Skip to main content

Update: OK-good. It’s recommended. Now, let’s take some action. DCG provides an excellent list of talking points to use when contacting the media—or even your friends and co-workers—about the Gallup poll. Let me go ahead and add my own talking point:

--Gallup openly attempts to bias media in favor of Republicans. Gallup admits that their LV model skews Republican, yet encourages media outlets to use the Likely Voter model instead of the Registered Voter model. Media outlets interested in balance should not play along with Gallup's pro-GOP game.

Further, slinkerwink provides an excellent list of media outlets to contact. When contacting people, be calm and cool like dcg’s talking points suggest. Let’s take a bite out of polling crime.

* * * *

I am cross pointing this from MyDD in its entirety, in the hopes it will reach the recommended section. From that point, hopefully it can reach some larger media. The level of Gallup's polling corruption over the past three months is staggering and needs to be exposed.

Gallup, even without their new poll, is without question the top outlying polling organization in this election. Since they began doing state polls on the 2004 campaign, one twelve occasions Gallup has had a poll in the field for at least one day when at least one other non-partisan polling firm has had a poll in the field. On eleven of those twelve occasions, Gallup's results where the most pro-Bush of the other non-partisan operations. On the other occasion, Gallup was actually the pro-Kerry outlier: (source)
(* = three way trial heat):

Florida, 7/19-7/22
	 Bush  Kerry   Date
Gallup	  50	 46    7/22
M-D*	  48	 46    7/21
Zogby*	  48.5	 48.4  7/23
IA	  46	 46    7/20
LAT	  45	 45    7/21
R2000	  44	 49    7/19

Florida, 8/20-8/22
	 Bush  Kerry   Date
Gallup*   48   46      8/22
Zogby*	 49.0  49.6    8/21

Minnesota, 9/11-9/14
	 Bush  Kerry   Date
Gallup	 48	 46    9/14
M-D	 46	 45    9/14
Star	 41	 50    9/13

Missouri, 9/3-9/6
	 Bush  Kerry   Date
Gallup	 55	41	9/6
Rasm	 48	42	9/3
Zogby*	 48.5	48.9	9/3

North Carolina, 7/9-7/11
	 Bush  Kerry   Date
Gallup	 54	39	7/11
R2000	 49	44	7/10-4

Ohio, 9/4-7
	 Bush  Kerry   Date
Gallup	  52	 43	9/7
SUSA	  50	 47	9/8

Ohio, 8/13-9/15
	 Bush  Kerry   Date
Gallup	  45	 47    8/15
Cincy	  46	 48    8/11-7

Ohio, 7/19-7/22
	 Bush  Kerry   Date
CD	  47	 44	7/23
Zogby	  48.1	 46.8	7/23
ARG	  45	 47	7/22
Gallup	  45	 51	7/22

Pennsylvania, 9/4-9/7
	 Bush  Kerry   Date
Gallup	  48	 47	9/7
Rasm	  46	 48	9/6-8
SUSA	  47	 49	9/7-9

Pennsylvania, 8/23-8/26
	 Bush  Kerry   Date
Gallup	  48	47     8/26
Rasm	  45	49     8/26

Washington, 9/3-9/6
	 Bush  Kerry   Date
Gallup*   44	 52	9/6
Zogby*	  44.2	 52.7	9/3

Wisconsin, 9/9-9/12
	 Bush  Kerry   Date
Gallup	  52	44     9/12
Rasm	  49	47     9/12
No other polling outfit, not even Zogby or Strategic Vision, is even close to being in favor of one candidate in state polls as frequently as Gallup has favored Bush during this election. Even the one time Gallup favored Kerry, it was still an outlier.

Looking at national polls, one finds almost exactly the same story. Since late July, Gallup has, save once, been the outlying poll. Here's how the last four Gallup national polls stack up against other non-partisan polls taken during the same time period (source):

National, 7/30-8/1
	Bush	Kerry	 Date
Gallup	 51	 47	  8/1
ARG	 46	 49	  8/1
Rasm	 45	 48	  8/2
CBS	 43	 49	  8/1
News	 44	 52	  7/30

National, 8/23-9/25
	Bush	Kerry	 Date
Gallup	 50	 47	 8/25
LAT	 49	 46	 8/24
Rasm	 46	 46	 8/24-6
Time	 46	 46	 8/24-6
Fox	 44	 45	 8/25

National, 9/3-9/5
	Bush	Kerry	 Date
News	 54	 43	  9/3
Gallup	 52	 45	  9/5
Rasm	 47.6	 46.5	  9/5

National, 9/13-9/15
	Bush	Kerry	 Date
Gallup	 54	 40	 9/15
Rasm	 49.3	 44.7	 9/15
Econ	 47	 46	 9/15
Pew	 46	 46	 9/14
Harris	 47	 48	 9/13
This last one is the most profound outlier of any national poll this year, leaving Newsweek and Time in the dust (maybe that is what they were trying to accomplish). Then again, maybe they aren't leaving Time in the dust, since CNN and Time are affiliated news organizations and frequently do joint polls.

It is pathetic and unacceptable for a "non-partisan" polling firm to be produce the outlying poll in favor of Bush in fourteen of its last sixteen polls. The odds of this happening at random are around one in 14,000. Considering those odds, the far more likely explanation for all these outliers is that Gallup's polling methodology is inherently structured in favor of Bush. Whether or not it is intentional, I do not know. However, I do know that Gallup's polls are connected to the largest news outlets in America of any poll, both in terms of print (USA Today is the largest circulation newspaper in the country) and cable news (CNN has more viewers than Fox, they just watch for shorter periods of time). I also know that sensational headlines sell. I further know that Gallup's chairman is a Republican donor.

This is a shameful state for the oldest and most respected polling organiztion in the country. Shame on you Gallup. Let USA Today know that Gallup always favors Bush by large margins, and that we have the numbrs to prove it.

Originally posted to Chris Bowers on Thu Sep 16, 2004 at 09:43 PM PDT.

EMAIL TO A FRIEND X
Your Email has been sent.
You must add at least one tag to this diary before publishing it.

Add keywords that describe this diary. Separate multiple keywords with commas.
Tagging tips - Search For Tags - Browse For Tags

?

More Tagging tips:

A tag is a way to search for this diary. If someone is searching for "Barack Obama," is this a diary they'd be trying to find?

