Gallup Scandal. Urged to STOP Polling Immediately.

[RhodaA]

Source: Daily Kos

by pronin2 at Daily Kos

Poll guru Simon Rosenberg at NDN has been hammering away at MSM misconceptions of the 2010 midterms. Today he wrote up a piece challenging Gallup to change its model because he argues its impossible and its hard to oppose the facts presented. Gallup claims it revised its screen etc under critiques of its projection of how massively conservative the electorate they project will be Nov 2. Rosenberg says its revision makes matters worse and renders Gallup polling not credible. How? Well ...

"According to a new report by Professor Alan Abramowitz, the new Gallup likely voter model has the non-white, non-black vote at 13 percent of the 2010 electorate, and coming in at 52% Republican and 42% Democratic. Simply stated these projections are not a possible statistical outcome in the 2010 elections, and draw into question the integrity of the entire Gallup 2010 elections polling project."

Read more: http://www.dailykos.com/storyonly/2010/10/21/912204/-Gallup-Scandal.-Urged-to-STOP-Polling-Immediately.-

[hlthe2b]

1. As much as we rail against Rassmussen, Gallup has always been

very "Right"- leaning. I guess that fact has finally become so obvious it can not be ignored.

[No Elephants]

2. They both lean right. Most, if not all, the well-known polls do.

[molly77]

13. Since US elections are fixed, the polling has to be fixed to match

Remember Alvin Greene in SC? NH is especially rotten. They are a joke and don't even try to follow the law re. secured ballots. NH has a large voice in who becomes our president...read blackboxvoting archives .

[loudsue]

39. You hit the nail on the head. You can tell which way the voting machines will go by what

mainstream media says, reporting poll "results" and "projections".

[CoffeeCat]

41. So true...they're only fooling those who are asleep...

This year it's just so blatantly obvious. The media just pound the notion,
over and over and over--that the Republicans are not only going to win, but win
so enormously big that it will knock everyone's socks off.

Every damn story. It's endless on CNN and other MSM outlets. It's
really sickening.

And this obviously distorted Gallup Poll only demonstrates that their
"polls" are just more propaganda used to support how they've all
ready determined the voting machines behave.

And...they're doing their best to suppress and depress the Democrats into
voting. Why vote Democrats, when you know you're going to lose?!

The media is supposed to be fair and impartial. They should not be
spending hours and hours of reporting this meme, because it can affect
elections and cause voter apathy in Democrats. They should not be blasting
this message every damn minute!

So yes, we get it. The powers that be want Republicans wins this time around.

So, they'll mess with the voting machines, disseminate ridiculously fake polls
and use the media to pound it into Democrats that you might as well stay home
on election day. WE GET IT! WE get what you want.

These thugs are getting sloppy with their evil plots. Reminds me of Dr. Doofenshmirtz...

[TheWraith]

44. So you believe every single poll is fixed according to a massive conspiracy?

Including the internal polls conducted by Democrats?

[Life Long Dem]

48. Maybe this is a reason why Obama and Dems are always telling us to vote early.

Early voting is definitely not matching the polls. Dems are winning in early voting from what I see.

[Larry Ogg]

55. Yup...

Long before the voting booths open, like some religious alter, the TV will have given most of the voters all the lies, propaganda and psycho babble they will need to cut their own throats on election day - while preserving the status quo.

[aquart]

56. This is a completely useless, defeatist outlook.

What does anyone gain by spreading this hysteria?

[madmax]

61. Bingo! Been saying this all along. eom

[PSzymeczek]

53. It's probably because

he was raised Mennonite. I'm basing this on my father's side of the family.

[Botany]

3. George Gallup has long been dead and the company that does the polling ...

.... and has his name has quit the science of accurate polling that he helped to start
many years ago.

edit .... no coffee yet

[druidity33]

4. Welcome to DU!

I notice this is your second LBN post to hit the Greatest Page...

You're ON IT!

Thanks for the quick links to today's stories... and welcome to DU!

A K and an R for you...

:)

[RhodaA]

27. Thank you for all the welcomes

I've been reading DU for a long time, but just recently starting posting. It's a nice play to be !!

[drm604]

5. A setup for election theft?

If they rig the polls also than the results will be less suspicious.

[live love laugh]

22. Bingo. nt

[Nothing Without Hope]

28. Exactly. A standard part of the operation. n/t

[SoCalDem]

31. and once the long-trusted exit polling was trashed after 2K, there's NO

checking system in place to point out the oddities..

