Patterico's Pontifications


CNN Poll: CNN Has a Lot of ‘Splainin’ To Do

Filed under: General — Patterico @ 7:24 am

That CNN poll showing Obama way ahead seems suspect. Here is the .pdf of the poll results with all the breakdowns.

The poll’s big news, reported yesterday, was that Obama was coming out ahead 52% to 43% among registered voters, and 53% to 41% among all voters, in response to this question:

Suppose that the presidential election were being held today and you had to choose between Barack Obama as the Democratic Party’s candidate, and Mitt Romney as the Republican Party’s candidate. Who would you be more likely to vote for — Barack Obama, the Democrat, or Mitt Romney, the Republican? (IF UNSURE:) As of today, who do you lean more toward?

But look at the numerical breakdown:

484 to 476, out of 1015 respondents.

484 out of 1015 is about 48%, and 476 out of 1015 is about 47%. (It can’t be registered voters because they say they have only 910, and 484+476 is more than 910).

This means the numerical breakdown, before weighting, is 48/47. Somehow their weighting process takes it to 53/41.

It’s not explained why the weighting is so dramatic. In fact, I can’t find any explanation of their weighting process.

This was pointed out to me by Kevin M, who notes that the poll is an outlier anyway. I see that Hot Air is performing the same analysis with the same results. Ed Morrissey says:

Pollsters weigh results in likely-voter models; in fact, that’s a vital part of the likely-voter modeling process. I’m not aware of the need to weight gen-pop polls unless the demos in the survey are wildly deficient, and since CNN didn’t disclose those, we don’t know — but it hardly imbues these results with confidence.

At the very least, CNN needs to answer some questions about these odd results.

61 Responses to “CNN Poll: CNN Has a Lot of ‘Splainin’ To Do”

  1. how dare you question teh narrative?

    you’re not even a journalist…

    redc1c4 (403dff)

  2. Hide the decline and massage the data until it fits what we expect it to be.

    Robocop (8efdaa)

  3. Why do they even bother with the polls?

    JD (becb30)

  4. You are assuming that every person who chose Obama or Romney was asked questions 3 or 4. What makes you assume that?

    Kman (5576bf)

  5. because they are still living off the rep they had with the Gulf War, when there was no competition,

    narciso (8d0f34)

  6. @Kman: why would they *not* ask the follow on question?

    redc1c4 (403dff)

  7. Scud Studs Unite!

    daleyrocks (bf33e9)

  8. “You are assuming that every person who chose Obama or Romney was asked questions 3 or 4.”

    Kman – You are trying to rationalize bogus poll results.

    daleyrocks (bf33e9)

  9. Good Allah, kmart.

    JD (becb30)

  10. there can’t possibly be that many Rs voting for Romney

    happyfeet (3c92a1)

  11. As I noted to Ed Morrissey on Twitter, that’s not the only screwy thing about the poll. Page 9 of the report gives crosstabs on the Obama-Romney matchup. The Margin of Error given for Democrats and Republicans is +/- 6%. The MoE give for Independents is only +/- 4.5 percent. Unless I’m missing something else, that would imply that CNN has fewer than 400 Ds and Rs, and somewhere around 500 Indies, in the RV sample. In actual elections, Indies are a smaller group than Ds or Rs.

    Karl (6f7ecd)

  12. Shorter CNN–It’s for the sake the polar bears and the chillun so whatever we have to do, it’s all goood.

    elissa (01e1a0)

  13. Kman– Because they only had 1015 people in the poll, and 5-6% responded no preference or other. 1015 less (484+476 or 960) = 55 which is that 5-6%.

    So, those 960 represented the class of adults who responded a preference, and they broke 48-47 in raw numbers. We have no info on how registered voters broke in raw numbers but it might have been FOR Romney since registered voters tend to vote more Republican.

    Note that Gallup and Rasmussen have Romney leading by 2-3 points among registered voters and you can see why this poll seems so, well, out there.

