Can anyone here demonstrate that national polls are wrong?

[MonkeyFunk]

I see it said all the time - I'm accused of being a Fox-news loving freeper because I believe the polls are largely correct.

Can somebody find 3 examples of a national poll, just prior to an election, that was wrong outside its margin of error on election day?

[Byronic]

1. I think the national polls are largely correct

but the problem is they are so mercurial, and can change so quickly.

You rarely see anybody dispute a poll which shows THEIR candidate ahead. The 'whining' comes from those lower down the polling scale. And, believe me, as a Joe Biden supporter, I say 'A Pox on the Polls!'. And I say it in a VERY whiny voice.

[MonkeyFunk]

2. Heheh....

yeah, that's the point.

If your candidate isn't doing well, the polls suck.

But the fact is, the polls are almost always correct. Clinton is leading (not by as much as she was), Edwards is fading, Kucinich is at 2%.

People hate when I say that, but they can't show that polls are unreliable.

[lostnotforgotten]

3. I Will Never Vote For A Bush War Enabler

So, I really don't care if she is leading in the polls or not.

Some of us are not lemmings blindly following others to our death.

[MonkeyFunk]

7. And that has exactly what to do with my question?

Try to keep up.

[ronnykmarshall]

41. It's hard for the one track minds to keep up.

[calteacherguy]

4. I think polls are overall reliable.

Of course, some are more accurate than others depending on the methodology used. Polls can be manipulated, and often are. However, if you take an average of recent polls that generally gives a pretty good idea of where things stand.

National polls aren't really important in a primary season, because obviously it's the state polls that count. In a caucus I think polling becomes really challenging because such a small percentage of the population historically attend a caucus. It comes down to who is going to be willing to go out in the dead of Winter and sit or stand for an (hour?) to support their candidate. And then there is the 15% rule where if a person's candidate is not viable, they go to their 2nd (and perhaps in some cases 3rd) choice. So, I think a caucus is particularly difficult to predict with polling.

The other thing is, unless Clinton really has a blowout in Iowa, which I think is unlikely, the polls will change dramatically in ways impossible to predict. So, polls are interesting now but certainly not determinative.

[Yukari Yakumo]

12. Totally disagree with the last paragraph.

For Edwards, he must win and by a decisive margin. With the NH primary right around the corner and him lagging way behind Clinton and Obama, he needs a big momentum boost to have any realistic shot at NH. Anything less and his campaign is over.

For Obama, he must win and by at least a "recount-proof" margin. He too trails Clinton in NH, but not by the margin Edwards trails her. Obama doesn't quite need as much of a momentum boost to make that race competitive.

For Clinton, if she wins or comes in a very close second, it's over. However, she can afford to lose Iowa for as long as it is only by a small margin.

[Life Long Dem]

5. Wrong. Edwards is rising in the polls. Hillary is slipping.

[MonkeyFunk]

10. Not from what I've seen

Edwards is doing poorly. As he should.

[sss1977]

6. wrong as in inaccurate, or wrong as in immoral?

Here's how it works. The media starts mentioning a couple names over and over again as "frontrunners". Being good little consumers, we remember those names. Then some pollster calls up and asks something about the candidates, which is then reported on. Voila, you have yourself a positive feedback loop at work. It goes around and around as time goes on. Media coverage leads to polls, polls lead to new media coverage, new media coverage leads to new polls, on and on we go. The wheels on the bus go round and round.

So go ahead and frame your question on the accuracy of polling. It won't change the fact everyone being polled is getting their news from an ever-shrinking list of massive corporate interests designed for profit. And those interests already have in mind how they want the electorate to vote.

So you ask, "are they accurate?"

I ask, "should we even fucking care about polls so much if we want a real democracy?"

[MonkeyFunk]

8. It's simple...

find a poll that's wrong.

If you think Kucinich is really doing very well and Clinton is doing bad, the the votes will show that. But you'd be wrong.

The polls are right.

[sss1977]

21. did you even process what I wrote?

Let me make myself a bit more clear.

Let's say you convince a group of 100 people that the sun does indeed orbit the Earth. I mean we see the sun out there rising and setting every day, it's pretty obvious, isn't it? So then you poll them to see just how many people believe the sun orbits the Earth.

So I ask again. Who cares if the poll itself is accurate or not. Don't you get it? Polls are just a tool. Don't confuse a tool for a brain.

