Skip to main content

Steve Soto looks at Gallup's latest effort, and finds it worse than the last.
Gallup has done it again. After supplying CNN and USA Today with a poll two weeks ago that showed a double-digit Bush lead amongst likely voters that turned out to have a significant bias in its sample favoring the GOP, Gallup did it again yesterday.

Except that yesterday, they not only did it again, they apparently felt that a 7% GOP bias wasn't good enough. So they perpetrated the same fraud upon the media (including their partners CNN and USAT) and voters and this time used a 12% GOP bias in their likely voter screen. I kid you not.

Just got off the phone with a reporter from USA Today who is writing a story on potential problems with the Gallup poll, and the liberal blogosphere's work in bringing attention to the issue.

Originally posted to Daily Kos on Tue Sep 28, 2004 at 12:15 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

  •  so you think the story will be (none)
    on page 14c?

    After the article on Brittany's latest divorce, right?

    "That government of, by, and for the people shall not perish from the Earth."

    by TheGryphon on Tue Sep 28, 2004 at 12:12:26 PM PDT

  •  use the new poll from IBB cristian science (none)
    as evidence.

    this is your mission: TERMINATE the Bush presidency

    by nevadadem on Tue Sep 28, 2004 at 12:13:22 PM PDT

    •  asdf (none)
      I don't follow the IBB/CSM poll.

      Quoting PoliticalWire:

      The new Investor's Business Daily/Christian Science Monitor poll has Kerry leading Bush 46% to 45% among likely voters. Among registered voters, they are tied at 44%.

      Is this a significant deviation from the standard LV model? I thought that LV models traditionally eliminate more D votes than R votes.

      •  You must remember this.... (none)
        Is this a significant deviation from the standard LV model? I thought that LV models traditionally eliminate more D votes than R votes.

        Keep in mind that folks who voted in the last election are "likely voters."  Also keep in mind that Gore won the last election......

        My preznit went to Iraq and all I got was this crummy economy!

        by roxtar on Tue Sep 28, 2004 at 04:06:14 PM PDT

        [ Parent ]

  •  USA Today (none)
    whoa, USA Today is writing a story on how the company they commissioned to do their polling is so fucked up?

    who is Gallup working for, anyway? what is their motive to skew the polls as they have been, if their bosses themselves are unhappy with it? (or am i overestimating the importance of USA Today in the whole process?)

    regardless, it's great to see this issue coming to light in the national press.

    "George Bush has no strategy for Iraq.  I do." - John Kerry, 9/20/04

    by thirdparty on Tue Sep 28, 2004 at 12:13:30 PM PDT

    •  The poll taken two weeks ago (none)
      was not sponsored by USA Today or CNN. Seems Gallup paid for it themselves.
      •  Sigh (none)
        That would explain why CNN last night was hurling Gallup like the holy grail - a lot more than usual.

        Just my luck to tune in the moment Candy Crowley is mockingly saying "The Kerry campaign says Gallup is an outlier poll," while she smirks tongue in cheek not taking the Dem strategists seriously.  I couldn't change the channel fast enough.

        First Bill Schneider, now Gallup - my boycott of CNN continues...

        •  I have simply stopped watching (none)
          them--not consciously--but just the way I don't watch Fox.  Who cares.  It's a pack of shallow thinkers spouting any ol' manure that comes into their head.  Dailykos and the web are how I stay informed now.

          The ignorance of one voter in a democracy impairs the security of all--John F. Kennedy

          by DCDemocrat on Tue Sep 28, 2004 at 12:59:43 PM PDT

          [ Parent ]

          •  CNN (none)
            Yes, it's very sad actually.  I used to appreciate CNN's straight-shooting approach (though Crowley was irritating in exactly the same way during the 2000 election).  At the gym I go to, above all the stairmasters, is a row of TV's.  One of them was always tuned to CNN, for years.  A few months ago somebody would periodically turn it to Fox.  Then it would be switched back to CNN (always when nobody looking).  In recent weeks it's been on Fox all the time.  I'm saddened to realize I don't care.  What's the difference?  (Except Fox's polls are more trustworthy!)
          •  CNN (none)
            Yes, it's very sad actually.  I used to appreciate CNN's straight-shooting approach (though Crowley was irritating in exactly the same way during the 2000 election).  At the gym I go to, above all the stairmasters, is a row of TV's.  One of them was always tuned to CNN, for years.  A few months ago somebody would periodically turn it to Fox.  Then it would be switched back to CNN (always when nobody was looking).  In recent weeks it's been on Fox all the time.  I'm saddened to realize I don't care.  What's the difference?  (Except Fox's polls are more trustworthy!)
            One more comment: I've travelled some overseas this year and CNN International remains excellent.  It's amazing what they can do, when they're not hiding in fear of the Repub screamers.  (Everybody knows the Repubs don't travel.)
      •  Whoaaaa! (none)
        Gallup paid for it themselves? Someone, quick, follow that money! Somethin' smells rotten in Gallup's Poll!

        I am angry that so many sons of the powerful and well-placed ... managed to wrangle slots in Reserve and National Guard units - Colin Powell

        by Pescadero Bill on Tue Sep 28, 2004 at 02:23:04 PM PDT

        [ Parent ]

    •  just for the hell of it (none)
      is there an innocent explanation for what they are doing?

      *****************

      also, does Gallup ask party affiliation before or after they ask presidential choice?

      •  Yes, actually (none)
        is there an innocent explanation for what they are doing?

        Corporate inertia. They have a product (their Likely Voter model) which has always worked for them and made money for them. Therefore, by the Logic of Senior Management of Large Companies, you never have to change your product.

        The world has moved on and their product doesn't reflect it. Oldest corporate story in the world.

        End result: your newer, more agile competitors (Zogby for instance) get your market share. Watch for a shakeup of the poll industry post-elections.

        Except that there's also a lot of corporate inertia in the large media who SUBSCRIBE to polls.

        •  Yes, But (none)
          The larger Bush lead this time is among registered voters.  Ruy has some pretty interesting (and funny) things to say about this.

          This aggression will not stand, man

          by kaleidescope on Tue Sep 28, 2004 at 01:03:29 PM PDT

          [ Parent ]

        •  Need to turn the word Gallup into a farce (none)
          I say we start using the word Gallup to refer to all things implausible (like Jodi Wilgoren, like Judy Miller, like Wolf of late). Use it as a word to embarrass not only the polling company, but the people associated with it.

          Let's re-brand Gallup as all that can go wrong with polling.

        •  George Gallup Jr.'s (none)
          quote in the moveon.org ad mentioned upthread is telling:

          "Gallup, who is a devout evangelical Christian, has been quoted as calling his polling 'a kind of ministry.' And a few months ago, he said 'the most profound purpose of polls is to see how people are responding to God.'"

      •  Gallup has a good reputation & long history (none)
        and so news organizations trust it. Unfortunately, it looks like they've been riding on that. Their techniques are old and almost certainly out of date.
  •  asdf (none)
    it is obscene what they are doing.

    how does Gallup explain these samples?  

    •  The word "bias" is misleading (none)
      It is my understanding that Gallup doesn't consciously decide what the partisan makeup of their samples should be. In other words, they don't rig the polls. Rather, it is an error of omission. They make their calls -- using a methodology, apparently, that tends to emphasize the response error in all telephone-based polls -- and then just present them, without correcting for party ID. Right now, there is a severe bias in favor of Republican ID'ed voters among the poll of people available and willing to respond to traditional polls. I have no idea why. But most polling outfits are seeing the same thing. It's just that some (more than half) correct for the error by attaching weights. Gallup doesn't. Either does CBS or ABC, or AP.
      •  OK, how's "incompetence"? (none)
        If Dem.'s turnout on average 4% more voters than the GOP, and if we're using these polls to predict an election, rather than just looking at the purty numbers, then anyone with half a brain would adjust their proportions accordingly.  

        Right?  

        As it is, it's a 16 point skew, which is producing an 8 point lead.  

      •  yeah this is what i thought... (none)
        i agree.  but i've heard that they do weight
        their results for gender and age ... just not
        political affiliation. (i don't know that to be
        true)

        so it is a situation where most every other
        polling firm...led by Zogby, is weighting based
        on political affiliation and Gallup still
        refuses to do so.