We USE exit polling to keep elections fair in foreign places where we "help" them run their elections, but somehow here in the good ole YooooSA, we cannot figure out how to do it anymore..:grr:

Sketchy, skewed polls, ballyhooed endlessly on corporate tv+no exit polling+weirdo vapor voting= no one really knows who actually won

[Kweli4Real]

32. Or, alternatively ...

If they rig the polls ... the rightwings' heads will explode (i.e., be whipped up with the prospect of voter-fraud to "revolt"), when democrats suffer few, if any, losses.

[kestrel91316]

54. Exactly.

[nradisic]

6. Great post...

Welcome to DU :)

[indimuse]

KNR! n/t

[indimuse]

7. KNR! n/t

[pnorman]

10. Me too!

[Bluenorthwest]

8. Polls are poison and always so in Sound Bite Land

They exist only to promote the media narratives, to provide seeming dynamics to the months of coverage they insist upon for each cycle. Polls also serve the pundits and the ad sales people, they give 'consultants' (a profession that always has more failure than success) 'facts' to point at as they pitch their schemes to their hapless clients.
When looked at in great detail, a precious few polls can actually tell a person a few things. A few of them, when taken in detail. That's it. Polls help Nate buy things, help Katie have something to talk about, they help others sell 'plans' to Meg.

[Dover]

9. I've been questioning polls for quite awhile on DU as regards their use in manipulating

perceptions and opinions. Our entire voting system is in question, and the best secondary gauge - the exit poll - has also been tampered with. Controlling information, perception, the message, is the name of the game and so I've just assumed that it too has been compromised along with corporate owned media outlets and voting machines.
If they can convince enough people that the minority's opinions, candidate, etc. are in fact the majority, they can in essense custom-create reality.

And quite honestly, I haven't seen much impetus from Repubs OR Dems to question or correct it.

[Beacool]

11. Many may not like Rasmussen, but they are far more accurate than Gallup.

I find that Rasmussen is usually on the money.

:shrug:

[dionysus]

29. got somethin for ya bea!

:hi:

[droidamus2]

12. Always questioned

Even after taking a statistics class in school or maybe because of it I always questioned the results of polls. The one thing that just didn't seem to sit well with me is the idea that if you poll your 'random sample' and the distribution of a particular group, in this case non-white, non-black voters, didn't match your 'expected' distribution the pollster would then 'adjust' the numbers to reflect their vision of the correct distribution. I don't see anyway this can be done without personal bias skewing the results. Since many of the 'adjustments' are based on past elections and polling this also means polls don't normally pickup on large changes in the electorate until after the change has become obvious. The best you can hope for with polls is to look at a lot of them and sort of do your own averaging by getting rid of the outriders (those that are just too wrong to be believed) and doing a little studying on the remaining pollsters methods. One thing my wife and I have noticed in the realm of non-scientific polling, in this case the CNN online polls, that it seems everytime the question is in the area of a conservative vs. liberal question what we perceive as the liberal position ususally wins in the realm of 60-65 for the liberal position verses 30-35 conservative (or 35-40 if there is no third option). Along with this there seems to be a core of about 20-24% that are absolute Republican/Conservative that will vote the supposed conservative way even if their answer to a particular goes against their early positions.

[Igel]

23. And those are the right questions to ask.

If you don't know the assumptions used in construction of the poll, if you don't know how the actual demographics of the respondents "map" onto the expected demographics of the voters, if you don't know the language used and the ordering of the questions, if you don't know the sources of error and how to account for them, the poll's just a meaningless claim.

For example, if you construct a poll that rotates the questions, manages to make them absolutely neutral for both (D) and (R), manage to nail the model so that those actually voting are perfectly accounted for (by gender, party affiliation, age, race), you can still mess up if your (D) are all from Louisiana and your (R) are all from Massachusetts (vs. all the (D) being from Mass. and all the (R) from Louisiana). The results can still be skewed if all your Latinos are Cubans or Central Americans.

Error in the model, how you adjust polls, is reduced once the precinct election results start coming in. You find out how your sampled precincts actually turned out and that can provide one kind of adjustment to your exit polls. You also find out how they voted, so you can do a regression to find out how groups voted and adjust your model for that.

Most polls are built on projecting the past into the future while trying to make some guess as to what's changed. It's a fool's game, to be honest, and there have been some real cock-ups in electioning polling history.