    Kevin M (bf8ad7)

  14. I am pretty sure that kmart’s question was not remotely in good faith, Kevin. You gave him far more credit than he has earned.

    JD (becb30)

  15. So Willard is tied with Dog months out, ahead of the conventions, when we know there will be a ten point rush to the challenger beginning Halloween as DWTS and Idol is crowded out.

    So why is the Juggernaut worrying about Indies and wimmen? The move to the left is supposed to come after convention.

    The party of Stupid is really hitting the canvas hard.

    gary gulrud (1de2db)

  16. 11. Population versus turnout?

    gary gulrud (1de2db)

  17. 13, 14. Thanks for the dumbing down, anyway.

    gary gulrud (1de2db)

  18. Karl–

    Yet another odd thing is the 473 “Republicans” they mention on page 8, with a 4.5% sampling error. I guess they were counting wannabes, but it’s really unclear. I also note on page 9 that “Tea Party Supporters” have a lower sampling error than Republicans which implies the Tea Party has more support than they might otherwise let on.

    Kevin M (bf8ad7)

  19. I am pretty sure that kmart’s question was not remotely in good faith, Kevin. You gave him far more credit than he has earned.

    Just the same….

    Kevin M (bf8ad7)

  20. Gary–

    They are worried because they fear a landslide. As they should.

    Kevin M (bf8ad7)

  21. The media is doing what it did when Ronald Reagan went up against incumbent Jimmy Carter.

    AZ Bob (1c9631)

  22. If the voters with no preference vote (45 it seems) they likely go Romney 2:1. As I see it if you like POTUS you like him. You are not undecided. Those 45 don’t like POTUS.

    So if it goes 2:1 then assume 30 more votes for Romney vs 15 for POTUS = 506 MR vs 499 POTUS.

    Anyway, bigger issue is that we are tied, economy not getting better, Romney has yet to aims his guns at POTUS in any meaningful way or speaking the undecideds in any meaningful way..

    Bill (af584e)

  23. What I take from this poll is one way or the other this election is a referendum on Obama and his Presidency. There are many more people participating in this survey that are either voting FOR or AGAINST Obama than there are voting FOR or AGAINST Romney. The bottom line is if this is the best candidate the GOP can run it’s not very inspiring.

    Ralph Gizzip (5ab3ea)

  24. they didn’t just cook the results, they fried them
    in deep fat, like they do with everything in Iowa,

    narciso (8d0f34)

  25. Mittens isn’t the best candidate the GOP can run, but he’s the one the RNC and the MFM want to run.

    redc1c4 (403dff)

  26. Well, that’s a given, but you see what they do to him regardless,

    narciso (8d0f34)

  27. 19. I get that of the Beast’s base, Race Hustlers, Union bosses, and GaiaDruids will grace the polls with bells on, most of the rest, not standing in line for free stuff, will be wearing paper bags.

    gary gulrud (1de2db)

  28. #23: Another way to look at is is that Romney has low negatives compared to Obama.

    Kevin M (bf8ad7)

  29. 23. The WildRose party in Alberta, founded 2007, has a 20 point lead over the Tories closing in on elections.

    We can easily dispatch Obamanation and launch a Wildcat run in November.

    gary gulrud (1de2db)

  30. I always trust content from CNN. Why do we need to trudge to the polls and spend billions in these antiquainted elections, when an honest, honorable news organization such as CNN can give us the results by intwerviewing a 1,000 or so people at no cost to us except a fraction of our cable bill?

    nk (875f57)

  31. The obvious explanation is that CNN is “weighting” the numbers to allow for all of the dead people, illegal immigrants, and people who will undoubtedly be voting two, three, and four times for Obama.

    For obvious reasons, those “voters” cannot be included in the tally of “registered voters” in the sample.

    Elephant Stone (0ae97d)

  32. Why don’t they show their work?

    SarahW (b0e533)

  33. it’s all very opaque

    happyfeet (a55ba0)

  34. The pollsters work in the same way as appellate judges: they know the outcome they desire, so they go looking for a way of getting to that predetermined end result.