      •  Maybe that is why Gallup is so fucked up (none)
        It is my understanding that Gallup doesn't consciously decide what the partisan makeup of their samples should be

        If that is true, then Gallup has no clue about how to conduct polls. If Gallup doesn't adjust the sample to fit the facts on the ground, then Gallup is just asking questions that bear no semblance to reality

        Consider that a sample of 1000 people from Montana would poll different than 1000 people from Massachusetts. Gallup, and other polling groups, ARE SUPPOSED TO ELIMINATE THIS BIAS in their results

        Gallup apparently acts as if the Bias doesn't exist

        Somebody from Gallup should take a basic statistics course. If the sample does not resemble the whole group, then the sample is meaningless

  •  We should riot over this (none)
    The entire Patriotic Opposition should create a large and long noise against these Propaganda Polls.

    Why is it that the right wing can put "forgery" into every news story within 12 hours, while things like this are basically unchallenged?

  •  So Kerry is up by 4 points? (none)
    What is the effect of weighting the poll accurately? Does the Gallup poll, by removing the additional GOP votes, reflect a 4 point Kerry lead?

    Just asking.

    If the poll shows Bush up 52% to 44% and 12% of respondants were GOP, how does this affect the poll? Does Bush lose 12% or is is more evenly distributed by some mathematical means?

    •  check some of the diaries on the topic (none)
      Everyone had slightly different takes, but assuming that the other statistics are right, you can crib from pollingreport to find the % of Democrats voting KE, vs. the % of Repubs voting BC, then making up some split for the undecideds/otheres, and it comes out to be about a dead wash.  'Cept Gallup surely didn't include all those Ohio voters who used the wrong paper....

      The baby AND the bathwater-- all of it!

      by persimmony on Tue Sep 28, 2004 at 12:21:59 PM PDT

      [ Parent ]

    •  Hard to say. (none)
      Statistics show general bias in the sample, but without having a reliable breakdown- especially amongst unaffiliated voters- it's hard to reweight accurately.

      Traditional turnout numbers show roughly 38-39% are Democrats, 32-33% Republicans with the balance considered independents.

      If the latter split evenly- which is a big assumption- you're looking at an adjustment to something like 50.5% Kerry to 49.5% for Bush, which works out to just about completely call it a wash- which, in turn would bring it in line with so many of the other polls.

      Especially when visions harden into dogmatic ideologies, they become inhuman, cruel and dangerous. - Arthur Schlesinger, Jr.

      by wingedelf on Tue Sep 28, 2004 at 03:40:19 PM PDT

      [ Parent ]

      •  Just a little methodology input (none)
        Here's how I reweighted the samples, with the assumption above:
        If the 25% who aren't Republicans or Democrats split their vote evenly, then roughly 12.5% of their vote allocates to each candidate. This is a little of a reach, since the initial 55-42 numbers still have a 3% undecided factor.

        Subtracting 12.5 from the Bush 55% returns 42.5. Adjusting this score downward the 7-8 percent the poll overweights the Republican representation, returns 35.5. Dividing this by the sample percentage returns that 88.75% of the sampled Republicans plans to vote for Bush. This is in line with most poll results broken down along party lines, possibly a little on the low side.

        Adding the half back in brings the number to a flat 48%.

        Adjusting the Kerry numbers the same way, 42% less 12.5% is 29.5%. The Democratic sample is underweighted 7-8 percent, returning 36.5%. Adding the 12.5% back in makes it a flat 49%. Splitting the remainder of three percent just to round it off to an even hundred, 1.5% is allocated to both, resulting in 50.5-49.5%

        As noted above, this depends pretty heavily on the results each candidate gets from his rank and file (to try and not over-weight this, I've used the lower figures for both, but that's less than completely reliable), and to unaffiliated voters split evenly. If the unaligned voters are skewed one way or the other, they can easily tip the balance.

        For what it's worth, you can use this type of approach to normalize any poll results if you have numbers reflecting the partisan breakdown of those surveyed. Numbers exist to show how the normative percentages which turn out on election day are divided, so it's just a matter of a little arithmetic from there.

        Note, also, that this doesn't take into account unusual circumstances, so it will not predict outstanding Get Out The Vote efforts or voter repression, so you're on your own there. ;-)

        Especially when visions harden into dogmatic ideologies, they become inhuman, cruel and dangerous. - Arthur Schlesinger, Jr.

        by wingedelf on Tue Sep 28, 2004 at 03:59:08 PM PDT

        [ Parent ]

  •  New Poll: Kerry up by +1 (none)
    The new Investor's Business Daily/Christian Science Monitor poll has the following:

    LIKELY VOTERS
    KERRY 46(43)
    BUSH 45(46)

    KERRY 45(42)
    BUSH 45(45
    NADER 2(3)

    REGISTERED VOTERS
    KERRY 44(43)
    BUSH 44(44)

    (stolen from Rimjob)

    http://www.dailykos.com/story/2004/9/28/132631/439

  •  The odd party affiliation numbers (none)
    by themselves aren't evidence that Gallup's polls are screwy - it is, in fact, feasible in and of itself that the GOP has gained a bit of party ID advantage. The more curious point is that no one else appears to be showing it. We know that some other polls aren't weighting for it, so they just don't appear to be getting the same numbers at all. Or something else could be at play - good polling outfits have to employ some massaging to account for demographic issues, and maybe Gallup is doing it all wrong. Unfortunately, info on that appears to be proprietary, so one can tell what they're doing.

    Whatever it is, yes, it's certainly annoying, and Gallup needs to start reconsidering their model if they want to maintain their reputation.

    •  No Change Possible (none)
      Soultaco, I respectfully differ.  Gallup's numbers add up to 99% of the electorate.  That means that 1% of the population either answered third party, refused to answer, or answered don't know.  Hardly likely, seeing as 5% usually answer "don't know" to every poll.  Want to know more? The Gallup Poll Is Rigged.  For Real.
    •  more than that (4.00)
      Coming up with 43% GOP to 31% Dem, when the numbers in 2000 were 39% Dem 35% GOP, isn't some small, explainable change.  It's a realignment along the order of the New Deal.

      And, as Ruy Teixeira has painstakingly pointed out, this hasn't occurred in response to a significant event.  Had the shift taken place right after September 11th, we'd at least be able to accept it MIGHT be true.  But, back then, there was only a small shift to the GOP, and it faded fairly quickly as the country retreated to status quo ante.

      This change, on the other hand, has been enormous, and has taken place in a matter of WEEKS.  Prior to the GOP convention, polls were showing nothing like this turnaround.  Is it really conceivable that a few speeches by Guiliani and Schwarzenegger could have had that kind of zeitgeist-altering affect?

    •  A bit of party ID advantage? (none)
      That's quite an understatement. For the GOP to gain that much party ID advantage would be on the order of the New Deal realignment of national politics. All because of a speech by Ahnold and a few ads? I don't think so.

      No, there's more at work here than simple random variance. It's too consistent in what Gallup does, and the party numbers fluctuate much more than the final horse-race numbers. That shouldn't be true.

      So, the party ID changes, but always comes out with a similar result . . . hmmmm.

      I smell somethin' . . . somebody cookin' around here?

    •   it is... feasible ... GOP has gained ... advantag (none)
      No, it isn't

      Facts on the ground show that democratic registration has always been higher than Repuglican registration

      and "new" registrations are going to the Democtrats too (BYW, new registrations do not count as "likely" voters by pollsters)

      All of the facts show that the repuglicans comprise about 35% of the voting public (at best) and the Democrats comprise about 39% of the voting public

      This is compiled from hard data, not opinion polls

      •  It is possible ... (none)
        that there would be a party ID change, particularly in the short term. Party affiliation can change. Right after a party's convention, it is in fact quite possible that enthusiasm will be high and more people (esp. independents and undecideds) will identify with that party.

        Let me note that I don't actually believe the party numbers that Gallup presents represent the actual voter turnout at the polls in November. I think their data is way off. But I do think that there could have been a bump in stated GOP party affiliation right after the convention.

      •  It is possible ... (none)
        that there would be a party ID change, particularly in the short term. Party affiliation can change. Right after a party's convention, it is in fact quite possible that enthusiasm will be high and more people (esp. independents and undecideds) will identify with that party. It's also true that events such as 9/11 could change party affiliation, and it apparently did, but not in the long term.