Use a person's full name, without any title. Senator Obama may become President Obama, and Michelle Obama might run for office.

If your diary covers an election or elected official, use election tags, which are generally the state abbreviation followed by the office. CA-01 is the first district House seat. CA-Sen covers both senate races. NY-GOV covers the New York governor's race.

Tags do not compound: that is, "education reform" is a completely different tag from "education". A tag like "reform" alone is probably not meaningful.

Consider if one or more of these tags fits your diary: Civil Rights, Community, Congress, Culture, Economy, Education, Elections, Energy, Environment, Health Care, International, Labor, Law, Media, Meta, National Security, Science, Transportation, or White House. If your diary is specific to a state, consider adding the state (California, Texas, etc). Keep in mind, though, that there are many wonderful and important diaries that don't fit in any of these tags. Don't worry if yours doesn't.

You can add a private note to this diary when hotlisting it:
Are you sure you want to remove this diary from your hotlist?
Are you sure you want to remove your recommendation? You can only recommend a diary once, so you will not be able to re-recommend it afterwards.
Rescue this diary, and add a note:
Are you sure you want to remove this diary from Rescue?
Choose where to republish this diary. The diary will be added to the queue for that group. Publish it from the queue to make it appear.

You must be a member of a group to use this feature.

Add a quick update to your diary without changing the diary itself:
Are you sure you want to remove this diary?
(The diary will be removed from the site and returned to your drafts for further editing.)
(The diary will be removed.)
Are you sure you want to save these changes to the published diary?

Comment Preferences

  •  we need to rip (none)
    these fuckers!!!

    http://colorpolitics.blogspot.com http://colorpolitics.blogspot.com

    by bigassbrando on Thu Sep 16, 2004 at 09:44:14 PM PDT

  •  It's shocking how Gallup is blatantly misbehaving! (none)
    Truly shocking.

    Gallup has no shame.  It's clear.

    From a scientific perspective, this is just obvious use of polling data to misinform.

    Gallup did the same thing in 2000!

    Why do we let Gallup choose the debate audience?

    Unless something is done, it will be a 60% Bush audience!

    •  We live in Alice in Wonderland World (4.00)
      The more blatant your corruption, the less likely anyone is to question it or believe it.  And Democrats, meanwhile, are so bombarded with outrages every day coming from every direction, so clear and obvious, that we don't even know where to aim first.
    •  Why Do We Let the Presidential Debates Commission (none)
      ...have anything to do with the presidential debates.

      It's high time some NON-partisan organization takes them over again (the Citizens Debate Commission had given me some hope, but I've begun to feel that they should really call themselves the Nader Debate Commission...sorry, Ralph, it's not all about you).

      In a "safe" state? Consider a vote for David Cobb, the Green Party's candidate for President.

      by GreenSooner on Fri Sep 17, 2004 at 05:36:59 AM PDT

      [ Parent ]

  •  Gallup is a right wing organization (4.00)
    Somehow, the meme has gotten out there that Gallup is the grandaddy of them all, an old, venerable polling organization.

    But after Mr. Gallup died, his name was "borrowed" by some right wingers posing as Gallupers.

    It's like 4 musicians getting together in 2004 and calling themselves "The Beatles."

    Claiming the name Gallup is just a lie.

    •  Look what happened (none)
      to Arthur Andersen.

      "I don't want to wake up on Nov 3 with George Bush still president and ask myself if there was anything more I could have done." -- Moby

      by lapis on Thu Sep 16, 2004 at 10:01:25 PM PDT

      [ Parent ]

    •  Gallup is now owned by right winger, not Frank G. (4.00)
      I now have no doubt that Gallup is a partisan pollster attempting to manipulate the election.  I think tht should just be our constant refrain.  We should write to USA Today and CNN and demand they fire their pollster. We should email MSNBC and the networks as to why does Gallup show such a persistant bias this year even when other polls show something else entirely.

      It is no longer owned and run by Frank Gallup, a man of integrity but by a right wing CEO who supports extreme right wing candidates.

      http://www.dailykos.com/comments/2004/9/16/223847/282/9#9

      Obviously Time magazine replaced their pollster Harris with the equally partisan firm of SBRC. Look at this admission intent to manipulate by Schulman of SBRC:
      http://www.dailykos.com/comments/2004/9/16/223847/282/13#13

    •  Gallup = GOP (none)
      Grand Old Poller
    •  I know that to a certain extent polls lead voters. (none)
      But if the "leading" and most widly quoted poller deliberatly sjews there results, does that not mean that on election day, the results are going to be skewed way the other way, when people vote the way they felf all along, as oppsose to the skewed samples?

      Just a thought, but it occurs that skewing polls, (unless yoou have a secret way to alter the actual vote) is a two edged sword.

      Mr Tek

      •  a certain % of voters vote for a canidate that (none)
        they percieve as strong or winning.  Thats the whole purpose - to convince people the bushie isn't a complete fuckup.

        "I just live forever, there just is no end / I just trust the oppression like I trust yr friends." (I got a) Catholic Block, by Sonic Youth

        by Demosthenes on Fri Sep 17, 2004 at 09:21:34 AM PDT

        [ Parent ]

  •  What I want to know (4.00)
    If anybody actually looked at these polls critically (of course, that would require California Dreamer to give his poll shopping lecture) is to ask Gallup to explain whether their national polling is outlying, or if its just their state polls. Because it has to be one or the other.

    Even if I accepted Gallup's LV model (and I don't, and it's really not b/c it's unfavorable to Kerry--WSJ/NBC is unfavorable also, and I think they're quite reputable), consider their LV data:

    MI: Kerry +6
    WI: Bush +8
    OH: Bush +8
    MN: Bush +2
    PA: Bush +1

    Let's suspend reality and take all those at face value and consider the Gore/Bush 2000 race as a "tie.". The ABSOLUTE WORST difference that Gallup is suggesting is in WI, where it thinks Bush is has a margin of +8 from where he was in 2000. So where the fuck of national lead of 14 points comes from is beyond me?

    Those who know me know that I'm not a big conspiracy guy. I don't think that most pollsters run around maniputaling their numbers; as discgraceful as the SCLM media, I don't really think that there's a huge conspiracy, I just think that the standards are awful and the right is better at playing the game. But the tone CNN has taken on, convenienatly coupled with Gallup going absolutely nuts basically since the end of the DNC really makes me wonder.

