Why Bush pollsters (CNN/Gallup, FOX, CBS/NYT)..will keep his ratings up...

[TruthIsAll]

They will not let his ratings dip much below 50%, because they know that for him to steal it again, he can't appear to be too far behind.

Only Zogby (a Democrat) and maybe one or two others will have honest poll numbers. How else do you account for the discrepancy between Zogby (45%) and the Bush pollsters (aprox. 58%)?

For Bush to steal it, the following conditions must exist, regardless of whether he Wags the War Dog, has another Trifecta, or raises the terror alert code:
1) His favorability rating must not drop too low.
2) No exit polling can take place (just as in 2002).
3) There must not be voter verified paper trail.

Otherwise,

B
U
S
H

I
S

M
E
L
B
A

[UpstateNYDem]

1. Nice try...

There are very explainable reasons why there is a discrepency between Zogy's polls and the recent ABC/WP poll. First, Zogby's question wording is ambigious in one of the catergories of disapproval, rating the president as fair is considered disapproval. Second, Zogby only polls likely voters asking all adults or registered voters adds more status quo respondents. To accuse Zogby or any other pollster of bias it is way off base, the only bias there is in polling is accuracy, its a competitive business and being the most accurate is whats important, not an agenda.

[papau]

3. Likely voters favor Bush more than Dems usually than Reg or all

So that does not explain it -

But the 52% the yes/no folks have gotten verifies the 45% Zogby as a 4our choice poll like Zogby usually gives result in favor of current president that are 5 to 7 points lower than the either or poll.

Todays ABC's poll's pro bush jump above 52% is of more interest as to Bush bouncing up after the 45% Zogby "low" point.

I do not think so - it seems just a bad ( stat outlying number)number.

[diplomats]

5. The Post polls are always about 5 points higher than others

so that means the 52, 53 percent number is probably closest to correct. BTW, see http://www.pollkatz.com (approval graph) to see how high the Post polls are.

[UpstateNYDem]

7. Incorrect

People who don't vote favor the status quo or dont pay attention to politics and thus support the status quo. There is a big difference between RV's and adults who resemble Democrats who don't vote and those that do; voting.

I do data analysis and methodology for a major polling firm so I think I have a little better idea of what I'm talking about then you.

[Beetwasher]

9. I believe it was a Gallup poll

That also asked all the respondents who they voted for in 2000 and the majority claimed Bush by a sizeable margin outside MOE, so that should tell you something about their sampling methods. BTW, Gallup has him at 52%. That being said, polls are an indication, but by no means dead-on accurate and, IMO, weighted towards the right. We don't see their true methodology and how they're formulating their weights.

[UpstateNYDem]

10. Actually

That is quite typical, people don't like to be losers, so in retrospect they say they voted for the winner. Look at almost any poll asking that, the winner almost always polls winning by a larger margin. Of course some might argue that is a variable that should be used for weighting, though I'm not sure any firm is doing that yet.

Zogby actually weights towards Republicans, because he believes Democrats are more likely to answer the phone. But irrespective of that all of these firms belong to AAPOR or NCPP and can demand any pollster reveal their methods or face a fine like Frank Luntz did over Contract with America polling, and then you risk your credibility.

[Beetwasher]

13. Just out of curiosity

how often does AAPOR or NCPP demand methods?

Also, are pollsters methods protected as trade secrets?

I actually worked in epidemiological research and am familiar with similar data analyses, but that sort of research is regulated differently I'm sure and I'm not familiar at all with the regulations on the commercial/political end...

[UpstateNYDem]

22. I'm not sure

There is presently something going on between Zogby and NCPP because of the recent Iraq poll they did. Of course Mitofsky is behind it and besides being the a brilliant pollster, hes a gaint pain in the ass.

NCPP and AAPOR have the ability to go through every piece of your methodology if they want. Besides, the weighting is a familarly simple part and most pollsters know each others.

[Beetwasher]

24. Zogby's being investigated?

Or audited? What was wrong w/ the Iraq poll? Do you have a link w/ more info about this?

Who calls the shots at these organizations and decides a review is necessary?

[DemocratSinceBirth]

27. There Is An Association Of Pollsters

who adhere to a set of recognized standards...

www.ncpp.org

[Beetwasher]

28. Unfortunately

That site doesn't really answer the questions I have.

For instance. Is the methodology handled and/or considered a trade secret? I've never seen how they construct their weights for instance. Is this information only handed over the NCCP if a firm is being investigated, and even then is it kept confidential among the NCCP committee members? It seems to me that it's only the Disclosure Committee gets to see this material and then they make a decision.