        Let me note that I don't actually believe the party numbers that Gallup presents represent the actual voter turnout at the polls in November. I think their data is way off. But I do think that there could have been a bump in stated GOP party affiliation right after the convention.

  •  we need to make sure the NYT picks up on this (none)
    And we need to organize a large media drive in the way of LTTEs to the editors of the small town newspapers in swing states.

    That's what I'm thinking we should do, and we need to discredit Gallup somehow. If the right-wing can discredit CBS, why can't we discredit Gallup?

    What's madness but nobility of the soul at odds with circumstance?

    by slinkerwink on Tue Sep 28, 2004 at 12:19:14 PM PDT

    •  Hard copy of NYT was all over it today (none)
      But only because MoveOn paid for a full page ad.  It's beyond depressing that the only way to get critical investigative journalism into Pravda on the Hudson is to do it yourself and pay for it.
  •  gallup also showed (none)
    Bush ahead the day after the DEM CONVENTION ----BY 5 POINTS if you look at all the other polls at the time they are about 10 points off in favor of Bush, I seriously wonder if Rove has a plant with Gallup or if thier in cahoots.

    this is your mission: TERMINATE the Bush presidency

    by nevadadem on Tue Sep 28, 2004 at 12:21:12 PM PDT

  •  Why? (none)
    Why is Gallup doing this? I must say, I haven't seen a convincing reason yet. And did this just start with the last few polls?

    "We must uphold a familiar commerce together in all meekness, gentleness, patience and liberality." I have a blog!

    by Marshall on Tue Sep 28, 2004 at 12:22:21 PM PDT

    •  Gallup has been doing it for a while (none)
      I became keenly interested in Gallup prior to the 2000 election. They were way, way off in their predictions.

      They had Bush up to win by 9 points only a few days before Gore won the popular vote by 1 pt. AND Gallup had stated a MoE of +- 7 ( that's right, SEVEN ).

      Pay no attention to Gallup. Most polls show it dead even od a slight Bush lead, but you can toss Gallup. This may be the nail in the coffin for Gallup's business, with Zogby and Rasmussen out there. This election might shake a lot of pollsters from the trees, and Gallup will be one of them.

      •  Right. (none)
        I've reached the conclusion that Gallup is pretty much irrelevant until they adjust their methodology.

        However, the general public may not be so skeptical, and I'd hazard a guess that many people get their polling information from one source and one source only.

        While I know that Gallup is most likely off, Joe Swinger in Ohio may not, and we've all seen the studies showing that people often vote for who they think will win rather than who they identify and agree with most.

        •  General public believes in Gallop (none)
          For me everything is filtered through the Rovian lense.  He believes that many Americans do not hear the background noise of truth.  They tune in to their favorite local news channel for 15 mintues, and if Gallop says Bush is way ahead, then for that voter it is fact.  No need to look further.

          We can bleat about the truth all we like.  Rove does not care.  We are voting for kerry anyway.  The republicans choose to believe Gallop, and the others are too busy to find out the truth.  They have been brainwashed and rove intends to keep them that way.

          This has to become a full PUBLIC story before it affects the regular voter.

      •  aoeu (none)
        +-7??  You might as well be throwing darts!

        My turtles laughter
        was loud when the Yankees lost
        22 to zilch

        by TealVeal on Tue Sep 28, 2004 at 12:46:30 PM PDT

        [ Parent ]

      •  Rasmussen? Maybe not (none)
        Rasmussen is right sometimes, but I'm not ready to put them on the same level as Zogby. Only Harris is another poll I respect right now.
        •  Rasmussen learns (none)
          He was wrong in 2000-because he didn't weigh for party ID.  He learned from his mistake, and now does so-and his numbers are very close to Zogby's now.  I trust them both-their names are on the door, thier reputation is on the line.
    •  Well... (4.00)
      Part of me seriously thinks that they're doing this so that when the Diebold machines report a Bush win, it won't look too outlandish. I mean, if every single poll was reporting Kerry ahead and then Bush "won" on election day, there would be a tremendous amount of suspicion. But if the polls are all over the place...

      I know this sounds like a wacked-out conspiracy theory, but the last four years have convinced me that there is nothing that's beyond these thugs. Nothing.

      •  you beat me (none)
        to the punch.

        "John F. Kerry: More than just a big cock and a nice ass."

        by drjimcooper on Tue Sep 28, 2004 at 12:30:12 PM PDT

        [ Parent ]

      •  Bingo! (none)
        It's not whacked-out conspiracy theory when manipulation of electronic voting has already been achieved.  Any cursory study of this subject and Bev Harris's seminal work shows that Republicans manufacture the machines, vote them into use, discourage paper trails, and then manipulate the results.  Simple as that.  We need to lift the blinders off and realize that this isn't just a "possibility", it is a concerted effort on behalf of the GOP to retain power and institute a one-party theocracy.

        "Conservatism makes no poetry, breathes no prayer, has no invention; it is all memory." - Ralph Waldo Emerson

        by reef the dog on Tue Sep 28, 2004 at 12:40:27 PM PDT

        [ Parent ]

        •  I would add (none)
          the observation that exit polling was "voluntarily" not used during the 2002 midterms, as a "response" to the media's "error" in predicting that Al Gore would win the popular vote in 2000. This enabled a history-defying GOP sweep to be more easily swallowed.

          For me, the discontinuing of the exit polls was the first red flag. As long as they were there to give conflicting results from the cooked voting machines, there would always be questions.

          And since (presumably) you can't fake people's responses to the simple question, "So who'd you vote for?" the exit polls just had to go.

          •  Will exit polls be conducted this time? (none)
            I hope so.

            You know, I too am starting to believe that Gallup (and others) are rigging the polls to allow the rigged voting machines showing a Bush victory to be believable.

            Which means we are no longer living in a democracy, but a dictatorship.

            :(

      •  This Article says the same thing (none)
        Gallup polls: Conditioning the public for vote rigging?

        Here is an excerpt:

        "It is interesting to note that James Clifton, who bought the Gallup organization, is a big Republican donor. He gave thousands to Right Wing Republican Georgia Senate candidate Herman Cain. (See http://www.opensecrets.org) Cain ran as a huge backer of cutting taxes for the wealthiest Americans. This is essentially the same tax position supported with vigor by the Bush-Cheney ticket."

    •  Because (none)
      if the polls show a Bush lead, it is easier to steal the election.

      "John F. Kerry: More than just a big cock and a nice ass."

      by drjimcooper on Tue Sep 28, 2004 at 12:29:27 PM PDT

      [ Parent ]

    •  Why (none)
      Why aside from the fact that they may actually be employed by BushCo?

      Why, for Rove, would be easy. If you can present the impression (bolstered by the vague notion that most Americans would have that Gallup is the most esteemed pollster in America) that Bush is ahead by 10 points, then you can:

      • Diebold some votes such that Bush ends up 10 points ahead
      • Call outrage when the actual vote comes back showing Kerry one point ahead, and call for recounts and challenges that will eventually land in Antonin Scalia's lap again
      •  Come on (none)
        Obviously it's good for Bush et al to have skewed polls in their favor. But what would Gallup be doing that? For the record, I am totally convinced by the evidence that Gallup is doing it. The question is why. And don't tell me because it's good for Bush--if you have evidence of Gallup being in league with Bush, produce it.

        "We must uphold a familiar commerce together in all meekness, gentleness, patience and liberality." I have a blog!

        by Marshall on Tue Sep 28, 2004 at 12:47:17 PM PDT

        [ Parent ]

  •  This kind of bias (none)
    can make an undecided voter look at that number and decide the Bush lead is so big, why get off my butt and vote!

    This has got to stop!

  •  Don't give up and don't give in (none)
    I am so glad to hear some explanation for these polls because they absolutely make me crazy. And then to hear them at CNN quoting them like they are words from God . . . .
    The purpose has to be that they want a low voter turn out . . . trying to discourage us from voting . . . AIN'T GONNA HAPPEN!
  •  Ridiculous (none)
    And it's sorta self-feeding.  Once the first slanted poll shows a 13 point Bush lead, they weight to THAT POLL and continue to crank the numbers up falsely.  Eventually they should be able to reach 100% for Bush with 100% Republican weighting.