[TexasObserver]

14. Pollsters work for the same people major media do.

They're hired guns who make their living doing the bidding of the same powerful interests who support the RNC and the likes of Fox news.

Gallup hasn't been trustworthy for years. They "build to suit" those who hire them.

[Warpy]

15. The only way those numbers hold up is if Gallup only polled

Batista Cubans over 70. Even then, I'd tend to take them with a pound of salt.

[ProudToBeBlueInRhody]

16. Someone needs to explain to me "internal polling"

How is it that every candidate seems to have a lead in their own "internal polling"? What is the purpose of this bullshit? To make someone feel good?

[phleshdef]

24. No. And I'd trust most internal polling more often than independent polling.

Internal polling is likely to be more accurate because the campaigns need to know how they are really doing and which areas they might need to hit harder. They do their own internal polls because its the smart thing to do if you are a candidate that actually wants to monitor your own progress.

[jeff47]

42. Because saying "We're gonna get crushed" doesn't help your campaign. (nt)

[fascisthunter]

17. Rasmussen should be next

[creon]

18. Something is not right.

This polling data does not fit with my reading. I see, with a pretty clear eye, how these political matters work out. I do not see a "tsunami" or "Cat4 Hurricane" for the GOP. If people get out to vote, that will not occur.

[robcon]

19. What nonsense. Gallup is very reliable.

n/t

[SkyDaddy7]

33. Nate Silver is basically saying the same thing as Gallup...

I wonder how many here on DU think he is also a mouth of the Right?

October 20, 2010, 8:52 pm G.O.P. Odds of House Majority Now 3-in-4
By NATE SILVER
The latest FiveThirtyEight projection for the House of Representatives shows little overall change from our previous update, released this weekend, but the Republican position has improved slightly. They are now given a 75 percent chance of winning the House on Nov. 2, up from 73 percent previously. During an average simulation run, the Republicans finished with a total of 228 seats (up from 227): this would reflect a net gain of 49 seats from their current position.

As I have warned repeatedly in the past, we believe that the uncertainty in the forecast is intrinsically quite high, stemming from the unusually large number of seats in play, and from differences of opinion among pollsters in how to calibrate their likely voter models to account for the so-called “enthusiasm gap.” There are only 170 seats that the model thinks of as “safe” Republican — those where their chances of winning are 95 percent or higher. However, there are only 151 seats that the Democrats are at least 95 percent assured of winning.

The slight movement toward Republicans this week is not the result of shifts in the polling. Our estimate of the generic ballot remains unchanged, showing about 6 point lead for Republicans among likely voters. And some of the local polling has been decent enough for Democrats in the past few days, like a nonpartisan survey showing their incumbents ahead in two tight races in Michigan, and a poll suggesting that Ben Quayle, a Republican, could lose in Arizona’s 3rd Congressional District — although our model is skeptical and still gives Mr. Quayle an 89 percent chance of prevailing.
Some polls can also be prone to misinterpretation by those who lose sight of their context. The large batch of polls released by The Hill this week shows many Democratic sophomores trailing. Clearly, this is not good news for Democrats — but in most cases, these sophomores are in difficult districts and had already appeared likely to lose in a universe in which Republicans were poised to gain 50 or so seats.

But there are other indicators that have unambiguously broken Republicans’ way this week. Third-quarter fundraising reports trickled in over the weekend, and Republicans bettered Democrats there — and not just because of contributions from outside groups that some Democrats have been complaining about. Instead, aggregate individual contributions for candidates who will appear on the House ballot in November totaled $74.3 million for Republicans, versus $54.3 million for Democrats. Although Democratic candidates generally had the stronger fundraising numbers early in the cycle, and in some cases have more money in the bank, this is nevertheless an auspicious figure for Republicans, particularly since they have fewer incumbents and fundraising can be a more cumbersome task for challengers.

Meanwhile, Cook Political again downgraded their rating for Democrats in a number of key seats. Dramatically, for instance, they now regard Raul M. Grijalva’s race in Arizona’s 7th Congressional District as being a tossup; whereas they had scored the race as safely Democratic mere weeks ago. (Our model now thinks that Mr. Grijalva has a 21 percent chance of losing his seat.)

For the time being, we are still in a universe where Democrats could probably hold the House by having the coin come up heads in a sufficient number of tossup races.

We may not be far from the point, however, where their chances would boil down, in essence, to there being systemic errors in the polls, which could potentially affect a large number of races — or there being some sort of last-minute change in the macro environment.

.