    And CNN does this, because the initial headline (‘Obama ahead by 9!’) is what registers among people paying less attention. And this is important because of the bandwagon effect… people who are somewhat undecided will be swayed knowing that most people pick Obama.

    We don’t trust CNN’s news coverage, why should their polling department be any different?

    steve (369bc6)

  35. You are assuming that every person who chose Obama or Romney was asked questions 3 or 4. What makes you assume that?
    Comment by Kman — 4/17/2012 @ 7:49 am

    — From the mouths of babes . . . Yes, why would you assume that they would be honest & forthright & consistent & true?

    They’ve had no training for that!

    Icy (ad6a1f)

  36. Why don’t they show their work?
    Comment by SarahW — 4/17/2012 @ 9:39 am

    — It’s a thing with peoples on their end of the spectrum: the transcripts are sealed; gotta pass it so you can see what’s in it, etc.

    Icy (ad6a1f)

  37. the transcripts are sealed

    It’s funny that Obama didn’t have that same respect for sealed documents when he was running for the Senate — as Jack Ryan (no not that Jack Ryan, this one) can attest.

    Kevin M (bf8ad7)

  38. Hell, this is nothing. Wait until after the Dem convention. CNN will report at that time that Obama is up by 15 points. The irony is that a material percentage of conservatives actually will believe it.

    Incidentally, the day prior to the 2004 election CNN/Gallup/USA Today came out with their final polls. They had Kerry up both in Florida and in Ohio, thereby winning the election. In Florida, specifically, they had Kerry over Bush by a margin in excess of the poll’s MOE. Oddly enough, however, Bush managed to eke out Florida that year, 52-47. Go figure.

    Tsar Nicholas II (cb2d5b)

  39. Yees, they lie, water is wet, yet Soledad is ‘soaking in it’

    narciso (8d0f34)

  40. At the very least, CNN needs to answer some questions about these odd results.

    Three words:
    “Fake but accurate”

    Smock Puppet, Official CNN Poll Analyst (8e2a3d)

  41. #23: Another way to look at is is that Romney has low negatives compared to Obama.

    True, but, really… how hard is that?

    I mean, a starving, feral tiger astride your prone body with his paws on your chest and his hot smile in your face, has “low negatives” compared to the President of All 57 States.

    You might be able to shoot the tiger and survive…

    Smock Puppet, 10th Dan Snark Master (8e2a3d)

  42. #23, 41: Yet another way to look at the numbers is just how absolutely polarizing Barack Obama is. Few people have a neutral opinion; they either love him or hate him.

    And to think McCain called him the most partisan of Senators. Obviously a smear.

    Kevin M (bf8ad7)

  43. The abo’s will do all the heavy lifting. And make this a landslide.
    Cnn, laughable.

    sickofrinos (44de53)

  44. Yeah, it wouldn’t surprise me if the CNN methods are sealed.

    The birth certificate is sealed, the college transcripts are sealed, the applications for college loans are sealed, the college senior thesis is sealed, the manuscripts for the books that Bill Ayres (first one) and various staffers (second one) wrote are sealed, the videotape of Obama toasting Rashid Khalidi is sealed, Nancy Pelosi’s “we have to pass the bill to find out what’s in it” but until then the bill is sealed, the Solyndra papers are sealed, the Fast & Furious correspondences are sealed…everything in this man’s life is sealed—except for Obama’s mouth, when he’s unwittingly getting caught talking trash in an open mic about Netanyahu or about post-election “flexibility” regarding missile defense.

    …but just remember, this is “the most transparent adminstration,” evah !

    Come next January 20, I hope he’ll be taking a vacation…to the Maldives ! I hear they speak “Austrian” there.

    Elephant Stone (0ae97d)

  45. I found that if you add the pro Obama votes (484) and pro Romney votes (476) you get 960, which leaves 55 undecided voters. If you add those 55 undecided voters to Obama’s Total you get 539 or 53%. If you subtract those 55 undecided votes from Romney’s total you get 421 or 41%. I cannot help admiring there creativity.

