Gallup defends results against MoveOn.org attack -USAToday

[Rose Siding]

snip>
USA TODAY said Tuesday it remains confident in Gallup.

The firm, said USA TODAY Polling Editor Jim Norman, "has been doing USA TODAY's political polling since 1988. ... I'm impressed by the constant testing and retesting they do of their methodology to make sure they get it right."

CNN spokesmen were unavailable for comment.

In recent weeks, complaints about the USA TODAY/CNN/Gallup polls had mostly been aired in "blogs," or online diaries, on the Internet. Tuesday, the issue spilled into the "mainstream media." MoveOn.org, a liberal advocacy group, paid $68,000 to run a full-page ad in The New York Times. The ad's headline: "Gallup-ing to the Right. Why does America's top pollster keep getting it wrong?"
.....
Critics say the debate over Gallup's work is important because the media's reporting of polls can affect the dynamics of a campaign. "We need the most accurate information possible. Next week the stories could be 'Kerry's surging in the polls,' but would that be true?" asks Markos Moulitsas Zniga, who's dailykos.com blog is a popular site for liberals..........lots more........

http://www.usatoday.com/news/politicselections/nation/president/2004-09-28-gallup-defense_x.htm

[RoyGBiv]

1. Well, he's easily impressed ...

Personally, if I review someone's work and realize that they are continually testing and retesting their methods and still coming up with screwed up results, their ass is gone.

[qanda]

2. It's quite obvious from the chart they have there

that the only time Bush has an advantage is when they poll more Republicans than Democrats.

[yardwork]

6. LOL! I just clicked on the chart

The difference between the two candidates is always exactly the difference between the percentage of Dems and pukes in the sample! In every case!

ROTFL!

[qanda]

8. I really don't see how they can defend this

And I remember when one poll was skewed the other way and the Republicans about had a hissy fit. I believe ever since the Republicans complained it's been the other way around. I'm going to find an article on that.

[jdj]

7. yes, I've read elsewhere this is what they are doing

they are sampling larger numbers of republicans than are predicted to vote in Nov, therefore skewing the results to the right. "they" meaning all the pollsters, not just gallup.

[yardwork]

3. I'm so glad that Moveon did this

Put the Repukes on the defensive for a change.

[featherman]

4. make the bastards deny it

Hardly a hard hitting article, though.

[AlCzervik]

9. usa today uses gallup as well so i'm not shocked

they didn't indict gallup but i'm really glad this is at least getting some coverage. Zogby has been the closet in the last couple of election, more "News" outlets should use them once and awhile.

[llmart]

26. Have you ever seen USA Today......

do a hard-hitting article? They used to be called the McDonald's of newspapers and rightfully so. They're a piece of crap that isn't even good enough to line the birdcage with. However, it's the sort of crap the average American reads and then thinks themselves knowledgeable.

[Bill McBlueState]

34. hold on, hold on

USA Today goes very nicely under our cat's box to sop up all her "mistakes." Don't slander that paper!

[llmart]

39. Well, I'm sure hoping you aren't paying.....

any of your own money for that "liner" under the litter box:)

[K. F. Gibbons]

5. According to a recent Gallup poll, the American people highly

favor there method of given Bush a reach around.

[Media_Lies_Daily]

10. This article tells me a lot about USA TODAY Polling Editor Jim Norman...

...and it's not a very pretty story.

[enough]

11. Here's CNN-Gallop poll on OCTOBER 20, 2000.

Found on this thread:
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x912337

[Media_Lies_Daily]

16. Bingo! Nice find!

[Jack Rabbit]

28. Enough said

I'd say the actual results were outside the margin of error.

[Mr_Spock]

47. Holy mackerel!!! Gallop is a friggin JOKE!!

Never, ever for the rest of my life will I trust their polls as anything more or less than just simple propaganda!!!!

[RevolutionStartsNow]

12. The silver lining in all of this...

is that news organizations are starting to pay attention to the "buzz" on the Internet, including in large part this site and DailyKos, and other major sites. Of course that means they also pick up on the tripe on RW sites, but I am glad that all this typetypetyping and incredible hard work that people do here to investigate and broadcast the truth is starting to get through...

[LastDemStanding]

13. Didn't Gallup have Kerry ahead prior to the last few weeks?

Was there any moveon.org complaint or other complaint during that time?

[LisaL]

14. Hah?

Whatever gave you that idea?:eyes:

[murielm99]

15. Why don't you check that out and report back to us?

Try Google.

[Manky George]

19. In the face of ignorance...speculate!!!

As a UK political researcher I declare every wish that the US ceases humiliating itself by electing idiots like Bush. But reading these pitiful attempts to challenge the integrity of the Gallup poll is not going to help you. You all miss a fundamental point in polling methodology - every single poll has more or less of each political group sampled. But this is then factored in the projected figures - work it out...it's not a straight line extrapolation.