    "Conservatism makes no poetry, breathes no prayer, has no invention; it is all memory." - Ralph Waldo Emerson

    by reef the dog on Tue Sep 28, 2004 at 12:24:17 PM PDT

  •  Here's hoping (none)
    that the article WILL appear and will NOT be on page 14q or 32n.  

    I already sent out another round of letters.  I'm thinking the info@ap.org is a good one-stop shopping letter destination.  If they put out an article, it seems to end up everywhere.

    But also, FAIR has a good list of major media contacts.

    And also the Kerry site has a good Letter to the Editor tool.

    I am strongly urging your organization to cover this story.

    The Gallup is now clearly trying to influence the upcoming Presidential
    elections.
    They are doing this by living in a fantasy world where they decide who votes
    and who doesn't.   Two weeks ago, they decided that vastly more Republicans
    would vote than Democrats.  Since Republicans say they are voting for bush
    9-to-1, Bush got a 13 point lead.  Sweet dreams for Gallup, but in contrast
    to nearly every other poll in existence at the time.
    Now Gallup continues to live in the land of make-believe with their new
    poll.  Here they dream that even more Republicans and fewer Democrats will
    vote than in their last fairy-tale.  Even then, they could only grant Bush a
    much smaller lead of 8 points. They don't seem to want to face what seems to
    be their nightmare:  In the last two Presidential elections, vastly more
    Democrats actually voted than Republcians.  Applying the true voting
    patterns to their data gives a Kerry lead.

    But since they have the name Gallup, they are trying to influence voters:
    "Don't bother voting, Democrats, it's not worth trying."

    Why does any respectable news organization take them seriously anymore?

    •  WOW!!! (none)
      I had no idea the head of Gallop was an evangelical Christian.

      Folks, we need to create a coordinated email campaign right here and right now! Can someone post a few links to feedback forms for us to use? We should all be bombarding CNN, USA Today and MSNBC with this stuff NOW!

  •  Somebody (none)
    check the temperature in hell
  •  Moveon.org ad (none)
    Moveon has a full page ad today in NYT.

    Headline - Gallup-ing to the right.

    Why Does America's Top Pollster Keep Getting It Wrong?

    "... they are refusing to fix a longstanding problem with their likely voter methodology."

    "...Seven other polls of likely voters were released that same week. ... No one else came close to Gallup's figures."

    Gallup, who is a devout evangelical Christian, has been quoted as calling his polling "a kind of ministry."  And, a few months ago, he said "the most profound purpose of polls is to see how people are responding to God."

    "We thought the purpose is to faithfully and factually report public opinion."

    Lots more.  Sorry no link.

  •  bias? (none)
    Quick someone call buckhead!
  •  Thanks to all who have brought media attention (none)
    to this issue.

    My question is.. at what point does the reputation of either Gallup or the media outlets that use it get damaged here, or does accuracy even matter - if nobody remembers what's reported about the polls or points out their flaws?  

    <CYNICISM>
    In other words, could a bunch of us theoretically get together, form some crack polling outfit that has a wide enough sample, have someone run some numbers to make up a "likely voter" model, and make shit loads of money getting some media outlet to pay us to produce the polls?  I mean, we would probably be wrong - as wrong as Gallup is wrong - but maybe the point of these polls is secretly just to give TV talking heads something to talk about, and not to be accurate or representative or informative about the actual electorate in any way, shape or form.
    </CYNICISM>

  •  good (none)
    Geez, Kos, you get any more press, this blog is going to implode from traffic!

    "Hope was on the way," her husband Gino Camabello said. "But she was too damn thirsty to give up her Diet Coke."

    by Hollywood Liberal on Tue Sep 28, 2004 at 12:30:34 PM PDT

  •  i wonder about these polls in general (none)
    and even if they're not accurate, then so what?  the only poll that matters is the one on 11/2 (barring any inaccuracies a la diebold).

    okay, so say that bush is leading in the polls, regardless of accuracy... what bearing does that have on anything?  is it that swing voters will bet on the perceptual winning horse?  i would think that if bush was leading, it would inspire those who want to vote against him more than those who would vote for him.

    an inquiring mind wants to know...

  •  Rasmussen can also be cited.. (none)
    ..to the media... after all.. he is supposedly a Republican pollster.. yet his daily poll shows only a 48-46 Bush lead.

    Our hopes are high. Our faith in the people is great. Our courage is strong. And our dreams for this beautiful country will never die. - Pierre Trudeau

    by tribe34 on Tue Sep 28, 2004 at 12:33:43 PM PDT

  •  How many independents support Kerry (none)
    I did a calculation...using the numbers from the article that Kos linked to.

    If one assumes that 90% of Republicans support Bush and 90% of Democrats support Kerry (or at least that those numbers are pretty close) around 60% of independents support Kerry.  If the % of Republicans who support Bush is greater than the % of Democrats that support Kerry, that would seem to suggest that even GREATER THAN 60% of independents support Kerry.

    If you keep the breakdown % and switch it to a 35% R 35% D and 30% proportion, Kerry would actually lead 53-47.

  •  speaking of taking on republican bias (none)
    David Brock is striking pre-emptively at Luntz and MSNBC;

    "Letter to MSNBC urging disclosure of pollster Luntz's GOP ties, questionable standards."

    http://mediamatters.org/items/200409280002

    Thank you Mr. Brock.

    (Clymer, Krugman and DailyHowler all pre-nailing the SCLM before the debates is all good).

    •  I knew luntz's focus group was a little fishy (none)
      A couple things I noticed.  I was extremely surprised by how they felt Bush was specific, yet Kerry was not.  I saw nothing of the sort.  

      The most striking thing I watched during Luntz's focus group segment: after Bush's speech, one of the members of the focus group said, "I really liked how Bush gave us his website to go to, so we can look at what he is proposing in greater detail".  He said this right after criticizing Kerry for not being specific enough, and not telling people where to go to for a better idea of Kerry's proposals.

      I don't know about you, but one of the specific things I remembered from Kerry's speech was that he GAVE HIS WEBSITE OUT AND URGED PEOPLE TO GO TO IT.  Did anybody else catch that?

  •  OT (none)
    Is anyone else finding the site to be glitchy today?
  •  Will USA Today Publicly Fire Gallup after this? (none)
    That's what should happen if they do find Gallups methods to be grossly negligent or biased on purpose.  

    "Reality" is the only word in the English language that should always be used in quotes.

    by LionelEHutz on Tue Sep 28, 2004 at 12:37:11 PM PDT

  •  Its a setup - now they fix the vote (none)
    What they are doing is priming the polls so a Bush Victory does not look strange...they will continue to do this until the last minute so when the fix is in the results won't look strange...

    JMHO

  •  This explains a recurrent right-wing meme (none)
    I laughed myself silly the first time my winger "friend" told me that the Rethugs had become the "majority party." This was right after their convention and in response to my scornful comments about Guiliani, Pataki and Schwartzenegger putting a centrist face on an extreme right-wing body.

    But since I first heard that "majority party" idea from my friend, I've been hearing it from other  right-wing and right-leaning types. The word seems to be out that the Rethugs have won the hearts and minds of the electorate and will now be the Big Tent Party for all eternity.

    Of course I don't think this is the case. There are several million rabidly anti-Republican gay voters, just to name one minority that isn't near ready to join this supposedly "majority party."  I honestly think there is more hostility toward Republicans among everyday voters than at any time in my lifetime. BUT this seems to be the Republican lie du mois. If this is the case, Gallup is just following marching orders.

    Watch the rest of the SCLM follow suit.

  •  Notice the trend though (none)
    You can see it here

    He still lost 3% even with the favorable sampling. I wonder if they noticed this early and decided to push the bias in favor of W?

    "We will not succeed in destroying freedom's adversaries if we are divided from freedom's friends." JK

    by Blacklantern on Tue Sep 28, 2004 at 12:38:33 PM PDT

  •  Here's the question (none)
    The magnitude of the change is historicand unprecedented. Should not Gallup be wondering if something is off-kilter?  As this is not recent as someare surminsing. It started in July.

    Some reasonable questions I think - what demographic is switching - white males, females, black males or females, hispanics? Asians?  Where is ths coming from?

    What region?  Why so many?

    IF you take the June numbers,which I believe were slightly majority Dem - in the 35-33 range, how can ou explain 12% of voting age Americans, 30+ million people deciding that they are now Republicans after being Dems?!?!?  It simply makes no sense.