[wowimthere]

38. I think he is getting PAID. That's the reality. When the price is right people will say anything

that sounds reasonable. The media wants the House to go Red and that's just how it is. They also don't mind if this president fails because it makes him much more credible than many other past presidents. He's done more in two years than any president before him and we're supposed to believe that he's almost as bad as Bush. Give me a break!

[jeff47]

43. Nate isn't doing his own poling

He's using data from other pollsters. If they're wrong, Nate's analysis is wrong.

[OHdem10]

20. What is more pernicious---How the Media uses the polls to change

Public Opinion. The 24/7 breathless announcements telling
the People over and over--The Republicans are going to win
so why don't you Democrats just give up now and for Gosh sakes
do not vote. This is MSM and Fox 24/7 Message.

Polls are used to sell ideas.

[Igel]

25. I must say, I have no anecdotal evidence to back that assertion up.

The best that I can come up with is when an Obama supporter that was thinking of voting for Hillary so as to not "throw away my vote' a couple of years ago decided to vote for Obama when polls showed that he was probably going to win. Personally, I think he'd have voted for Obama anyway.

On the other hand, I'd point out that despite polls showing Obama was down early his poll numbers picked up. Now, were those fraudulent numbers intended to sell Obama, to get us to vote for Obama? Or was that the result of a lot of people ignoring poll resuls and making up their own minds?

The first alternative says that Obama was installed in a kind of poll-related coup, foisted on the American populace because the MSM and pollsters so wanted him in power. The second says that poll numbers matter, at best, around the edges.

You have a hypothesis. The prediction's entailed. It has to be the former.

I'd have to go with the latter, although I think that "at best" cannot be over-emphasized.

[Rozlee]

49. The operative word is "anecdotal"

That's been my observation in what they're calling the non-white, non-black population. My sister is a big time winger, the only one in our large Hispanic family. But, she's sitting this election out because of the anti-immigration bias and racism that's become so blatant in the Republican party the last two years. She used to be an avid Fox viewer and right wing talk show listener. She basically just started with an anti-abortion stance then got sucked into the hard right by the tangential groups that associate with it. She's still right-wing, don't get me wrong. The Kool-aid's too strong. But, the teabaggers ruined the Republicans' previous attempts at "wooing" Hispanics with their hardline, in-your-face bigotry. And I can't believe that my sister is an isolated case. She was never a casual voter. You can thank her for helping to vote in crazy creationists to the Texas State Board of Education. But, if Faux has been losing almost 29% of it's viewers, some of them are probably Hispanics that are repulsed by their anti-immigration stance like she is. And polls tend to drop many Hispanics, esp. the younger or poorer ones as "unlikely" voters. Like I said, it's anecdotal. But, I can't believe it's isolated. There's got to be a lot of resentment in Nevada against Sharron Angle in the Hispanic community right now. I'm hoping it's an anger that will vote instead of just stay home and stew in it's juices because they lack "enthusiasm" as the pollsters say.

[deutsey]

65. That's a well-worn psyops tactic

According to David Halberstam in The Fifties, the CIA "deftly created a fictional war over the airwaves, one in
which the government troops faltered and refused to fight and in which the liberation troops were relentlessly moving toward Guatemala City."

According to the Church Report, in Chile leading up to the '73 coup "press placements were attractive because each
placement might produce a multiplier effect, being picked up and replayed by media outlets other than the one in which it originally came out."

[femrap]

21. The MSM are trying to

'create reality' by simply stating that the election is over and the GOP won. It's a great way to discourage democratic voters.

If I read one more article about the 'new speaker of the house,' I'm gonna :puke:

Get out and vote....fuck the pollsters. They've been lying for years now. I swear they just make up numbers so TPTB are happy.

[Scurrilous]

26. K & R

:thumbsup:

[Kweli4Real]

30. My, admittedly naive, question would be ...

unless you are a candidate (or working/planning a campaign) or you just want to vote for the "winner", what is the relevance of a pre-election poll?

The poll numbers are NOT going to influence my vote one bit.

[soryang]

34. I saw a Gallup geek on CSPAN and noticed this..

Every statement made by callers was "consistent" with his findings or experience no matter what view was expressed. In other words he could harmonize any view with his ostensibly Solomonlike objectivity.

After he characterized a finding of strong support of a (democratic policy) 31 percent, strongly oppose 34 percent, somewhat support 32 percent, he drew the conclusion that most people likely to vote oppose the policy and that the democrats have lost substantial ground on this issue.

