in another thread, Dem voters tend to be harder to accurately poll, and there's the problem of a changing information age society that doesn't answer phone, and some cell phone users don't even have a land line phone. The polls themselves can influence an election (everyone wants to vote for the winner), therefore if they very inaccurate - that is a very big problem, for which we'd need laws to address how the polls are used by which media outlets, etc etc.

This is a much bigger situation that has less to do with the scientific method of polling, and much more to do with how polls are used - just as you stated in your previous poll. Above all else, polls should be used in such away that they don't influence the election itself.

[ChavezSpeakstheTruth]

4. I've been saying that for months!

[DemocratSinceBirth]

5. What Should We Rely On?

[ChavezSpeakstheTruth]

6. What do you actually get from polls? You rely on them?

How do they change the campaign? They are a media sham

[DemocratSinceBirth]

7. I Have Some Idea Of How The Race Is Going...

Without the polls to guide us it would be like trying to find Pin Point, Georgia without a map...

[ChavezSpeakstheTruth]

9. But if the polls are inherently flawed and you aren't getting a true

sampling - what the hell good are they?

[DemocratSinceBirth]

11. I Think They Give A Decent Picture...

This race is close and fluid...


That's why pollsters are having a hard time...

[ChavezSpeakstheTruth]

13. I'm not so convinced its so close - granted I live in CT but I find an

unusually strong anti-bush sentiment.

[Stuckinthebush]

14. Social science is inherently flawed

yet it is used all of the time.

Most statistical analyses and probability theory rely on true random sampling. It rarely happens in social science. There is no way to insure that every member of a large population has an equal chance of being in a sample. All social scientists (and pollsters) know this. So, they do their best to get the most representative sample they can. With polling, it isn't just the sample you have to worry about, it is the turnout as well. So, pollsters have to make an educated guess on what the turnout will be. Some do a better job than others.

The public often believes that polls should be right on the money. It won't happen that way. There is always a band of error surrounding the results. Pollsters describe this error using the MOE or margin of error. The MOE is based on the assumption of random sampling, so it is never quite accurate either. Pollsters know this, too.

The big problem is in the use of polls by the talking heads. They don't understand the methodology, they don't understand the science, and they don't understand the MOE.

So, look at a poll as a rather good measure of the electorate's feelings at a specific point in time, and mentally place that MOE band around all data. Polls help us understand the general mood of the country, but they will rarely ever hit the final election numbers on the head. Unless the pollster is really, really lucky.