Let me put it another way. If polling companies simply sampled specifics voters for each party, wouldn't the results exactly replicate the percentages chosen? This isn't happening here. And the evidence? The polls keep fluctuating despite consistent sampling.

This isn't an easy message to take but it might just be that John Kerry is less popular that Bush right now - that is what they are telling every polling organization. That's unlikely to be a view that's prevalent on this site for the obvious reason that we all want Kerry to win.

Oh, finally, the conspiracy notion is not credible...it really isn't. However, I'm always open to incontrovertible proof as opposed to outraged posturing. Let's stick to the facts.

[Diogenes2]

21. The voters are NOT telling every polling organization

that they prefer Bush. That is simply not correct. And yes, Gallup IS skewing their results by using sampling trickery, projecting voter turnout in Republicans' favor, always choosing "likely" voters (a nebulous concept at best), etc. Nice try at being a pundit, though.

[Carolab]

23. The proof is out there

you just haven't looked hard enough. Lots of credible persons have been researching this a very, very long time and one of the key factors now coming out, based on the outright lies told in the 9/11 Commission's "whitewash" regarding the NORAD/FAA "timelines". Senator Mark Dayton from Minnesota brought this very fact up to the "commissioners" during the DNC. I watched it myself on C-Span. NOBODY disputed what he said. The testimonies are contradicted by the facts. The "drills" on that day are contained in Peter Lance's new book "9/11 Commission Cover-Up" and also in Mike Ruppert's "Crossing the Rubicon".

[Ediacara]

36. I don't think you understand what's going on....

Gallup and the pollsters for Time and Newsweek overweight their sample in favor of Republicans.

They have weighted Republicans at 40% of likely voters, and Dems at 32% of likely voters.

As an example, say they have 414 Republicans they got on the phone, 89% want Bush, they have 528 Democrats, 86% want Kerry, and they have 451 Independents, 53% want Kerry, but they WEIGH the data based on skewed numbers:

Republicans: 89 x 0.40 = 35.6% Bush, 11 x 0.40 = 4.4% Kerry
Democrats: 14 x 0.32 = 4.5% Bush, 86 x 0.32 = 27.5% Kerry
Independents: 47 x 0.28 = 12.2% Bush, 53 x 0.28 = 14.8% Kerry
Bush 52.3, Kerry 46.7

Even though a fair assessment of the same sample would yeild something like:

Republicans: 89 x 0.33 = 29.4% Bush, 11 x 0.33 = 3.3% Kerry
Democrats: 14 x 0.34 = 4.8% Bush, 86 x 0.34 = 29.2% Kerry
Independents: 47 x 0.33 = 15.5% Bush, 53 x 0.33 = 17.5% Kerry
Bush 49.7, Kerry 50.0

And just adding up the raw numbers from their sample, they get 53.1% for Kerry!

Gallup's assumption is completely wrong based on history (Dems are a larger percentage at the polls), and common sense (Dems hate Bush passionately, and will come out in huge numbers).

Those polls that actually weight their sample based on history and common sense, or just plain fairness (about 1/3 each to Republicans, Democrats, and Independents), show that Kerry and Bush are W/I the MOE and Kerry is gaining.

[Terran]

40. Pray tell, then

How do you explain Gallup's egregious 12 percentage point error in polling in 2000, as shown in post #11? Eh? And only two weeks before the election. Rather far outside the margin of error.

[funnymanpants]

49. You are giving us a lot of BS

You are not a political researcher, or you wouldn't have written such a stupid post. For example, you write "I declare every wish that the US ceases humiliating itself by electing idiots like Bush." How can you declare every wish? Such poor phrasing doesn't indicate a real political researcher, who would have a better grasp of the English language.

But then when you try to explain the obvious results of the poll, you are completely incoherent. "But this is then factored in the projected figures - work it out...it's not a straight line extrapolation." I am sorry, but this explanation is pure obfuscation. It makes no sense. As a former college instructor, I can tell when someone really knows what he is talking about and when someone throws around jargon to cover up their ignorance, as you are obviously doing. What type of extrapolation is done? How can you correctly extrapolate if you have a faulty data set to begin with?

The rest of your post is almost humorous.

"This isn't an easy message to take but it might just be that John Kerry is less popular that Bush right now - that is what they are telling every polling organization."

You have a faulty pronoun reference in this sentence. You switched subjects right in the middle--not exactly the sign of a literate political researcher. (And just what is a political researcher, anyway?) Further, you don't cite any polls to back up your point. Why not? Because if you had, you would realize that you contradicted yourself. The Zogby poll (which was the most accurate in the last election) has both candidates at a dead heat. So do other reputable poll.