    Gallup's LV model is the only model that does this; magically lumps numbers onto Bush's total. Now, I don't buy their RV numbers either, but Bush +8 I could look at and at least find it within the possible realm of sampling error. +14, not a fucking chance--it's of course possible if they interviewed mostly Repubs, and maybe they did. But you won't hear Gallup asking any questions.

    As Ben P pointed noted to me last night, Gallup hasn't really pegged a presidential race since 1984. They admit, in their book, "Polling Matters" that their LV numbers skew R. Pick this book up if you want some insight into the mythical Gallup LV screen. It's based on  8 questions, that border on the absurd, including whether the respondent knows where his/her polling place is; but that's not the end of it. Once they get all that data, a computer more or less "ranks" the sample and cuts it off at 55% (for 55%). Do y'all realize how many steps away that is from the original data, no matter what it says? No other pollsters use LV screens like that--none.

    Gallup missed the Dem Convention by about 10 points compared to the average. Not even close. I'd trust a Fox poll now before I'd even look at a Gallup LV.

    Harris and Pew are two of the best in the business. I wish I had the link off the top of my head that gives the history of presidential polling. Harris (contrary to the myth of Zogby) nailed 2000 most accurately; Zogby was up there in 2000, and dead on in '96. Gallup hasn't nailed it since 1984.

    •  Gallup admits to not doing much state polling... (4.00)
      Matt,

      Frank Newman, who comes across as if he attended Bob Jones University, is the main Gallup representative.

      On WNYC, interviewed on the Brian Lehrer Show recently, admitted that the "Gallup" organization didn't do State polling.  He also admitted that they should.

      It's weird.  Such a supposedly vaunted polling firm (actually, they disappeared for awhile then reappeared in a new guise, years after Gallup's death as I recall)....

      Not doing State polls until the past two weeks?

      And it's true.   Since when does Gallup do State polls?

    •  Fox Poll is FAR Better Than Gallup (none)
      I'd trust a Fox poll now before I'd even look at a Gallup LV.

      Fox polls occassionally ask some ridiculous questions seemingly designed to get bullshit propoganda points for the rightwing.  But in terms of their horserace and favorability numbers, they seem to be fairly sound compared to most of the rest of those jokers.  They probably a skew a couple points Repub, but they are consistent in their skew, so when taken with a couple of the other polls that seem to be sound, such as Harris and CBS/NYT.

      Matt, I'm sure you're familiar with Pollkatz, but those who aren't should consider poking around his site.  He's got great comparisons that show how skewed and erratic Gallup is compared with everyone else.  ABC/WaPo is probably more Repub, but consistent.  Gallup is just a joke.  

    •  Here's the site (none)
      Hi Matt!

      With Kos' search engine, ah, "overloaded" I figured I'd come to the rescue, as I had posted this site before in one of my early diaries:

      www.ncpp.org/poll_perform.htm

      It is a fascinating site.  You're right, Zogby's reputation came from '96, as he was one of the only to say Clinton will win by single digits.  But yes, Harris got it almost exactly right in 2000.  

      Kevin

  •  Fair and balanced reporting (4.00)
    Imagine if the Wall St. Journal, Pew, and Harris polls didn't come out first -- the media would be calling Bush the winner already. Oh wait, they already have.

    On Paula Zahn today, they showed their current electoral map projecting that Bush would win 290-248. Then Howard Dean comes on and her first question is: "If the election were held today, George Bush would win. What do you think of that?" Dean smartly brushed it off and said who cares.

    But what kind of question is that? What happened to: "According to our polls, President Bush has a slight lead in the electoral college. Of course, many states are tied or too close to call." Zahn then went to ask Dean essentially the same question six times in a row: Why does John Kerry suck? Like Dean's gonna withdraw his support and call Kerry a crappy candidate.

    One more Zahn moment. They showed the current Pew poll, which is tied at 46%. How did she introduce it: "John Kerry may be experiencing a slight rebound." No mention that the previous Pew poll had Kerry down 12 points.

    Fox News doesn't even discuss the election anymore. It has now become America's official 24-hour "Memogate/CBS & Rather Must Pay" network. I don't know what's worse.

  •  Since this is recommended (4.00)
    I'll post here, too. I don't know internals of latest poll, but their previous one (Bush up 7 in LV, 1 in RV) combined these half-samples:

    Based on 265 Republicans or Republican leaners and 234 Democrats or Democratic leaners. 

    and

     

    Based on 239 Republicans or Republican leaners and 231 Democrats or Democratic leaners. 

    So that's 504 Repubs, 465 Dems. Doesn't match national stats, and I'll bet the current poll is much more lopsided. Can't wait to see - but by then the media will be in post orgasmic dreams...

  •  Nevermind Dan Rather's resignation (none)
    I think Gallup needs a huge amount of heat.  

    Either they are:

    1. Staggeringly incompetent (which shocks me about no Republican organization anymore), or
    2. They are trying to create an impression of invulnerability for Mr. Bush to depress Democratic voters and keep them from the polls.

    •  Yes (none)
      Additionally, they're trying to make the "big Diebold election 2004 fraud" more plausible. You can't just skew the DRE results without having some polls showing that an outcome favorable to your candidate is possible.
      •  A more sinister motive? (none)
        Actually, that's exactly what I've been thinking.  If ALL pre-election polls show Kerry winning, all eyes (and accusations) will fall on Diebold if Bush somehow pulls out a papertrail-less victory.  However, if Gallup continues to project a strong Bush victory -- even if they are multiple percentage points away from every other polling outfit -- then the GOP can simply point to Gallup and say that Gallup's "methodology" must have been right while everyone else's was wrong.
  •  I recommend flooding the media with letters (none)
    Certainly, we can complain to CNN, USA Today, and Gallup itself.  But more importantly, we can complain to their competetors, and embolden them to attack Gallup.  The almighty buck will motivate them if they see blood in the water, I think.

    It's 1 am and I have already put in my 8 hours for John Kerry, so I'm hoping someone else can post a list of where we should be sending mail.  People actually write if you provide them the information without them having to look it up.

    •  Good idea (none)
      Sending emails, faxes, or letters to their competitors telling them Gallup is trying to embarrass them or make their polls look like frauds ought to do the trick.  Politely of course.
    •  We are taking the first bold step here. (none)
      By calling Gallup on their willful deceitfulness!