In the scientific research community, because of peer review and the openness and publication of results, methodology etc. in minute detail there's an honest assessment of the work and therefore a higher level of confidence in it's efficacy...It doesn't seem to me that this is the cases for pollsters. Take a look at the officers for that organization.

[DemocratSinceBirth]

29. It's All Based On Inferential And Descriptive Statistics...

[Beetwasher]

30. Yes, I know, which makes it very subjective

That's why it's difficult to really tell the validity of these data. Knowing how the weights were constructed would help...

[DemocratSinceBirth]

31. It Is Somewhat Subjective

because it's social science not natural science where you have much more control over the variables but polling uses the same mathematical assumptions(inferential statistics) as geneticists use in dna testing...

Here's a simple test...

Take a coin....

Flip it five times...

You might get four heads but flip that coin one thousand times and I'll bet you get damn near five hundred heads and five hundred tails.... Polling is based on the same method.... I don't need to interview every single voter to know what they all think...

[Beetwasher]

33. Thanks but

I'm aware of that...I'm just trying to find out more about the weighting and whether the methodology of the pollsters, or at least portions of it, is considered secret and/or protected to some degree...

[DemocratSinceBirth]

45. Alot of It Is Art

Polling is an art and science...

With the proliferation of fax machines, cell phones, call blockers, and answering machines you might have to make one hundred calls to get one valid response.


Pollsters match their response to the general population, i.e. you need the correct percentages of Dems, Reps, whites, blacks, jews, hispanics, etcetera... If you don't do this your poll will be badly skewed...

If you are interested in learning more about polling, e-mail or call any of the polling organizations and am sure they will help you...

[DemocratSinceBirth]

21. Do A Poll Of Polls

or throw out the outliers, i.e. the highest and lowest polls...

As someone who has done post grad work in Government and done a fair share of statistical analysis it makes me sad to see polling dismissed as voodoo...

[TruthIsAll]

4. Nice try. I never accused Zogby of bias. But as for FAUX, you be the judge

Zogby was right on in the 2002 Senate elections, but he was made to look bad because the Repukes stole at least 4 of them with unbelievable turnarounds of 8% or more.

I believe Zogby is honest; it's the corporate media pollsters who are shameless in applying teflon and propping Bush up, who I don't trust one bit.

Your statement is quite naive; Bush pollsters want to look like they got the polls "correct" because they know Bush will cheat and make up the difference which they have already padded for him.

In other words, if Bush really pulls 48%, but pollsters have him at 51%, and he fixes 3 million votes to "win" at 51%, then the pollsters will appear to have been accurate.

On the other hand, Zogby, who got it exactly right, will be made to look like an amateur.

In order to cheat, rule number one is: don't make it appear that you needed to cheat to win.

[UpstateNYDem]

18. Ugh

You people here and your damn conspiracy theories. No Republican won a race that the trends were indicating otherwise, based upon what goes on behind the scenes, the trends actually indicated that Thune would win, that was the only suprise of the election. All trends pointed to even the Republicans who were behind winning and Carnahan and Talent dead even.

[TruthIsAll]

25. Conspiracy theories, you say? GA, MN, CO, NH, TX?Are you a DEM?

Trending towards Repukes? Quite the contrary.
Are you new to DU? Do you follow BBV? Do you know about Diebold?

Repugs didn't cheat? Twenty-two thousand voting machines software "patched" in Georgia the weekend before the election.

Four out of 10 critical senate elections ALL falling for the Repukes when the latest polls had the Dem ahead by MUCH more than the (+/-3%)MOE?

Chances of that being just a "coincidence": Less than 1 out of 40,000

As a Dem, you are quite naive in believing the Repukes didn't fix the elections to gain control of the Senate..

[UpstateNYDem]

41. Not trending

Using advance statistical means with data from previous elections I came up with the winner in all but two of the statewide elections we polled in 02, South Dakota and Missouri, in September.

I am not going to even get into that idiotic MoE argument I've seen here. It shouldn't take much more than common sense to understand why that is completely irrelevant.

[TruthIsAll]

42. You may have called them correctly..but only because they were stolen

Your limited knowledge as well as your Repuke agenda have been exposed.

A true Democrat would NEVER disparage Dennis K. as you just did.

Who did you vote for in 2000? Betcha it was Bush. OK? I'd say chances are 95% you voted for the Chimp. Tell me I am wrong.