[MonkeyFunk]

9. If your candidate is not doing well,

then the feeble-minded should not care.

But reasonable people should pay attention.

[Joe Chi Minh]

33. Thye don't allow them in France until after the last vote has been made.

But what would the Frenchies know about politics, Mutley Monkey.

I remember rebutting posts of yours earlier, which I don't think concerned polls, but did seem very right wing.

[MonkeyFunk]

35. Well your recollection is faulty

I'm nothing close to a right-winger, but people here seem to think that living in the reality-based community=right-wing. It's sad.

[Joe Chi Minh]

54. Why would people favour Hillary and the DLC, when they echo the Republicans,

"We can't afford to permit the top 1% not to keep trebling their income every ten years, while the people work longer and longer hours for less and less?" You, in the reality-based community? You're just sick to compound your folly by saying a thing like that. Read the item below, and spend the rest of your life as a penitent. You've not only been found out, but your tosh is ridiculed mercilessly:

http://comment.independent.co.uk/commentators/mark_steel/article3179607.ece

[fasttense]

11. Yes, there are some polls that have systemic problems.

The first and most obvious way they skew polls is how they word the questions. For example on using force in Iran, they will say if all diplomatic and political action fails, should force be used to prevent Iran from gaining nuclear weapons. The problem is that the US is not using diplomatic efforts and refuses to sit down with Iran to come to a resolution. There is also no clear evidence that Iran is close to building a bomb. So the whole question is totally hypothetical. But they tout the poll as being an accurate indication of the American opinion on using force in Iran.

Next they skew their samples. They poll many more republicans and conservatives then they poll democrats and progressives. The truth is, we have more people leaving the republican party and yet they are over represented in the sample groups.

The more subtle way of skewing the polls is attaching weights (supposedly based on US demographics) to certain demographic groups' responses. For example, they poll only 30 women in a sample size of 100. So they will give the 30 women's opinions a number to multiply their response to reflect that females make up 51% of the population. These multipliers are hard to find and most polls don't publish them. So you never really know how accurately they weigh the sample population. It's not so bad when you are dealing with the male and female ratios but when you get to the conservative vs liberal ratios it frequently is bias in favor of the conservative opinion.

[Perry Logan]

13. I think all of the polls are biased toward the Republicans.

[Joe Chi Minh]

34. Precisely. They're a well-known political tool among the politicians

and journalists. It's not regarded as a close secret. On the contrary, when they refer to them among themselves, the understanding of them as political tools is implicit. No question about it, still less a need to explain.

[glowing]

14. Reality is that the very first polls should just be published to general

pop. now. In 2004, they were already doing polls on Clinton's electablity. Actually, the moment her husband left the office, they were specualting on Hillary Clinton's inevitable road to the white house. AND no matter what they tried to play by telling the people of NY that she had no plans for the Whitehouse, she had puposely moved to NY to run as a senator so she would have the name recognition on a larger scale.

Has anyone noticed how much time they would give to her speaking from the floor of the Senate, and she was a freshman. The whole thing has been rigged for quite a while. This has been her intention all along.

Since she's been talked about for the last 7 yrs as a presidential contender, she happens to be on the front minds of most people. Some people don't pay attention until the last minute anyway. That's when they decide. There is a huge margin of undecided. In some states it doesn't matter what party you are registered to, you may still vote for whichever person you want to.

Polls are probably more important for the person running. They are horrible for the general pop. I've said for a while I'm stuck between Kucinich and Edwards.. wish they'd announce a ticket with the 2 of them together. I think they would be great co-workers and they would work towards peace, stability, strong job market, and universal healthcare.

[tkmorris]

15. What does the "polls being wrong on election day"

Have to do with polls months away FROM election day? If you were attempting to influence voting behavior through bogus polling data you would fudge it completely early on in order to get people to climb on the bandwagon, and as election day approached you would bring your data increasingly in line with reality to maintain the illusion of integrity.

I'm not saying I believe the polls are fudged, merely that your question doesn't address it either way.

[MonkeyFunk]

26. Because

I believe people have a basic misunderstanding about polling.

A poll months away from election day doesn't claim to predict what will happen on election day. It only claims to measure the public sentiment at that time.