      More and more people will judge the data for themselves, and when they see the facts, they will realize that Gallup is not neutral, but highly skewed.

      Scott Rasmussen is a much better pollster, if you want to pick a Republican pollster!

      Gallup has galloped away from the truth!

    •  Here's My Letter (none)
      Most national polls report that the presidential race is a statistical dead heat, or with Mr. Bush no more than five points ahead. Yet, the Gallup poll shows George Bush ahead by 14 points. You would think that over at Gallup, someone would wonder what the hell was going on--especially after the debacle of Gallup's flawed polls in the last Presidential election. You'd at least think they'd be a little cautious and try to redeem their reputations since the once well-respected Gallup pollsters haven't called an election accurately since 1984.

      I suspect the problem over at Gallup may be their likely voter model--but of more concern to me than Gallup's problems is the media that it's linked to. Because the Gallup poll is also the USA Today and CNN poll, it gets disproportionate news coverage even when, as it is in this case, an outlier. Even though 13 out of 17 national post-convention polls show the race a dead-heat, tomorrow's headline is going to be about the Gallup poll and this points to a real problem in media consolidation's effect on the Democratic process.

      Is there a better way to handle election polling and will your network be pursuing a story about it?

  •  All over the place (none)
    Kerry or Bush in the lead? Depends on the day, the poll

    Statistics experts say all of the polls could be right because the differences among them fall within the margins of sampling error. All were conducted on different days, which can affect a poll's outcome, and all used a different formula for identifying likely voters.

    This story regards the Minnesota state polls, but it applies to all at this point.

    •  But that ignores the SYSTEMATIC nature of bias. (none)
      Gallup is CONSISTENTLY over-estimating the Bush vote, and they did in 2000, too!

      There is something wrong with their sampling procedures....

      ...and it isn't an innocent mistake, because they've been consistently wrong and they have shown that they do not care.

      We should petition for Gallup to not be allowed to select the debate audiences!!!

      They will go for Bush, big time!!!  Important!

  •  On the other hand... (none)
    ...with this and the Rasmussen numbers, we need to brace ourselves for a disturbing possibility: that Bush may have seized a very real three or four point lead in the last couple of days. Nothing will counter that lead except for a massive wave of effective political advertising, something which this candidate is doing, but just not well enough.
  •  This is fraudulent (none)
    How do they justify running this headline with all the other new polls showing a very tight race? The previous three polls to come out (9/12, 9/13 and 9/14) had a TIE, Kerry +1 and Bush +1 -- BUT NOW IT'S BUSH +13 USA Today and Gallup look fucking ridiculous.

    I would think CNN and USA Today would start looking around and wondering what's going on at some point. How logical is it to tie all of your major poll coverage to ONE POLLSTER anyway? The headline should have read: Latest presidential polls mixed. And then they should have fired Gallup.

  •  Debate audience? (none)
    I just read on another thread that Gallup picks the audience for the debates -- anyone know if that's that true? And if so, a concern? Or tinfoil hat material?

    <<This space for rent>>

    by daunte on Thu Sep 16, 2004 at 10:47:48 PM PDT

  •  Maybe instead of (none)
    railing about the weekly results of each polling outfit, you could analyze all the pollsters at this point in the races in 2000 and 2002 compared with the actuals in November.

    My quick review of 2000 indicates that most of them except the LA Times reported Gore leading and by far too many points.  Newsweek, Pew, Harris and Zogby were way too high.  Gallup not off by as much.  The LA Times was way off in the other direction with GWB showing a large lead.  Investors Daily Business reported a dead heat with 11% undecided - IOW they didn't push the leaners.  

    I was reading MyDD and dKos in 2002 and most of the comments then were similar to what I see today -- "the polls are all wrong."  Yes, the polls were wrong if one expected DEMs to win that year, but not wrong at all when the votes were counted.

    GWB doesn't have a ten or fourteen point lead.  But he has a lead.  And if he is at 49 or 50%, it's looking bleak for our side.  

    •  Your analysis is incorrect.!!! (none)
      Are you just ribbing us?

      Don't expect us to take you seriously!

      If you picked a particular week, Gallup may have been accurate occasionally, like a broken clock is correct twice a day.

      But overall, Gallup overestimated Bush's vote by a wide margin, and you know it if you really checked the 2000 results all the way through October!

      Don't lie.  Don't pick a specific date, ok?

      •  this (none)
        person is wrong big time...

        http://colorpolitics.blogspot.com http://colorpolitics.blogspot.com

        by bigassbrando on Thu Sep 16, 2004 at 11:04:58 PM PDT

        [ Parent ]

      •  Do you always accuse (none)
        someone who presents an historical fact of lying?

        Since I even specifically stated that GWB wasn't up by 10 or more points, it should have been clear that I wasn't accepting Gallup's numbers or even suggesting that their polls are accurate.

        I do, however, think that the polls in the first week of September (when the conventions  and Labor day holiday are over and more people are aware of the fact they we are having an election) are a good baseline.  In 2000 all of them except the LA Times had Gore winning which he did do.  The later polls until a few days before the election were all over the place but mostly had Gore losing.

        Things can happen between now and November.  Or nothing can happen.  Right now GWB has a lead -- we just don't know by how much (and when I hear reports as I did tonight that DEMs on the ground in MN are concerned, a state that should easily be in the DEM column, I get very concerned).  GWB is the worst POTUS in my lifetime and makes Nixon look good by comparison, but the voters at this time have not rejected him -- and if they haven't rejected a man who has plunged this country into debt and a war based on false claims, exactly what is it going to take for them to do so?  (A stock market crash in October or seeing him falling down drunk would probably do the trick.)

        You are free to interpret the polls anyway you want, but don't accuse people like me of lying when we present data and prior experiences.  You can come back and gloat after the election if Kerry wins handily.  And if GWB gets a landslide I will have been equally wrong.  But I lived and breathed those polls in 2002 and right now it's feeling like deja vu.

      •  Check the 2000 poll (none)
        mumbers again - see "Had It's reference below.  And while Gallup did have some high margins for GWB in some weeks, in others they were lower than other pollsters.
    •  Gore was way too high? (none)
      At this point in 2000 he was pretty far ahead, and then, as we all know, came the debates....