[leftofthedial]

11. "the only bias there is in polling is accuracy"

Ha ha ha, BWaaahhhhaaaaaaahahahahahahhah

'scuse me haahhahhahahahahahahahahhaahahahahahahahahaha


Did you really say that?

Thanks, Ineeded a laugh.

[UpstateNYDem]

19. Ouch

That really hurts coming from someone out of touch with reality enough to support Dennis Kucinich.

Partisan pollsters have to worry about helping their clients win, and if you want to win you need the most accurate picture of the situation you can get.

Independent pollsters have to worry about getting clients and the more accurate the more clients you will get.

Media pollsters have to worry about selling newspaper/getting viewers and the more accurate you are the more you get.

Go read about the history of exit polling and see what the lengths the media outlets were going to be the first and the most accurate.

[TruthIsAll]

26. Seem to recall that media-run VNS did NOT exit poll in 2002..

The media was afraid we would conclude the elections were fixed.

You just gave yourself away, bud. You badmouthed Dennis.

I will refrain from telling you what I really think (I do not care to be flamed here). You do NOT sound like an outraged DEM to me; more like a "Get over it" Repuke.

Having DEM in your moniker doesn't make you one. You have been exposed.

[UpstateNYDem]

37. Ugh

The VNS system crashed in 2002.

Some of have to function in the real world, why don't you armchair Dems stay out of it so the real Dems can win elections. A conspiracy theory has never won an election.

[TruthIsAll]

43. VNS crashed? Oh, that's why. Get Real.

Democracy crashed. You are in the wrong Forum, bud.

So what makes you a real Dem? A real Dem would not belittle DEnnis K. and believe all the Repuke bullshit propaganda regarding VNS. Crashed? You crashed.

I asked this question before, but did not get an answer from you.

Who did you vote for? Bush? I'm betting you did.

[TruthIsAll]

32. A reply to each bogus statement....

That really hurts coming from someone out of touch with reality enough to support Dennis Kucinich.

- Very nice sentiment. Who do you support? Bush?

Partisan pollsters have to worry about helping their clients win, and if you want to win you need the most accurate picture of the situation you can get.

- Partisan posters are paid to be partisan and are paid accordingly.

Independent pollsters have to worry about getting clients and the more accurate the more clients you will get.

- Independent posters (like Zogby) are more inclined to poll fairly.

Media pollsters have to worry about selling newspaper/getting viewers and the more accurate you are the more you get.

- Media pollsters (like WP) shill for their candidate by inflating his numbers.


Go read about the history of exit polling and see what the lengths the media outlets were going to be the first and the most accurate.

-Exit polling has been very accurate. That's why VNS decided not to show us the numbers in 2002. VNS is a consortium of the Repuke media (WP,NYT,WSJ, etc.) VNS did what they were told - don't produce the true results which would only expose the massive fraud.

[leftofthedial]

36. the PURPOSE of most polling is bias

Are there straight up polls? Of course. Are most polls straight up? Hell no.

Push polling predominates in today's political polling.

Polling is specifically a component of the cold war era disinformation campaigns that Reagan turned against the public.

Polls, by and large, are tools of propagandists, not some objective scientific search for truth.

Market research has a vested interest in finding what people really think. Since Goebbels, political polling's vested interest and primary function is in manipulating opinion.

[UpstateNYDem]

38. ::Smacks forehead::

It is clear you know nothing about polling, so just stop posting on it. I'm Dennis is glad to have looneys like you backing him.

[leftofthedial]

40. next time use a hammer

[leftofthedial]

39. How is supporting any candidate during the primary "out of touch"?

you DO understand what primaries are, don't you?

Partisan pollsters have to worry about helping their clients win, and if you want to win you need the most accurate picture of the situation you can get.

That is not the primary purpose of political polling anymore. Successful (that is, neocon) candidates use polls to spin issues, to uncover points that can be spun on issues, and for damage control. They also use them to frame the debate. Third, they use them to counter the opposition's polling results. They are at best marginally useful in determining actual widespread public opinion.

Independent pollsters have to worry about getting clients and the more accurate the more clients you will get.

Fiirst, almost no political pollsters are "independent." Second, they make their money by accomplishing the functions described above, not by discovering the truth. Third, they get hired mostly to tell their clients what they want to hear.

Media pollsters have to worry about selling newspaper/getting viewers and the more accurate you are the more you get.

Not true. The more your poll reflects the biases of the people who read newspapers, the more you get. It's entertainment, not journalism.

Go read about the history of exit polling and see what the lengths the media outlets were going to be the first and the most accurate.