    Has there ever been a party with over 40% ID in the voting population?  Recently, last 30 years, no.  Even the infamous LATimes poll had Dems at 38 and the GOP at 25 - the big move was a loss in ID for the GOP, not a forward movement by Dems.

    A good research organization would be taking a close look at this - Gallup has chosen the ostrich approach.

    "We're not criticizing Bush for going after terrorists, we're criticizing him for NOT going after terrorists." - Paraphrase of J. Kerry, tip to Gen. Clark

    by Armando on Tue Sep 28, 2004 at 12:41:19 PM PDT

  •  CNN Just Had It! (3.66)
    Woodruff just showed the moveon ad questioning Gallup and then had a representative on from Gallup who said all the polls show a 6-8 point lead, and "we are used to this skepticism".  Judy then said proudly "CNN and USAToday are partnered with Gallup this election season.".

    THE FIX IS IN.

    "Conservatism makes no poetry, breathes no prayer, has no invention; it is all memory." - Ralph Waldo Emerson

    by reef the dog on Tue Sep 28, 2004 at 12:43:05 PM PDT

    •  Keep up the emails (none)
      Folks, I just saw Frank Newport squealling like a pig underneath a fence on CNN about the complaints coming in about the Gallup poll. Now is the time to keep the pressure up. KEEP WRITING CNN, MSNBC, ABC NEWS AND everyone else. We have to get the word out that Gallup is an arm of the GOP - just like the rethugs made it out that CBS was for Kerry!
    •  Lie (none)
      All the polls show 6 to 8 point lead? What about the IBD/CSM poll today showing Kerry up 1? We should bombard CNN with that poll and question why Woodruff didn't follow up with the Gallup liar with that poll.

      GDoyle

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

      by GDoyle on Tue Sep 28, 2004 at 01:02:55 PM PDT

      [ Parent ]

      •  Woodruff just put up (none)
        a "poll of polls", with an average from about 8 new polls, showing Bush with an overall 6 point lead.

        So the reason for the Gallup is at least partially to skew the "poll of polls".  

        "Conservatism makes no poetry, breathes no prayer, has no invention; it is all memory." - Ralph Waldo Emerson

        by reef the dog on Tue Sep 28, 2004 at 01:07:15 PM PDT

        [ Parent ]

  •  This could be great news... (none)
    If gallup is raked over the coals and forced to change their methodology it will show Kerry making a mongo jump over the stretch of one poll.

    While informed poll watchers know it is not true, the general public will view it as Kerry being on the move!

    The Kerry camp will also be able to spin it this way.

    It is imperative to hold accountable.

  •  Gallup/Rove Poll (4.00)
    I think we should start calling it the Gallup/Rove poll.  Maybe Gallup will shape up if they see its once pristine name branded anew with a more appropriate, less non-partisan title.

    I can still remember Karl Rove touting Gallup when he was interviewed with Brit Hume prior to the Rethug convention - saying how Gallup has consistently shown Bush ahead and it is the best poll around.  Makes TOTAL sense now thatthis has come to light.

  •  Frank Newport Just Responded... (none)
    ...On INSIDE POLITICS.

    Woodruff brough up the ad from MOVEON & then gave FRANK NEWPORT, the GALLUP editor in chief, 2 minutes uninterrupted to respond to the charges. Newport never specifically responded to the party ID question but said...

    • George Gallup Jr.'s religious beliefs have no impact on the polls

    • The GALLUP poll is measure of where the electorate is today, & the party ID has no bearing on it

    • GALLUP's likely voter model is "very" accurate & nothing is wrong with it

    I don't agree with or believe Newport's assesment, but that's what he said. I'll post the exact quotes when the transcript is available.
    •  hahaha (none)
      he responded to nothing but straw men. MOVEON.org got them on this one. We need to expose these crooks!!

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

      by bigassbrando on Tue Sep 28, 2004 at 12:50:12 PM PDT

      [ Parent ]

    •  Deflect do not address... (none)
      The problem isn't Juniors beliefs or the likely voter model.

      The problem is that they have a sample that is weighted towards the Republicans.  Newport refused to address this issue.

      If I call 10 people at random I may get 5 Republicans and 5 Democrats, but I might also get 2 Republicans and 8 Democrats.  So polling companies use larger numbers to try and even that out (the law of large numbers will tend to bring the numbers back to a 50-50 split, if that were the breakdown of Repbulicans to Democrats in the country, i.e. leaving Independents out of this example).

      The people that Gallup called were 43% Republicans and 31% Democrats. This can happen every once in a while due to statistical fluctuations even in a purely random sample where Republicans and Democrats are at about even numbers. But this has happened regularly in Gallup polls (the last poll had Republicans outnumbering Democrats by 7%).  This makes it look like they are using a method for choosing their samples that increases the odds of getting a Republican respondent over a Democratic one.  This is not to say that they are intentionally choosing to do so as some part of a conspiracy (conspiracies in large companies would be impossible to keep quiet); it is just that some sampling practices are better than others.

      Since over the last three presidential elections Democrats have either outnumbered Republicans (39% to 35% in 2000, and 39% to 34% in 1996) or had even numbers (34% each in 1992), the Gallup numbers just don't strike people as representative of what is going to happen in this election.

  •  CNN Inside Politics (none)
    Just turned on my TV and what do I see but Judy W. giving the Gallup spokesman a chance to respond to the moveon.org ad.

    Of course there was absoulutly no followup questions by Judy or someone from Moveon.org.

    It was unbelievable.

  •  Is Limbaugh saying.... (none)
    ...what I think he's saying?  Help me out here.  I heard briefly a couple of things while waiting in the paint store today, to wit:

    Limbaugh opines that the polls are wrong.  Why?  Because there is a massive surge in new voter registrations, particularly among Democrats.  He went on to explain how these people wouldn't be taken into account by pollsters when doing their poll "weighting."  

    THEN he went on to talk about the newly registered Democratic voters.  He began to refer to them first as "illegal voters" and then as "cheaters"  -- the typical Limbaugh Orwellian use of a phrase which is morphed into different meanings.  New voters = illegal voters = cheaters.

    I heard only a small part of his diatribe -- maybe you heard more and can amplify? or correct me if necessary?

    •  Was he high? (none)
      If not, that would be unusual.

      GDoyle

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

      by GDoyle on Tue Sep 28, 2004 at 01:04:04 PM PDT

      [ Parent ]

    •  I wouldn't doubt it. (none)
      Limbaugh isn't a complete idiot.  In fact, in the area of propaganda he's quite effective.

      You have to remember that he's preaching to the Republican base.  Sure, some Democrats & moderates listen, but the Democrats just listen to get their blood pressure up and the moderates, hopefully, are more turned off than attracted by his bombastic ways.

      So, given that he's preaching to the base, he should be talking down these poll numbers.  If Bush is showing a significant lead going into the election it can depress turnout for both sides.  If he whips up the fact that it could be a phantom lead, he energizes the base to go out and vote.

      Just because a thought escapes his lips doesn't  make it a lie.  Good propagandists know when, where, and how much to bend.

  •  Atrios sez (none)
    Newport from Gallup was just on CNN "defending" itself against the Move On ad in today's New York Times. The incredible thing was that he talked about a bunch of mostly irrelevant details in the text of the ad, without addressing its central charge -- that Gallup is hugely oversampling Republicans.

    And Judy did not ask him about it.  CNN is in on it, this is the very definition of a conspiracy.

    Let's hammer 'em.

    "Conservatism makes no poetry, breathes no prayer, has no invention; it is all memory." - Ralph Waldo Emerson

    by reef the dog on Tue Sep 28, 2004 at 01:00:15 PM PDT

      •  CNN Contacts (none)

        Phone number:
        (202) 898-7900 (ask for Tom Hannon or the comment line)

        Feedback:

        General TV shows and anchors:
        http://www.cnn.com/feedback/cnntv/  

        Judy Woodruff
        http://www.cnn.com/feedback/forms/form4.html?18

        •  Ha ha! (4.00)
          I just talked to Tom Hannon in person!  First of all - what a fucking dick.  I mean first-class A #1 PRICK.  He says "Yeah this is Tom."  I say "Yes I'm calling about the Gallup polls you're using?"