      To first time voters: Voting is just like driving; if you want to go forward, you put the lever on 'D'. If you want to go backward, put it on 'R'.

      by AWhitneyBrown on Fri Sep 17, 2004 at 01:19:35 AM PDT

      [ Parent ]

      •  Many of the polls had it (none)
        tied or Gore with a one point advantage at this stage in the race.  A couple of the tracking polls had it right as did Investors Daily.  That suggests to me that the debates had little to no impact as the undecideds broke evenly.
      •  Anytime a (none)
        candidate takes a big jump when previously polling at a much lower number for a long period of time, that jump, in the absence of some powerful factor -- like the opponent being caught in a scandal -- is a recency effect that fades over time if not reinforced with whatever factor caused the rise in the first place.  The debates didn't so much hurt Gore as they didn't help him.  He was unable to reinforce what the public saw in him in his acceptance speech.  Or maybe it wasn't Gore at all but Lieberman's performance that gave people second thoughts.  None of the polling is anywhere good enough for us to determine the strength of all the variables that came into play; so we can only guess and hope that we are more right than wrong.
    •  You are Wrong, Proof is here: (none)
      I'm not being critical because you are presenting facts.  I'm being critical because you are NOT presenting facts.

      Go here for a pretty long list of 2000 polls.

      Look at the last column ("Bush minus Gore") which is the Bush lead.  From Oct. 8th onward, almost every poll had Bush with a very good lead.

      Where can anyone look at that data and say "Gore was in the lead"?

      Sorry but wrong.

      •  AH, I take it back (none)
        you did say "At this point in the race."

        September 2000 is true.

        Sorry

      •  As I said, (none)
        you were correct, but I lost the point you were making here:

        you could analyze all the pollsters at this point in the races in 2000 and 2002 compared with the actuals in November.

        My quick review of 2000 indicates that most of them except the LA Times reported Gore leading and by far too many points.


        bold mine

        I read too quickly and the majority of the post sounded like an overall statement of 2000.

        Apologies again.

        So this means that the debates are Key?

  •  Well Researched Chris (none)
    I've always thought political operatives have probably considered infiltrating the polls. There is a known bandwagon effect with these things. You would think strategists who desires to control as much as the campaign environment as possible would have considered trying to influence the polls. I am somewhat familiar with the industry and have thought it wouldn't take much to get analysts to screw up sample and weighting schemes, a tabber to f with the data or an interviewer to mess with the responses.  

    I did a diary entry in August about my Gallup suspicions. My Crazy
    Uncle Jake might have been right. I might even take credit for it now that it doesn't seem so nuts.

    •  There are no recognized standards (none)
      I've always thought political operatives have probably considered infiltrating the polls. There is a known bandwagon effect with these things.

      Given the way the news media like to report on campaigns, the desire for headlines, and the lack of transparency it is not at all unreasonable to suspect that the results are being manipulated for political purposes.

      But of course just the desire to sell headlines could explain it too.  The idea that a large percent of Americans are changing their minds about the race day to day according, presumably, to what they see on TV and read in the papers also feeds in to the egomania of those in that business.  

      Polling is a complex thing its practices do not even have the sort of "generally accepted principles"  that accounting does, and look at how much you can manipulate accounting.  And if you want to put on your tinfoil hats look at how well this sort of "black box polling" fits into the sort of "black box voting" that is being introduced.

      •  Electronic Voting Needs Many Safeguards (none)
         And if you want to put on your tinfoil hats look at how well this sort of "black box polling" fits into the sort of "black box voting" that is being introduced.

        Exactly. The thug element in the political world found ways to mess with mechanical voting machines when they were introduced. The element is still with us (i.e. Rove) in this information age.  

        Just because you are paranoid ...

    •  The thing to keep in mind (none)
      If you were Karl Rove and price and scruples were no barriers, would you make an attempt to subvert Gallup, the "most trusted name in polling" and an organization that has a big hand in the debates?

      The better question is, "how could you resist NOT doing this, if it were possible?"

      Visit the Diary of the Lying Socialist Weasels, for commentary from the Original Progressive Web Warriors!

      by Jonathan on Fri Sep 17, 2004 at 04:16:05 AM PDT

      [ Parent ]

  •  this should not be surprising (4.00)
    gallup's role here is to overestimate bush support levels and therefore disguise pro-bush election fraud on nov 2.

    fraud is more blatant when election day results overtly contradict the latest pre-election day poll numbers as was the case in GA's 2002 senate race, but is less obvious when it appears to confirm pre-election day polling - especially polling by the best known name in the business.

    i did not know before reading this diary that gallup had been bought by a bunch of republicans, but such a move fits the republican agenda for election fraud perfectly. there is therefore little point in politely asking gallup to stop skewing their numbers, just as there is little point in politely asking the republican establishment to not engage in election fraud at every level they can get away with.

    the best defense is to ceaselessly highlight gallup's republican connections in an effort to lower its legitimacy in the eyes of the public. yeah its a duanting task but the gallup name is now tainted and we should shout it from the rooftops until their republican connections are as widely known as diebold's are.

  •  Shift sure looks remarkably smooth (none)
    9/13-15/04: 55 42

    9/3-5/04: 52 45

    8/23-25/04: 50 47

    8/9-11/04: 50 47

    7/30-8/1/04: 51 47

    7/30-31/04: 50 47

    7/19-21/04: 47 49

    7/8-11/04: 46 50

    Given the margin of error is +/- 4%, one would expect some bouncing up and down a little bit just by random chance. But these show Bush's numbers very consistently rising and Kerry's numbers very consistently dropping. I'm no mathematician, and I don't claim this proves anything, just looked odd to me.

  •  very salient info on Gallup methodological flaws (4.00)
    Kossack Jon B Good wrote a diary that explains the technical reasons why Gallup is getting such skewed results.  As a survey professional, I believe his explanation makes great sense.

    Gallup is being more blind and stupid than merely partisan.  Their methodology is flawed, and they ought to be smart enough to realize that.  But maybe the Repub leanings of the top brass are fighting with their own good sense.

    Anyways, I highly recommend Jon's diary if you want to know why Gallup is getting these cracked results.

    Also, check out this diary from pontificator, detailing criticisms of Gallup's methodology by respected academics.

    Lastly, something from the experience of one of my former students, who was getting her Ph.D., who has taught quantitative reseach methods, and who did part time phone surveying work for Gallup about 8 years ago.  She said that she and her co-workers often laughed at how poorly the questions were worded.  Sometimes they had to violate their instructions and read the questions differently just so the respondents could understand them.