Objective exit polling WAS an accurate early read on election results. There were no exit polls in 2000. This is a tactical, isolated form of polling that is intended to predict an outcome that will be verified factually within hours. Yes, they work hard to make it accurate. This has almost nothing to do with the conventional practice of political polling day in and day out.

[GOPBasher]

14. Yes, very true

Zogby has four categories: excellent, good, fair, poor. Now I don't know why he does that, because the word "fair" is pretty subjective. Many people who answer "fair" would say they "approve" of his performance, while many who give the same answer would say they "disapprove." I think any presidential approval poll should only have these choices: approve, disapprove, don't know/no answer.

[Evanstondem]

17. Any opinion choice is subjective

Zogby's response categories are atypical for political polling, and I agree this makes it more difficult to compare his data to results from other polling firms.

However, there is nothing intrinsically wrong with his response categories, and they are much more widely used than approve/disapprove in the more general field of market research. Based on my experience, with scales of this type, I would expect most people who rate Bush as "fair" to also "disapprove" of his performance.

[diplomats]

2. Zogby is the only one who polls likely voters

However, I must admit I found it little bit odd that so many recent polls had Bush at exactly 52 percent. I don't ever recall that many at one time coaalescing around the same number. I hope that's not his "bottom" figure.

[DemocratSinceBirth]

23. The Closer You Get To Actual Elections

the more pollsters poll likely voters...

And determining who is a likely poster is a difficult prospect because folks lie..... Nobody wants to say they don't vote or aren't registered...

That's why pollsters ask several questions such as where do you vote, , who did you vote for in the previous election, etcetera...

[ButterflyBlood]

6. I'd hate to rain on everyone's parade, but

the reason Zogby's are so difference is because he uses a different format. Rather than asking if they approve or disapprove, he puts Bush in 4 categories, excellent, good, fair and poor, and makes the approval and dissaproval ratings by lumping excellent and good together, and fair and poor together. Only problem is fair doesn't neccesarily mean "disaprove". So I'll guess that Bush's real approval ratings are probably around 50%, give or take a few points either side. Which still isn't much more than voted for him remember, and it can't get any better.

of course, I bet a lot of people who say fair that would've said approve if asked in that format are the people who don't really approve of him, but are scared in this day and don't want to think someone incompetant is in the White House, so they stick their head in the sand and pretend Bush is doing an "OK" job, but when election day comes around they'll drop him like a hot potato.

[Evanstondem]

8. "Fair" is a negative rating

I work in marketing research for a living, and when we use similar response categories we always tell our clients that being descibed as "fair" is bad news.

[ThoughtCriminal]

12. A problem with sampling

On occasion, a poll will also ask "Who did you vote for in 2000?". Typically this will show 60-70% claim to have voted for Bush. Since we know less than 50% actually did, this means one of two things:

1. The poll sampling is seriously flawed and overwhelmingly favors Bush supporters.

2. People lie. This is a well known phenomena - claiming to support the winner, but it may also a smoke screen for bad sampling.

So either the sampling is bad, the answers are not honest or a combination of both.

[Evanstondem]

15. Sampling Problem

Lying is possible, but non-random sampling is the more likely problem. This is especially the case with online polls; these do not reach low income adults, who normally skew Democratic.

[Deja Q]

16. No exit polling and no verified paper trail?

Whoever's behind all that is guilty of treason. Real treason. Undeniable grade-A treason. Pure unfettered treason.

[DemocratSinceBirth]

20. All The " Recognized" 00 Polls Cept The Battleground Poll And The

Rasmussen Polls were dead on in 00. They all predicted a razor close race (<3% difference) with either * or Gore winning....


Polling is an art based on scientific principles but humans aren't lab rats...

[TruthIsAll]

35. Wrong! The Rasmussen Polls were the worst..

Check their tracking..as I recall, they had Bush way ahead until the very end. Rasmussen is a Repuke...

Only Zogby got 2000 right. He had Gore winning by 1%.

[DemocratSinceBirth]

44. Reread My Post

I said all the polls cept Rasmussen and the Battleground Poll were dead on in 00.

Maybe my age betrays me but dead on = spot on to a "old" guy like me...

And Harris, Newsweek, and the New York Times predicted a "tie" or a Gore win in 00. Actually Gore won the popular vote by 7/10 of a percent. NO polling organization can get that type of specificity....

[Hippo_Tron]

34. CNN's polls = not very accurate

They just kind of gave you a good sense about who was ahead when the primary process first started and everybody was pretty much equal except for Sharpton and Braun.