The only thing I could infer was his bias.

[mzmolly]

35. How long have they contributed to the "Democrats are in for a slaughter" meme

now? :grr:

[wowimthere]

36. SLAM!!! I don't know any Democrat that isn't voting! Polls can say what you want it to say nt

[LiberalFighter]

37. If polls require a model for them to use that reflects actual outcome

how does one that do it if they don't really know if it will 52% Republican and 42% Democratic or 49% Republican and 51% Democratic? Or that there will be a higher voter turnout of non-whites than normal.

[ProudDad]

40. Believe polls at your peril...

The "questions" can be manipulated to get whatever results you desire...

The "Sampling" algorithm can be jiggered to get whatever results you desire...(as in the OP)

Many of the most Progressive voters don't have land line telephones and so, don't get polled...

It's a shame that so many are "convinced" by the polls to go with the "winner"...

It's an even bigger shame that we don't have a democracy in the USAmerikan Empire...

[ProfessorGAC]

45. What A Shock!

Their methodology has been the subject of academic criticsm among statisticians for DECADES! They have never gotten the science down.

Their very hiring criteria is suspect.
GAC

[Beetwasher]

50. Yup, I remember a particularly egregious piece of garbage they did in IRAQ

When we invaded. They polled Iraqis and found a majority supported the US invasion.

[hay rick]

46. Another welcome to DU.

Thanks for the link to a good article. Contemporary polling is about as credible as oil industry-sponsored climate science.

Thanks also for the excellent post on Rick Perry- which I missed originally, but saw in your journal.

[emsimon33]

47. Gallup has no integrity! Check out their Teacher Insight assessment

They will not re;ease results to those who take the assessment and will not let others examine their technical manual--if there even is one!

These people are fakes and little better than Faux News polls!

[SnakeEyes]

51. Actually Nate Silver

has Gallup ranked more accurate than Rasmussen

http://www.fivethirtyeight.com/search/label/pollster%20ratings

[Aslanspal]

52. Polling today is a wreck and in chaos

What about polling people on wifi or cell??? can you really get those and let me tell you landlines are the way of the dinosaur.

I think polling today is not credible, somebody has to either capture new communications(cell/wifi) or go door to door.

Gallups day is dead and gone.

[madmax]

62. Do away with the damn polls

The only one that matters is on election day.

Stop trying to manipulate the herd mentality. Yes, it works for those who don't really pay attention or those who aren't sure of their own damn convictions, or are too stupid to live.

[democracy1st]

57. K & R

[tiptoe]

58. Richard Charnin exposed the Polling Scam first. 10/15: Check out Gallup's "Unlikely Voter" Dem share

10/15 Richard Charnin Midterms Forecast: MSM's Likely Voter Polls Preparing the Cover for GOP Fraud

Stay Tuned for Charnin's 10/22 Update. The Pre-election Polling Scam Slouches Onwards.

[anAustralianobserver]

59. I'd like to see a Michael Moore exposé on polling and election fraud. (nt)

[eagertolearn]

60. I lost trust in any poll after the 2000 and 2004 elections.

[Blue_In_AK]

63. The best advice is to ignore polls altogether.

They are nothing but voter suppression tools and certainly serve to manipulate elections results. I'm pretty sick of them myself.

That last CNN/Time poll that showed Lisa Murkowski and Joe Miller tied and Scott McAdams back about 7 points, the pollsters actually asked people if they were going to vote for Joe Miller, Lisa Murkowski, or Scott McAdams -- even though Lisa's name is not on the ballot. Talk about skewing a result! I believe I heard that there was about a 25% difference between polls taken with her name mentioned and polls mentioning a generic write-in.

And so Lisa's people use these skewed polls to scare moderates into voting for her to stop Evil Joe. They keep insisting that a vote for McAdams is a vote for Joe Miller. I keep saying that a vote for Scott McAdams is a vote for Scott McAdams, and since I can't be true to my ideals by voting for either Joe or Lisa, Scott it will be.

This election up here is going to be insane unless Scott McAdams wins outright. Joe Miller is going to contest every single one of Lisa Murkowski's write-in votes, and the Democratic Party is already threatening to sue if certain polling station "helpfulness" with the write-ins (as in pointing out the spot and saying her name) isn't cleared up soon. The rest of us can't wear political buttons in the polling booth -- they can't show people what name to put where.

[Blue_In_AK]

64. Sorry, dupe.