[mrgorth]

16. Dude,

I don't deny that Hillary might run the tables. But seriously, your national polls are shit. IA and NH will change everything. Maybe, as I said, Hillary clobbers in both and her polls just get stronger but all it takes is someone doing better than expected and your national polls 180. May as well know that going into Jan.

[Stuckinthebush]

18. Yes, I think national polls are not as reliable as state poll numbers

Because we have a primary system that allows a handful of states to influence the race - and media that allows those states to influence the race - then national polls are not as meaningful. Voters are influenced by the outcomes of those races and change their views to match the "wisdom" of their predecessors.

[William769]

19. Your right Iowa and NH will change everything.

Will give Hillary even a larger lad.

[mrgorth]

23. Larger than Bill?

He's pretty big isn't he?

[William769]

24. Yes he is.

Anyone out there would love to have his poll numbers, and he's not running! :rofl:

[MonkeyFunk]

27. Polls don't claim to predict what will happen months away....

can you demonstrate that the polls are wrong?

[Joe Chi Minh]

36. No need to. Who are you?

[MonkeyFunk]

42. Well

that was sort of the topic of this thread.

Try to keep up.

[Stuckinthebush]

17. Polls are correct if the right sampling methods are used

I think people are more likely to call out polls if they don't show the support for their candidate that they'd like.

Polling uses tried and true methods of statistical sampling. Each poll has a margin of error for a reason - the pollster knows that based on her overall "N" there is a level of probable error. This allows the pollster to set up bands surrounding the observed percentages. We see the margin of error but many times we ignore the margin of error when interpreting a poll.

Poll results fluctuate because of this margin of error and because, naturally, people's opinions of politicians fluctuate. Two months ago I'd have said "Strongly Agree" to a question about Hillary's leadership ability, but today I'd be at Agree or Disagree. Next month, I might be back up at Strongly Agree. Polls are simply a snapshot in time.

It is clear that some polling companies are better at sampling strategies than others. There have been studies on which companies use the best methods, but rest assured, the companies are constantly upgrading their methods based on the latest statistical studies. There is little incentive for a polling firm to run bad polls.

[MethuenProgressive]

20. Any poll not showing Obawards winning the Amazing Race was fixed by Rove and Clinton.

Everybody knows that!
:rofl:

[Clintonista2]

25. LOL!

[CK_John]

22. A poll is a snapshot of the time it is taken. Will it the predict future, is the real question?

Polling methods have not kept up with todays eWorld, so for some things that are no longer actuate for some topics. Example: if you were to poll about the most favored "ringtone" using only landlines, your poll would not be very accurate.

As far as politics goes, we need to determine if most voters still use landlines? IMO, most of the GOP still use landlines, DEM have both, and IND and non voters use cell only. I think they still are a fair snapshot for this election, but will need to change for the future.

[MonkeyFunk]

28. Of course not

and that, I think, is the basic problem people have understanding polls. They don't make any claim to predict the future.

But I don't know of any polls, the day before an election, that were proven to be wrong outside their margin of error. That's my only point: polls are largely accurate, and people who dismiss them are being willfully ignorant.

There are people here who actually believe Kucinich has more support than Clinton, and they believe the polls are all lying. That's a level of denial that's incomprehensible to me.

[CTyankee]

30. I agree but I do scrutinize polls more closely to see how big it is and

who is being polled. National polls tend to be big so have a smaller MOE. And polls that just poll Democrats about how popular HRC is only measure that and cannot be used to then apply it generally.

Well constructed, scientifically designed polls are, IMO, useful and true. However, some people here might be remembering the exit polling that was so off in Ohio in 04. The mathematicians who studied that said that the data was suspect which gave rise to the theories that votes were tinkered with. Do you have an opinion on those theories?

[Joe Chi Minh]

37. ...'cep the photoshopping is done before.

[CK_John]

43. Many people go though life unable or unwilling to make a decision, they rely on polls to tell them

who to vote for. The most honest of the decision challenged stay home and don't vote. It is those who are in denial, that vote for sure losers, just so they can rationalize their inability to make an informed decision with, "don't blame me,I didn't vote for them".

[Joe Chi Minh]

53. Interesting comments. Especially the last one!

[CTLawGuy]

29. it's not that they are not accurate

it's that they are not helpful because the primary election is not a one-time national election. The more relevant polls are the IA and NH polls, which are a lot closer.