    I was shocked by this at first.  My guess is that Gallup cranks out so many polls, that their research managers don't have either the skills or time to carefully test and adjust their questions sometimes.  Overall, I put less confidence in Gallup than I used to.

    Immune system recovery! American body politic rallies to reject the NeoCon-tagion.

    by Civil Sibyl on Thu Sep 16, 2004 at 11:07:10 PM PDT

  •  What happened to 54-40? (none)
    Originally Gallup was saying it was 54-40? Now USA Today seems to have amended their website to 55-42? When I read it earlier, about 2 hours ago, it said Bush 54-40. Now it says Bush 55-42.

    Originally this poll was suppose to be out today. They then held it back until tomorrow. Did they hold back their polling due to the inconsistency with the other polls out there to do more analysis? Did that analysis give Kerry an extra two points somehow? What gives with Gallup? Why did they change their poll result a mere 2 hours after the USA Today story appeared on their website. How could a poll, supposedly conducted days ago to show a 3 point change like that in a few hours(2 to Kerry and 1 to Bush) when all the polling was concluded between 4 and 2 days ago? That doesn't seem very scientific or convey confidence in their results to me if they can update their website in one night after 2 hours and say....ummm, wait a minute, we got a change here.

    http://www.usatoday.com/news/politicselections/nation/polls/2004-09-17-gallup-poll.htm

    GDoyle

    "Deserves got nothing to do with it"-William Munny, "Unforgiven"

    by GDoyle on Thu Sep 16, 2004 at 11:11:44 PM PDT

    •  The Nader factor (none)
      According to pollingreport.com the 54-40 results are if Nader, Badnarik, Cobb and Peroutka are included. Without them the result is 55-42.

      A Conservative Government is an organized hypocrisy. --Benjamin Disraeli

      by rogun on Thu Sep 16, 2004 at 11:22:19 PM PDT

      [ Parent ]

  •  When I became suspicious (none)
    I became suspicious when the polls went haywire after the RNC. It was obvious that Bush was improving in the polls during this time, but that doesn't explain the suddenly large differences that existed between various polls taken around the same time.

    This just came across as very precarious to me. The differences seemed too varied for professional and objective polling. My conclusion was that some of these polling services are biased because I have enough respect for their science that I can't believe that these large variances were caused by mistakes.

    A Conservative Government is an organized hypocrisy. --Benjamin Disraeli

    by rogun on Thu Sep 16, 2004 at 11:18:24 PM PDT

  •  Here's the link (none)
    to USA Today's feedback form. Let's flood 'em!
  •  Talking Points about Gallup (none)
    Because this diary is recommended and will stay on the front page, I'm going to shamelessly link to my diary, which attempts to distill Chris Bowers' and others great points into bulleted talking points we can use on this.

    http://www.dailykos.com/story/2004/9/17/21148/5896

  •  my email to news editor at CNN (4.00)
    Dear Mr. Jordan:

    Gallup has just done another poll that is so at odds with the rest of the polling community ; just as it has done all this election season, that one should be cautious in using it.

    There are 4 polls  -Pew, IBD/CSM, Harris, and the Economist which show a tie.  They are polling the same time frame.  If I were a news director I would not hinge a lot of analysis and coverage on a polling company that is so consistently off in its polling.   And remember this is not the same Gallup poll run by the late Frank Gallup.  It is owned and operated by a whole other brand of fish.

    Lately there has arisen in the academic polling community serious concern about the methodology that has been pushed by the new Gallup polling firm.  2 links

      http://www.dailykos.com/story/2004/9/17/04339/5885
    http://www.thehill.com/mellman/091504.aspx

    Second
    Analysis of Gallup's polling all year and how it was so at odds with the polling of other firms. Please note in these articles  that mention is made of the extreme disproportion of Republicans  and Democrats that are polled.  It defies the historical record and commonsense and what we know of the feelings of ABB to think that Democrats will vote in a smaller percentage in this election than they have voted in the last 3 elections.
    http://www.mydd.com/story/2004/9/17/03529/1649#readmore

    I again urge extreme care in using these figures.  Such results so far from others should distress an editor who did not want to place his organization  in the position of being embarassed.

    Yours,

    Debra Cooper
    NY NY 10023

    PS please note that the CEO of Gallup is not only a Republican donor but a donor to right wing Republicans.
    Publicly available info from the FEC, do the search:http://herndon1.sdrdc.com/fecimg/advindsea.html
    CLIFTON, JAMES
    WASHINGTON, DC 20037
    GALLUP ORGANIZATION
       CAIN, HERMAN
        VIA CAIN FOR US SENATE
        06/25/2003    $2000.00

    For those of you who don't remember who Herman Cain is:
    http://www.cainforussenate.org/

  •  polls (none)
    Polling is less accurate now than any time since Truman-Dewey...

    no one is home, there are call screeners, people don't answer their phone etc.

    I don't think polls have much of an impact on people. If anything, this will motivate the base.  

    I have no worries that Kerry will win the popular vote. I worry about Florida. That remains the key, and the Rethugs have shown they will steal it by any means possible.

    •  ohio (none)
      it's the new florida

      "No more watered down television trash" -- Fear

      by markymarx on Fri Sep 17, 2004 at 12:22:13 AM PDT

      [ Parent ]

    •  I dunno (none)
      It's comforting to think that polls don't really mean all that much. But if that's so, then why is it every time there's a new one released there are 17 diaries on it here, and twice that if it's bad for Kerry?

      I can't argue with Chris' numbers: Gallup has a consistent track record of favoring the Republican, and it's the only poll I'm aware of that's showing a double-digit lead for either candidate. All the other polls have showed this race tightening back up to a statistical dead heat, which makes Gallup's numbers fishy even without their demonstrable history of bias.

      But it's hard to stay calm and collected in the home stretch when every time I turn on the radio or get on the 'net, I'm hearing more doom and gloom from the pundits and more bogus numbers being reported as Gospel truth. I don't think I could survive the next six weeks by just shutting off all access to the news--but if I don't, I may wind up with hypertension and/or an ulcer.

  •  Why is this a bad thing? (none)
    Let the Republicans think they will win this no matter what.  It will drive less of them to polls.  Give them a false sense of security.  Why even go to vote when Drudge says my candidate will win?

    Keep it comin, Gallup.