          Hannon: (dripping with sarcasm)  "Oh, you mean the ones that DON'T show Kerry in the lead?  The ones that DON'T show Kerry up 8?"

          Me:  Um, excuse me, are you mocking me?

          Hannon:  What do you mean, am I mocking you?

          Me:  I just called to get a quote and you're mocking me!  You just did!"

          Hannon: Where are you from?

          Me:  I'm from Political News Magazine and I called for a quote and you immediately started mocking me.  You don't even know which side I come down on on the issue and you... you know you're representing CNN right?

          Hannon:  What, are you recording this?

          Me: Yes.  And you openly mocked me.  And you have a problem.  The Gallup poll is weighted incorrectly.

          Hannon: There is no weighting.  The poll has no weighting.

          Me: What are you kidding?  It's weighted with a 13% bias towards Republicans.

          Hannon:  No it isn't.  It's not weighted.  Are you saying Gallup is at fault?

          Me: No, I'm saying the media is responsible. It's irresponsible of CNN to report these polls without letting people know it's hugely biased towards Bush.

          Hannon: You aren't allowed to call CNN irresponsible.

          Me:  WHAT?!  Are you kidding?  You're representing CNN and you're telling me people are not allowed to call you irresponsible?

          Hannon: That's not what I said.  I said you're not allowed to call Gallup irresponsible.

          Me: No you didn't.  You said I wasn't allowed to call CNN irresponsible.

          Hannon:  Yeah well.  Is there something you wanted to say?

          Me:  Just wanted a quote.  I got some good ones, thanks.

          Hannon: (click)

          "Conservatism makes no poetry, breathes no prayer, has no invention; it is all memory." - Ralph Waldo Emerson

          by reef the dog on Tue Sep 28, 2004 at 02:53:44 PM PDT

          [ Parent ]

  •  new pew poll (none)

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

    by bigassbrando on Tue Sep 28, 2004 at 01:03:44 PM PDT

    •  Well... (none)
      I guess the debates are the last chance for Kerry to counter some of the stupid perceptions the GOP has managed to put out
      •  Breakdowns a Joke (none)

        Just look at the breakdowns. They claim that Kerry's not gaining, but practically all of his internals are up. And they don't provide a party ID breakdown. You can bet this is just as heavily weighted as the Gallup poll.

        Is the Christian Science Monitor the only honest news source left in America?

        CNN - about as "trusted" as a compass in an active MRI machine.

        by RHunter on Tue Sep 28, 2004 at 01:42:36 PM PDT

        [ Parent ]

        •  Naw (none)
          Look, MOE and high undecided numbers - there is really little difference between 46-46 and this result - the internals show great opportunities for Kerry.

          I like this poll actually.

          "We're not criticizing Bush for going after terrorists, we're criticizing him for NOT going after terrorists." - Paraphrase of J. Kerry, tip to Gen. Clark

          by Armando on Tue Sep 28, 2004 at 02:39:51 PM PDT

          [ Parent ]

    •  Interesting poll actually (none)
      First, the movement, over 3 polls in 2 and 1/2 weeks, - Bush's range 46-48,Kerry 40-46, Undecideds - 10-12 - harbors good things- Bush has a ceiling - and it is about 48.

      Kerry is in the ballgame precisely because of that.  

      And look at the internals - women only 45-42 for Kerry, men favor Bush by 52-37, 3 for Nader, 8% undecided. Kerry has 13% women undecided.

      The undecideds are there for Kerry - the swing is those jumping on with Kerry and jumping off - Bush is barely moving the needle now - hehas no more room.

      Kerry has 13% undecided blacks!  13% undeicded women!  Even 8% white males are likely fertile Kerry ground.

      This poll is a good poll,and an accurate reading in many respects, taking into account the margin of error of 3 for each candidate -thus 46-46, 48-40, in a poll that clearly does not push leaners - where Kerry is the challenger, and undecideds break for the challenger 3-1.

      "We're not criticizing Bush for going after terrorists, we're criticizing him for NOT going after terrorists." - Paraphrase of J. Kerry, tip to Gen. Clark

      by Armando on Tue Sep 28, 2004 at 02:37:00 PM PDT

      [ Parent ]

    •  Troubling (none)
      The Pew Poll is especially troubline because it has been showing movement to Bush.

      Just a few weeks ago it showed no more than a 2 point lead, IIRC

  •  let's protest Gallup!! (none)
    FRIENDS, the election may depend on this

    We need to hammer not just the media but Gallup til they beg us to stop. Let's show some strenght in supporting the truth.

    http://www.gallup.com/help/contactUs/

    General/Corporate Inquiries
    For general comments and questions about The Gallup Organization, contact:

    The Gallup Organization
    Corporate Headquarters
    901 F Street, NW
    Washington, D.C. 20004

    Corporate Campus
    1001 Gallup Drive
    Omaha, NE 68102

    Anyone up for a protest outside of their headquarters? Fronted by Ruy Texeira?

  •  By the way... (none)
    ...and with respect to the Gallup Poll unmasking, we need to start appreciating the effect of blogs on the news -- positive effect.  I'm not talking about the oh-boy-oh-boy talk in the NYTimes magazine article; I'm talking about a substantive shift in how the mainstream media are responding in general to public opinion thanks to some hardworking, knowledgeable and relentless bloggers.  I don't think Kos and kin are working in a kind of underground anymore; I think they've achieved the right kind of visibility.  Congratulations to Kos for the new Guardian gig -- but I hope it doesn't seem more important than what Kos is doing now because it surely isn't!
  •  Anyone besides me... (none)
    ...think it was a mistake for MoveOn to comment on George Gallup's religious affiliation? It may have something to do with the bias in their poll, but it is a comment that could be easily grabbed on to by the right and used to obscure the issue.

    "MoveOn attacks Christians!"

    Ancient Chinese curse: May you live in Interesting Times

    by Chris Andersen on Tue Sep 28, 2004 at 01:11:05 PM PDT

    •  No... (none)
      The ad isn't for the media or for Christian extremeists, it's for Kerry voters that are getting depressed by the polls. The right-wing noise machine is trying to keep them home by making it seem like the election is hopeless. If they learn the polls are bogus, they'll be less disheartened.
  •  Ahhh, quotes exposing gallup.... (none)

    SOURCE: Electing the President 2000: The Insiders' View. Election strategy from those who made it

    The book includes interviews from Bush 2000 campaigners Karl Rove, Mark McKinnon, Alex Castellanos, Matthew Dowd, Fred Steeper, and Lionel Sosa and from Gore 2000 campaign people Carter Eskew, Bob Shrum, Stan Greenberg, and Bill Knapp. Kathleen Frankovic from CBSNews was also a participant.
    ----

    Quote from Bob Shrum (p.112):

    SHRUM:
    ...But there was a general tendencey, especially with the CNN poll and the Battleground poll, to have huge massive unexplained shifts in one day, which the Bush people must have found amusing, too.  One morning, Tad Devine said, "I wonder if 19 million people got up this morning and changed their minds about who they are gonna vote for for president?

    The quotes from Stan Greenberg (p.113):

    GREENBERG: I watched Frank Newport of Gallup go on TV with a straight face and say the reason for this variation is because the voters change their minds from day to day. I will get you the quotes.

    There's also a quote where CNN and USAToday didn't do party ID weighting in 2000 either.

    Bill Knapp on reporting and polls (p. 172):

    KNAPP: One of the things coming out of this election that does need to be looked at is that if you spend the time polling the right way, if you spend the money to poll the right way, if you're patient to poll the right way, you can get the right numbers. Stan had the right numbers. I'm sure the Bush campaign had the right numbers. The networks, the news outlets don't have the right numbers. And they need to step back and figure out what they're going to do about it, or they'll continue to pervert the system. [...] I just want to make another point about the reporting. On September 11, Gallup had Gore 49, Bush 42. That's a seven-point spread as I count it. It was reported as Bush/Gore deadlock. Go to Ocotber 8: Gore 41, Bush 49, eight-point lead for Bush. Headline: Bush Maintains Solid Lead Over Gore. October 28: Gore 42, Bush 49, seven-point lead. Same as the September 11 lead, except reversed. Bush is reported as having a solid advantage.  We ahve a five-point lead on September 17. That's a "narrow lead." My point here is, the analysis is totally at variance with reality. I'm not picking on CNN ... Well, maybe I am. I'm picking them on purpose. But the poitn here is, you can see this in poll after poll and outlet after outlet.