Last, you sign off with the stupid signature of all dumb conservatives. "Let's just stick to the facts." Of course you didn't use any facts in your whole argument, but that is typical of conservatives who are too stupid and smug to realize that they don't have the facts on their sided.

Translation: you are an obvious fraud.

[Media_Lies_Daily]

18. That would be true only in "Reverse World".

[fearnobush]

20. Ahh,,, No. They gave him the only negative bounce of all the pollsters

after the DNC in Boston. All of your media, CNN, MSRNC, FAUX, ABC, CBS... all focused 95% of their poll coverage after the convention on that one skewed poll. Mean while, Zogby gave Kerry a 7pt bounce, all the other polls ... ALL OF THEM, gave Kerry a 3 to 6 point bounce. None of your whore media made mention of it. SO, We tell all people, at work, at parties, bars etc that the USA media is clearly bias to the GOP and that the Republicans pretend it is bias the other way.

[nine30]

38. Maybe..

..but then they decided that time was running out for helping their candidate. When all the polls are showing that * has a razor thin edge over Kerry, they figured they couldn't be entirely wrong for great exaggerating the margin.

[Mr_Spock]

48. Here lies a freeper!

[Nlighten1]

17. Anyone have the link to the MoveOn ad?

[high density]

22. Yup

http://www.moveon.org/content/pdfs/Final-Gallup-Ad.pdf

[Supersedeas]

33. high den - ty ty for the link to the ad

[rfrrfrrfr]

24. Take a look at the table they show in the article.

It explains why their polls show bush is ahead. Their sample is skewed with more republicans and independants then there really are. Talk about a smoke screen. They try and deny problems in their methodology and the eveidence showing they are wrong is posted right in the dang article.

[bringbackfdr]

45. Gallup simply proves an old truth

You can make numbers dance, sing and tell any tale you'd like them to tell. Gallup is a master at it - shameless hacks like CNN and USA Today pick them up and run them as gospel because they're too damned lazy to check them out.

[DeepModem Mom]

25. Their own chart proves MoveOn's case!

[Child_Of_Isis]

46. The chart is no longer there (eom)

[llmart]

27. Gotta love MoveOn!!!

Time to ante up some more donations to them. Every $5 or $10 helps get our message out.

[robbedvoter]

29. Was there a chart linked to the article? If so, it's gone!

[qanda]

30. It's still there

The moved it to another page, I guess because it was screwing up the formatting.

http://www.usatoday.com/news/politicselections/nation/president/2004-09-29-gop-poll.htm

[sidpleasant]

31. Re: "likely voters"

A coworker told me an intersting story a few days ago. Her son's best friend got a call from Gallup. This guy is 30 years old and has never voted before but he registered this year so he can vote for Kerry. He's absolutely positively going to vote on November 2nd. He told all this to the Gallup pollster who then replied that he didn't meet their definition of a "likely voter" and hung up on him.

[leftchick]

41. wow... that is some story!

unfreaking believable! :grr:


and welcome to DU!

[wishlist]

32. USA Today ran headline about Bush getting bounce from Dem convention

In contradiction to most poll results that gave Kerry at least a slight bounce, USA Today's front page on the Monday after the Dem convention said "Bush not Kerry gets bounce"! So they have no credibility on this issue.

[henslee]

35. Love it when the heavies are playing defense! Sumthins in the air!

[progressivebydesign]

37. They're over sampling Independents, too..

They do NOT make up almost 1/3 of the voters anymore. ANd.. they skew republican, because they are left over Perot-ites.

[MallRat]

42. Just look at the last 8 polls. Stunning.

July 8-11: GOP bias 0, poll Kerry +4
July 19-21: GOP bias +6, poll Kerry +2
July 30-Aug.: GOP bias +8, poll Bush +4
Aug. 9-11: GOP bias +6, poll Bush +3
Aug. 23-25: GOP bias +6, poll Bush +3
Sept. 3-5: GOP bias +7, poll Bush +7
Sept. 13-15: GOP bias +7, poll Bush +13
Sept. 24-26: GOP bias +12, poll Bush +8

This just in: the more GOP respondents you sample, the bigger Bush's lead gets.

The latest poll is really stunning. They oversampled GOP respondents by an additional 5 points compared to Sept. 13-15, and yet Bush's lead DROPPED by 5 points.

There is an untold story brewing here, folks. This race is tied, and Gallup doesn't want you to know it.

-MR

[daleo]

43. Yes, if you plug this into a spreadsheet and do an X-Y graph

It all becomes very obvious, what is going on.

[robbedvoter]

44. Kerry leads then if W is 8 pts ahead but was given 12% advantage

in the sample - as per their own graph. Whadda crock!