    •  It's a bad thing because ... (none)
      ... it affects public perception by making Kerry look like a White Elephant. Obviously, no one wants him. Look how badly he's doing in the polls.

      Now, look how well Bush is doing in the polls. Obviously, people want him. So, who are we going to get behind, a winner or a loser?

  •  A question to those polling experts out there (none)
    Are there studies on the psychological effect on certain poll figures? I am asking this because a Bush lead could work either way. It might frustrate the Kerry folk, or it might energise them. I am not sure which is more common.
  •  Gallup Mehodology (4.00)
    Survey Information:

    1.  The survey was conducted between 3:00am and 7:00am Eastern.

    2.  If the line was busy, or no one answered, that was tabulated as a vote for Bush.

    3.  When determining likely voters only those with incomes above $40,000 were taken seriously.

    Survey Questions

    1.  If the election were held today, which of the following candidates would you vote for:  

    George W. Bush
    John Kerry
    Ralph Nader
    Arnold Shwarzenegger (counted as Bush votes)

    2.  What best describes your view of George W. Bush's job as president:
    Excellent
    Really Good
    Pretty Good
    Not Bad
    O.K.

    3.  Who is likely to lead the country better, the PRESIDENT and commander in chief -- the esteemed George W. Bush, or some liberal guy from Massachusetts where they conduct gay marriages?

    The right to be heard does not automatically include the right to be taken seriously. - Hubert H. Humphrey

    by PBJ Diddy on Fri Sep 17, 2004 at 04:37:46 AM PDT

  •  Probably Been Said Before (none)
    But Jimmy Breslin has said exactly what I have been feeling for months.  I think that there are 10s of millions of Kerry supporters that will never be included in polling, myself being one of them.  

    At home I have VOIP (through Vonage) and outside I use cell phones.  Young people are far more likely to exclusively use cell phones and they are far more likely to vote for Kerry.  I think there could be a rout this year and the news media will be dumbfounded at being so blindsided.  

    I do not mean to take away attention to the worthy goal of educating the media on Gallup's lack of objectivity, but I thought this also was an important point.

    •  I posted this on a separate diary (none)
      But that diary has already fallen off recommended diaries, so I'll say it again...

      I think the cell phone issue does more than simply undercounting younger voters.  It also undercounts voters in lower economic brackets.  This is purely anecdotal, but my husband has taught in inner-city neighborhoods in Brooklyn over the past 5 years.  Most families he works with use cell phones exclusively.  This is because of credit issues connected with traditional phone lines and because the families can't afford both landlines and cells, so if forced to choose, they choose cells.  When you add in those families being excluded from phone polls as well, you can see further explanation for Kerry votes being undercounted.  

      My suspicion is that Gallup, as the oldest, most-established polling firm, would be the last firm to admit that their firm relies on a dying methodology.  I suspect that their conservative bias is a result of resistance to change more than political ideology.  

    •  I hope it is a rout but.... (none)
      you need to show old farts like myself that a group notorious for depressingly low turnout (so-called progressive young voters 18-34 would have turned the tide in 2000 had they SHOWED UP) is actually going to VOTE this time.  

      I hear the claim that young voters are energized and active and are working hard for progressive causes every election year and every election year the turnout percentage remains basically unchanged for this key group.  

      Even if young people's preferred person to person communication technology, VOIP and cell phones, were accounted for in polling a signficant number of the responses would not make the cut in most of the LV models anyway because of the past low turnouts.  

      If Kerry and Edwards make a daily point of begging the young voters to PLEASE COME OUT AND VOTE, that will be a BIG SIGNAL that the jig is up.  

      Also, there is nothing I would like to see more than the news media being dumbfounded and blindsided, but it must be recognized that the media really wants to AVOID that.  This desire transcends, I think, even demonstrable media "bias".  After all the 2000 election was, to THEM, a "blown story", by ALL of them, of historic proportions.  

      The media will cover their asses over and above all other considerations.  Since the media spin is now all going to Bush then they must smell blood on Kerry, but the media is essentially soul-less and will quickly turn the tables if the wind changes direction and they pick up a new scent.

      So get the cell phoners, the IM'ers, the Text Messagers, the VOIP-ers REGISTERED and to the polls!  I know we're going to need them.

  •  WSJ on likely voters (none)
    Mainstream press article by Al Hunt making the key general point.
    •  That's retrarded (none)
      He decries the way that questionable polling can skew coverage.....  and then he quotes Hamilton Jourdan making anti-Kerry analysis based on the same bad polling that Hunt just raised questions about (and doesn't note the irony).
      •  Irony (none)
        it seems, is left to the reader in modern journalism.

        Still, it is good to have the fragility of likely voter models out there in print by someone not identified as being a Dem operative.

  •  WSJ article on uselessness of likely voters (none)
    http://online.wsj.com/public/article/0,,SB109536730469420041,00.html?mod=todays_free_feature

    Key quote:


    Gallup explains it has what it considers a time-tested formula for determining most likely voters. It asks eight questions, such as current intensity of interest, past voting behavior and interest, and whether you know where your voting place is.

    "We've discovered that if we ask a set of more indirect questions, we can better predict who is or is not likely to vote," Frank Newport, editor in chief of the Gallup Poll, has said.

    But there is reason to suspect those criteria are outdated, especially in an election where both sides say the intensity level is much higher than four years ago and get-out-the-vote organizations are considerably better than ever -- few people on Nov. 2 will be in the dark on where the voting polls are.

    "A formula that made sense years ago may not recognize all the changes in society," notes Mr. Hart. "It gives more credence to past behavior and too little to current interest."

    "For low-turnout elections those old models work well," suggests Bill McInturff, a Republican, and the other WSJ/NBC News pollster. "But in today's presidential election those models tend to [tilt to] a little older, a little more white, a little more affluent and a little more Republican voters. They may miss some of the extraordinary activity going on in African-American and Latino communities."

  •  The Note (none)
    Even they couldn't resist being snarky about this poll:

    The best example of how degrading it can be to be a political reporter: USA Today 's classy Susan Page forced to write up the Gallup poll/joke "showing" the president with a mythical 13-point lead for the front page of her paper -- thus suggesting Gallup's 2000 track record of wild swings might be replicated this cycle.

    (Question it gives rise to: What's the real margin? A. 5-6 points, with Bush holding small- to medium-sized (surmountable) leads in most of the important battlegrounds.)