    ----

    Apparently, no one took Mr. Knapp's advice.

  •  Making some money (none)
    Anybody planning to or trying to take advantage of the skewed polls to make money wagering on the election?  I mean, of course not really, it's illegal to wager here in the U.S.  But if you are overseas or something, this is how you make a killing at the sportsbook -- when you have "inside" information that the public perception of the contest is skewed.

    I just happened to notice a futures wagering site offering Kerry at +200 (bet $1, win $2).  Seems like a pretty good bet to me -- Kerry is way more that 33% likely to win.

    Don't get me wrong, I'm not completely self-interested.  I'm also going canvassing for Kerry in the swing state next door.  But if Gallup is going to be misleading and corrupt about this, I think I should be allowed to make money off it.  If only internet gambling were legal where I live. :-)

  •  Let's show Gallup... (none)
    what the full might of daily kos can do! Blast them, folks! I haven't seen a form letter to send the media. Has anyone been using one that they could put up to share?

    I don't have time to write but I can make time to blast emails from work.

    LET'S BEAT THESE BUMS!!!

    -fink

  •  Parsing the Gallup Poll (with help from CBS) (none)
    Here's what I get when I map out Gallup's procnostigation (sic, a poll that bends scientific method over a log, a la "Deliverance") with the latest CBS's latest by-party breakout of Bush job approval ratings.

    (I ran the results through an Excel spreadsheet Solver, and yes, this is more a thumbnail sketch with numbers than real polling...but you make up your own mind)...

                                   Bush    Kerry   Undec
    Pub (43.6% of sample)     86.8%   11.9%     1.3%

    Dem (31.3% of sample)     13.5%   85.4%     1.1%

    Ind (25.1% of sample)     39.7%   48.0%    12.3%

    Resulting in              52.0%   44.0%     4.0%

    Personally, I'd love to see Gallup's actual by-party breakout and talk shop. In lieu of that, I'm making the guestimates above.

    Play around with the sample percentages. While I doubt that there really is a precise triune parity between Dems, Inds and Pubs, it's interesting to see what the results are in that scenario:

    Because it comes out Kerry 48.4% to Bush 46.7% when you play fair and weigh the factions equally.

    Check it out.

    Truth gone to seed is dangerous, but well-cultivated ignorance is deadly.

    by cskendrick on Tue Sep 28, 2004 at 01:41:05 PM PDT

  •  Gallup picks the debate audience. (none)
    Don't forget, Gallup picks the debate audiences. Guess how they'll be weighted..
  •  CNN can't defend its use of its own poll. (none)
    In a call to Tom Hannon, CNN's Executive Director of Political Coverage, Mr. Hannon was decidely aggressive when asked why CNN was reporting a CNN-USA Today-Gallup poll that oversamples Republicans by 12 percentage points (43% Republicans to 31% Democrats). The bias in the poll sample was pointed out by Tad Devine, but CNN continued to quote the poll throughout the day without making note of this bias. The biased sample in the poll, which unsurprisingly gives Bush a lead over Kerry, is questionable in that over the last three presidential elections Democrats have either outnumbered Republicans (39% to 35% in 2000, and 39% to 34% in 1996) or had even numbers (34% each in 1992).  

    Gallup sent out Frank Newport to deflect these criticisms of their poll (most notably made by a MoveOn.org add in the NY Times today). Mr Newport attacked the style of the ad, but not the ad's substance. Not once in his statement did he talk about the sample bias.

    CNN is on an shaky ground in using the poll, as was apparent in Mr. Hannon's who used the same method of attacking the questioner that Mr. Newport used, rather than explaining why they continue to use the poll without disclosing the sample bias. The sample bias is, in fact, not included in the information on the CNN website (as of 1:30 pm on 9/28/04), one must seek the information out directly from Gallup. Mr. Hannon asserted that only those with "expertise in polls" would understand why such a stilted sample bias was appropriate. But pollster John Zogby disagrees. In a statement he describes that the bias markedly affects the results, "Given the fact that each candidate receives anywhere between eight in ten and nine in ten support from voters in his own party, any change in party identification trades point for point in the candidate's total support."

    I urge you to call CNN and ask them to either objectively report the bias in the sample of their poll, or stop using the poll. CNN's phone number is (202) 898-7900, you can either leave a comment with the comment line, or ask to speak with Tom Hannon directly. Another option is to contact your local papers and ask them if they are going to look into either CNN or USA Today's use of the poll.  

    •  Done (none)
      I left a message on the CNN viewer feedback line.
    •  Good idea--push back (none)
      I've long wondered whether Gallup is genuinely confused about how to poll accurately, or they are basically conducting something like a push poll...pushing in the sense that they are trying to push the electorate reading their poll results onto a fictitious bandwagon.  On slate.com Chris Suellentrop has a good piece today on the Republican attempt to manufacture an aura of inevitability for their dimbulb.  Anyhow, the more I hear about the Gallup crowd the shadier they sound.

      Since the 1930s, they have pretty consistently underestimated the popularity of Democratic candidates, as e.g. Dewey vs. Truman.  Part of this may be their willingness to oversample Republicans, but there are other factors as well.  The most significant, I think, is their crude grid for determining 'likely' voters.  They are aggressive in eliminating voters, for example by chucking out the responses of people who do not happen to know where their polling place.  I've lived in cities where it was uncommon for a polling place to remain the same from one election to the next.  I'd guess that this part of their grid strongly disfavors urban voters' opinions.

      Anyway, there is another factor that I think is undermining the accuracy of nearly all polls.  They take about half an hour to complete, and there are more polls than ever this year.  It appears that fewer people than ever are responding to pollsters this cycle, though the polling outfits aren't happy to advertise that fact.  They have to call about 4000 to 5000 people just to get 1000 completed polls.

      Therefore the pollsters' 'margins of error' are a polite fiction.  Even a very small factor in making people unwilling to respond to a poll can have a large effect on the results they report.  For example, people holding down two jobs aren't likely to gab for 30 minutes to strangers.  Also, I would guess that in an election as bitter as this, more people would refuse to talk to pollsters.  The very worst of polls in 2000 were conducted by Rasmussen; these were computer generated calls, no human interviewer, and I'd bet they had a prodigious hang-up rate.

      In short: a long, cold shower for the horse-race numbers of any poll.  The best predictor continues to be the job-approval numbers for Bush, which fairly consistently have been under 50%.  That would suggest that Bush will lose unless the other 50% suddenly find something to like about him.  I can't think what that might be.

  •  They all go to the same poll spinmeister school (none)
    I heard Frank Newport of Gallup on NPR a week or so ago defending his poll and he reeked of Republican spin. Even mentioned the Swift Boat Liars (okay, he didn't call them that), just so he could get in another dig at Kerry.

    He must have been trained at the same time as the Newsweek poll spinmeister over Labor Day weekend, who was wondering whether Kerry could come back from his 11-point deficit.

    These poll spinmeisters even have excellent voices.

    Folks, these people will stop at nothing.

  •  Bombard CNN (none)
    force them to capitulate, like the freepers did to CBS. Go to cnn.com, scroll down to the "contact us" page, write a short letter, and fire away with your cut and paste function.

    The United States has a conservative political culture defending a liberal heritage. The modern Republican Party's problem is that it is neither.

    by Ben P on Tue Sep 28, 2004 at 02:33:48 PM PDT

  •  ok ok ok but... (none)
    what about
    WAPO-ABC News Poll: Bush Ahead by 7

    Bush leads Kerry 51-44 percent among nation-wide RV's, with 2 percent for Nader, 2 percent other/neither and 2 percent no opinion, according to a Washington Post-ABC News poll conducted 9/23-6.

    This poll is weighted about right, and the nthe PEw poll.

    Honestly, i like us going after Gallup, its good work, but i am going to continue campaigning like these polls are right and we are down significantly.

    Maybe its all a Liberal media plot to show kerry down a bunch, then after the deabtes, fix the internals and show him surging ? LOL

    •  We are down (none)
      But Gallup is overestimating the deficit. Go to pollingreport and do the math, and throw in Rasmussen number from today. The result: Bush 48, Kerry 44. So Kerry trails, but Bush is under 50. Don't make any bets or declarations yet.