[Beaverhausen]

31. DING DING DING!!! correct answer

[Joe Chi Minh]

38. Your second sentence is not a corollary of your first one. Since they

are political tricks of political chicanery, they are still irrelevant. Worse, they are massively misleading, as the actual landslide victory Kerry was cheated of demonstrated yet again.

[Bjorn Against]

48. And not even the Iowa or New Hampshire polls mean much at this point...

This is a primary election, and people are much more willing to change their vote at the last minute in the primary than they are in the general election. Many voters like multiple candidates, and there are not all that many voters who are strongly committed to one primary candidate and will not accept anyone else. Look at the polls from Iowa and New Hampshire in November 2003, and you will see Dean with a lead in both states and Kerry not doing so well. We all know what happened there though.

[KittyWampus]

32. IMO, Polls are part of an extremely dysfunctional system. They simply serve to reinforce what Media

has established in their narratives.

[Joe Chi Minh]

39. Spot on. Another component of the Republican noise machine.

[ronnykmarshall]

40. Because!

Because the "MSM" is a tool of the Klin-toon machine!

Hilabeast murdered Vince Foster!

She's a lesbian!

Chelsea is ugly!

Bill let Bin Ladin get away while Moncia sucked his dick!

:sarcasm:

Hey! Figured I'd like the Hill-bashers save some typing efforts.

They're all using right wing bullshit from the 1990's anyway.

[IA_Seth]

44. My theory..

They aren't wrong in the sense that they do reflect (accurately even) what a given sample of people feel at the time the poll was conducted.

However, I believe they are wrong in the sense that being a staggered primary/caucus season, each prior primary/caucus results influence to some degree the outcome of the following primary/caucuses.

So yes, they are accurate right now, but these results will most likely not be the overall results after all is said and done.

In conclusion, polls suck ass.

[Stop Cornyn]

45. The national polls aren't wrong, they are just irrelevant. National polls tell how the candidates

would do in a national vote, but there isn't a national vote in the primaries.

The primaries and caucuses go state-by-state or a few states at a time. The campaigns are only campaigning in a few states.

As a result, an Iowa poll, for example, is a poll of people who have seen months and months of TV ads, candidate mailers, candidate events, campaign news leading the local news on TV and in the paper, etc. A Texas poll, in contrast, is a poll where NONE of the poll respondents have seen more than a TV ad or two, haven't received even one single mailer, haven't been to any events, and who watch the local news without hardly any mention of the presidential election.

National polls survey the attitudes of people who mostly haven't been affected by any of the party's campaigns. Iowa and New Hampshire polls mostly survey the attitudes of people who are hip deep in the campaign process.

But most importantly, national polls are irrelevant because history shows that the results in caucuses and primaries are SIGNIFICANTLY effected by the results of earlier primaries and caucuses. For example, the difference between a week-before-primary polling in New Hampshire and the actual New Hampshire primary results can be affected by as much as 20% by the results of the Iowa caucus (winning the caucus could mean a 15% boost in New Hampshire and a third place finish in Iowa could sink a candidate by 5%).

[MonkeyFunk]

46. I agree

I think it's perfectly fine to argue that national polls have little to no predictive value.

I think it's silly to argue that they're deliberately rigged.

[harun]

55. They mean nothing at this point so I don't know why anyone even

wastes their time on bringing them up.

[Bjorn Against]

47. Well go find me a national poll from November 2003...

I think you will see the results of any poll you will find from that month look very different than the actual presidential primary results. It is not that the polls were wrong necessarily, it is just that they were meaningless because primary races depend so much on momentum and on the fact that people are far more likely to change their vote at the last minute in a primary than they are in the general election.

[MonkeyFunk]

49. I don't claim they have predictive value

I just reject the notion that they're rigged to support a certain outcome.

[Bjorn Against]

51. They aren't rigged, but it is pretty irresponsible for the media to base coverage on poll numbers

When they know full well that primary poll numbers mean nothing yet they continue to base the airtime they give each candidate off their poll numbers they create an uneven playing field, and yes work to rig the contest in favor of those who poll well. The polls themselves are not rigged, but media coverage of the election certainly is.

[dugggy]

50. Yes, 2004 presidential election

n/t

[MonkeyFunk]

52. All the polls

I saw showed a very close race right before election day.

Do you have examples of some that were off beyond their margin of error?