    "I don't want to wake up on Nov 3 with George Bush still president and ask myself if there was anything more I could have done." -- Moby

    by lapis on Fri Sep 17, 2004 at 07:24:26 AM PDT

  •  I'm Confused (none)
    Why hasn't this poll been trumpeted on CNN's Web site?  It's buried in there.  Wouldn't you think this would deserve a headline?  What's going on?

    There was never a good war or a bad peace. - Ben Franklin

    by AndyPA on Fri Sep 17, 2004 at 07:39:31 AM PDT

    •  You're right... (none)
      Odd that.  It IS on Gallup's site, so it has been released, I guess.

      How unusual.  

    •  See lapis' post (none)
      immediately before yours.
      •  Now It's There, But Carefully (none)
        Yeah, now CNN's site has the poll, but they're hardly presenting it with big flashing lights.  They're really being cautious with this one.

        There was never a good war or a bad peace. - Ben Franklin

        by AndyPA on Fri Sep 17, 2004 at 09:20:56 AM PDT

        [ Parent ]

        •  Good, maybe CNN realizes Gallup is unreliable (none)
          Maybe our blogging is waking people up!
        •  CNN knows there is a problem Gallop (none)
          Yeah, now CNN's site has the poll, but they're hardly presenting it with big flashing lights.  They're really being cautious with this one.

          On their political coverage tonight they seemed the same way.  They did not give the sort of play that they gave much smaller and slower shifts in Bush's favor.


          BTW as I under stand the "time tested" Gallop likely voter screen it works like this they ask a number of questions designed to measure who will vote, but unlike those who only ask how interested one is they also ask how interested you were in the last election, whether you voted in it and if know exactly where your polling station will be on November 2nd.  Anyone who did not vote last time, or was not very interested in the last election, or does not know the location of his polling place gets a low score.


          Then everyone who passed all those tests and who says they are "highly interested" in this election gets the max score and makes the cut for as a likely voter.  Then the results are adjusted to reflect the expected turn out.  In 2000 the target percent was 50% and so it max scorers did not come to 50% then they would start to count people who passed the other tests but only said they were likely to vote.  This year the only adjustment for higher voter interest was to use the number 55% and so count more second level most likely people.


          The obvious problem with this is that there is a bias against younger people, and mobile people.  All this is on top of the bias introduced by only counting people with land lines who answer the phone.

  •  My letter to USA Today (none)
    Dear editors:

    I frequently enjoy reading USA Today, and generally find its coverage to be admirably straightforward.  However, I'm afraid that your newspaper has embarrassed itself in terms of its coverage of the recent Gallup poll seeming to show a 14-point gap between George W. Bush and John Kerry.

    Coverage of polls--particularly in such an unusual electoral environment as this one--require context and scrutiny.  The real story here isn't the ostensible 14-point lead shown by this poll, but the wild divergence between this poll and other reputable polls (Harris, Zogby, etc.) taken at the same time, which show the race to be a statistical tie.  It's also worth noting that of all the major polling outfits, the Gallup poll has consistently skewed Republican, showing a Republican edge in nearly every state and national poll, despite numerous other polls' demonstrations of a virtual tie or marginal Democratic lead.

    This recent Gallup poll represented an opportunity to educate your readers about the nature of polling and the subtleties of polling methodologies.  Instead, we have received context-free and uncritical reporting of the results.  I believe that most of your readers sense intuitively that this presidential race is extremely close; with such general and uncritical coverage of the result, the only casualty is your own credibility as a news source.

    Please give more care to the way you report polling results.  It will be difficult for me to read your paper with the same degree of trust that I once did.

    Sincerely,

    loopster

    (okay, I didn't really sign it "loopster."

  •  Gallup is choosing TV debate audience!!! (none)
    We shouldn't be standing still and allowing a partisan polling firm to pick the debate audience!

    If Gallup picks the auddience, it's likely that the 60% Bush audience will cheer mostly for Bush, which will skew the debates!

  •  More and more people realize Gallup is biased. (none)
    It appears that it is not just we liberals and progressives who are putting the pieces together and realizing that there is a clear pattern of misleading polling by the "Gallup" organization.
  •  Keep in mind Gallup is working for TV!! (none)
    There is one thing everyone needs to keep in mind, which is that Gallup is now working for a TV network, which wants ratings!  And what is better for ratings than widely fluctuating polls!

    And there is a very easy way to produce that, LOW SAMPLES.  I'm not just talking about overweighting GOP voters, I'm talking about samples of only about 750 for their LV.  That gives an error ratio of 4-4.5%.  In contrast, check the history of the Economist polls.  The Economist has never shown anyone with a double digit lead, because their sample is MORE THAN TWICE Gallup's, at about 1800.  Their error ratio is only 2%.

    This is not the first time Gallup has done this.  Even in 1996, I was startled to see Clinton's lead over Dole fluctuate by 14% IN ONE WEEK!  And in 2000, some have already indicated that Gallup moved from Gore up 11% to Bush up 13% in just a couple of weeks.  In contrast, Zogby, which always had over 1000 in their sample, never had either with a double digit lead.

    So, keep in mind, Gallup is doing this partly for ratings, as smaller samples lead to wider error ratios, and thus, more "exciting" changes in trends.

    Kevin

  •  REALLY FUCK GALLUP UP (none)
    BY FUNDING OUR GET OUT THE VOTE EFFORT AT

    http://actforvictory.org/

    ACT will bring in new and uncounted voters if we take one second out from e-mailing CNN and Gallup and DONATE TO GET DEMOCRATIC VOTERS REGISTERED AND VOTING.

  •  one critical point being glossed over (none)
    regarding the institutional bias in gallup's polling sample methodology: not only do these results skew in favor of bush in the head-to-head matchup and the three- or six-way matchups with kerry, but they also show an artificially elevated approval rating for bush, an artificially high right track number, artificially high support among women, as well as artificially low support for kerry in all those same ways.  in fact, it says a lot that, despite the oversampling of republicans in the gallup poll, john kerry gets an approval rating of 51%, and his disapproval rating is only at 44%--exactly the same as bush's disapproval rating in the same skewed sample.
  •  Gallup in 2000 (none)
    It seems like Gallup's 2000 polling was fairly accurate. Were they not oversampling Rethugs in 2000, as they are now? If they were, we really might be in trouble this year.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site