      I'd wait until 11/2, especially this year.

      The United States has a conservative political culture defending a liberal heritage. The modern Republican Party's problem is that it is neither.

      by Ben P on Tue Sep 28, 2004 at 02:50:08 PM PDT

      [ Parent ]

  •  Pew (none)
    For what it's worth, the Pew poll shows ShrubCo ahead by 8 as well, according to Reuters.

    Wonder how much their skew is?

    "Flip-flopping is a Conservative insult meaning 'thinks about the issues.'"

    by Bob in Atlanta on Tue Sep 28, 2004 at 02:38:05 PM PDT

    •  Well (none)
      Its not a question of the horserace DIFFERENTIAL, per se. Bush's horserace number is at 48, which suggests their polling is more accurate. Bush's horserace number is what most matters in rating these things, IMO. I'll care more about Kerry's in a months time.

      Ben P

      The United States has a conservative political culture defending a liberal heritage. The modern Republican Party's problem is that it is neither.

      by Ben P on Tue Sep 28, 2004 at 02:47:35 PM PDT

      [ Parent ]

      •  Correct (none)
        High undecideds, and with MOE - 45-43 is just as likely, shen you consider the high undecideds, who the undecideds are, etc.

        One thing,if Kerry is up 45-42 with women, and Nush 52-37 with men, if it breaks down 50-50, then the number should be 47-41.  But if memory serves, more WOMEN than men vote, so it should be closer than that, in this poll

        "We're not criticizing Bush for going after terrorists, we're criticizing him for NOT going after terrorists." - Paraphrase of J. Kerry, tip to Gen. Clark

        by Armando on Tue Sep 28, 2004 at 03:05:18 PM PDT

        [ Parent ]

    •  What's up with the wild swings (none)
      in the Pew poll?

      Big Bush lead, then, tie, and now big Bush lead.

      No major events would have caused any of that wild fluctuation. That alone should tip any poll watcher off. I'm surprised Pew would release polling like that.  

      I'd bet money that the polling done by both campaigns show this race as a statistical tie nationally, with the electoral college pretty close, too.

      •  Not so wild (none)
        Actually within the MOE - Bush ranges from 46-48, Kerry a bit more 46-40.  The numbers move mainly from undecided to Kerry.

        No leaners, big undecided - I actually like this poll frankly.

        Last point, I think there is a math problem too - Kerry leads 45-42 with women, trails 37-52 with men - if it is 50/50 split, then the number should be 47-41.

        "We're not criticizing Bush for going after terrorists, we're criticizing him for NOT going after terrorists." - Paraphrase of J. Kerry, tip to Gen. Clark

        by Armando on Tue Sep 28, 2004 at 03:07:28 PM PDT

        [ Parent ]

  •  "Grassy Knoll Polls" (none)
    OK, buckle your seatbelts for a bumpy ride into conspiracy-land.  

    Did you ever think there might be a concerted effort to skew polls in favor of Bush before the election, so that when Kerry actually does win, it gives folks on the Right a reason to scream foul?  Think about it, they can say, "Bush was ahead by double digits on Sunday, but lost by 3 percentage points.  You simply cannot get a 13 point swing in three days..."  Or, if, God forbid, something happens, they can remove the of legitimacy from Kerry's win by blaming terrorists (domestic or international) for trying to affect the election by saying "Bush was leading right up to the attacks"?  

    Can't you just hear them on cable news now explaining that Kerry isn't the real president, and that foul-play must have occurred, therefore, nobody should accept the results?  

    OK, we are safely back from fantasy-land (with a nod and a wink).

    I used to be paranoid until I lost my self-esteem. Who is going to waste their time following me?

    by Mote Dai on Tue Sep 28, 2004 at 02:52:01 PM PDT

  •  Alrighty - Just how many Republicans are there? (none)
    There is a paper titled "Turnout Effects on the Composition of the Electorate:
    A Multinomial Logit Simulation of the 2000 Presidential Election", by a fellow named Michael D. Martinez.

    Whew. That's a mouthful!

    Table 5 of the aforementioned paper is a sketch of voter composition of the electorate as a function of turnout. The analysis is based on the 2000 election.

    Per the public record, there was 54.7% turnout that year. Per Martinez, the breakout by party affilation was as follows:

    Strong Democrats 22.2%
    Weak Democrats 15.6%  - Total Dems: 37.8%

    Leaning Democrats 12.8%
    Independent 7.3%
    Leaning Republicans 12.8% - Total Inds: 32.9%

    Weak Republicans 12.2%
    Strong Republicans 17.0% - Total Pubs: 29.2%

    Note: I've added the subtotals. Other parsings of the political population are possible. (Whew. That's a lot of alliteration!)

    Every other day I hear: The world changed on 9/11.

    Did the poles reverse that day? I doubt it, otherwise the current occupier of the Oval Office would have never sweated his prospects for staying in office, ever again.

    I'm thinkin'...yep. Gallup up and galloped into its own horse poop.

    Truth gone to seed is dangerous, but well-cultivated ignorance is deadly.

    by cskendrick on Tue Sep 28, 2004 at 02:52:05 PM PDT

  •  There are two explanations for these poll numbers (none)
    As usual, it comes down to either incompetence or evil (or possibly both).

    First, incompetence:

    That is, Democrats are harder to poll via the telephone.  This is quite possible.  I would think that Democrats are more likely to be college students in dorms without permanent phones, people with cell phones only (young and/or technologically with it), and/or people who don't speak English well enough to answer a poll.  So not weighing misses these people and provides a skewed poll.

    Second, evil:

    They are faking the polls to assist in Bush stealing the election.  Considering who runs Gallup (and some of the other polling firms), this is quite possible.

    Now, party ID can change, but events that happened in the past can't.  A Time poll had internals that, when they asked the respondents who they voted for in 2000, they voted Bush over Gore by 12%, instead of Gore over Bush by .5%, which is what really happened.  Not concidentially, the poll showed about a 12 point lead for Bush.  This poll had bias that couldn't possibly be argued about-since the D/R/I numbers were bad there too, this can probably be carried over to other polls like this.

    The fact that we can figure out what happened from the internals leads me to believe that incompetence is the main or only cause here-otherwise they could have faked the internals, too.  Hell, what proof do we have that Gallup or whoever called anybody?  They could have pulled the numbers from thin air if they wanted to, and showed Democrats weak for Kerry and Independents and Republicans strong for Bush, and the numbers might be more believable.  However, if the fix is in, these numbers, even if caused purely by incompetence, still help Bush get away with it.

    Now, what this also means is that if the fix ISN'T in, and Kerry wins big, the SCLM might realize that they are barking up the wrong tree, and they might start questioning the Thugs more, and show less of a bias towards them.

  •  The Plot thickens (none)
    I don't have a link but I just listened to a spot on NPR (Market Place) which covered the Move-On NY times Ad challenging the accuracy of Gallup.

    It covered the basic disgreeing views but finished with this tidbit:

    Gallup gets to pick the audience for the town hall style Presidential debate.

    This is a huge deal and I think we may need to start some sort of campain to protest there inclusion, or at the very least to pressure them into nutrality.

  •  asdf (none)
    I think it comes real close to fraud if they are claiming that their polling sample is likely election results if they know that that is not possible.  I believe there MOE is something like +/-3 or 4.  

    If there model has no basis in reality, then there motives become suspect.  If they are posting questionable results in the hopes of influencing popular opinion, I think you can argue fraud.

    Then again, I am most definately not a lawyer.

  •  Gallup's biases (none)
    Can anyone comment about the just-released Pew poll giving Bush an 8-point lead, up from 3 in mid-September?  At the same time TIPP released a poll showing Kerry 1 point ahead?  At the same time Rasmussen gives Bush a whopping 1.6% lead?  

    Can anyone intelligently comment on the biases of these 3 polls?

    Paul in Austin

    •  Well .... (none)
      It all depends on what poll you want to beleive. Remember that none of these polls are taking into account new voters. I cannot speak to the internals of the other polls, but Gallup fucked up - period - and I think it was on purpose. They weighted GOPers heavier then they did in the last poll, showing a 12% heavier advantage. I do not understand why the don't go with even numbers for DEMS, GOP and IND.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site