    Dave G (39b624)

  46. I much prefer CNN faking poll results than faking the news itself. I still can’t get my arms around the fact that our window into pre-2003 Iraq provided by CNN was fundamentally fraudulent. Worse, it wasn’t that Saddam was editing the news – it was CNN’s fear of getting kicked out driving them to self censor bad Saddam behavior from their reporting.

    Then they hid Saddam’s plan to kill a relative who had fled Iraq and was tricked into coming home.
    Now they hide the fact that the President they love is a likely one termer. Nobody dies ‘Now’.

    East Bay Jay (2fd7f7)

  47. Silly rabbit you’re asking the wrong question,

    narciso (8d0f34)

  48. the lies upon lies
    uninterrupted failure
    Big Zer0 is toast

    Colonel Haiku (1878e8)

  49. jimgeraghty “The fact that Obama ate dogs between 1967-1971 is irrelevant to this campaign, and I’ll tell that to every dog owner in America if I must.”

    Colonel Haiku (1878e8)

  50. jimgeraghty “This president should be evaluated on his record in office, not on his actions as a child that horrify and shock every dog owner in America.”

    Colonel Haiku (1878e8)

  51. iowahawkblog It’s an Obama-eat-dog world out there

    Colonel Haiku (1878e8)

  52. Dear People Who Still Get Their News from CNN:

    Please come over to my place for dinner this coming Sunday, so we can discuss why you still watch CNN. I’ll buy pizza and fried chicken for roughly six, I’m guessing. I have like eight chairs at my dining room table and I could pull two or three more from the living room, so if some of you bring along a spouse or a date we should still be OK.

    d. in c. (ac417f)

  53. “your own Book of Rules”
    Long Live Project Veritas!
    Death to Alinsky!… whoops… too late

    Colonel Haiku (1878e8)

  54. it wasn’t that Saddam was editing the news – it was CNN’s fear of getting kicked out driving them to self censor bad Saddam behavior from their reporting.

    Yes, that’s basically all I think of when I hear “CNN”. They catered to a tyrant. They were the opposite of the best ideals of journalism.

    Dustin (330eed)

  55. Muslims don’t eat dogs. Stay sane.

    nk (875f57)

  56. “Muslims don’t eat dogs. Stay sane.”

    nk – Obama’s ghost writer was lying?

    daleyrocks (bf33e9)

  57. figures don’t lie. but liars figure.

    tommy mc donnell (3bd3df)

  58. I have a new thread up on the dog eating.

    Patterico (feda6b)

  59. CBS/Purina poll*:

    Barack 0bama: 0%

    Mitt Romney: 100%

    * 995 licensed dogs sampled

    Colonel Haiku (1878e8)

  60. 4. Comment by Kman — 4/17/2012 @ 7:49 am

    You are assuming that every person who chose Obama or Romney was asked questions 3 or 4. What makes you assume that

    Or you could say, better, that not every person who answered the question, if you had to choose between Barack Obama and Mitt Romney who would you be more likely to vote for, also answered the question if it was more a vote FOR the candidate you picked or AGAINSt the other.

    But you can’t say that. These are not raw numbers.

    43% of 1,015 is 436. Yet 476 respondents said they planned to vote for Romney (or actually when pressed that they at least leaned that way)

    52% of 1,015 is 528. Here we have 484 respondents. There you could say 40 people didn’t answer the second question. But you can’t get an extra 40 for Romney.

    I assume they just didn’t put in their poll anyone who refused to answer the first question, but they clearly had to adjust the answers.

    A vote for Neither (volunteered) had they said, 5% and Other (volunteered) had 1% and No opinion had 1%.

    Sammy Finkelman (d22d64)

  61. CBS/Purina poll*:
    Barack 0bama: 0%
    Mitt Romney: 100%
    * 995 licensed dogs sampled
    Comment by Colonel Haiku — 4/18/2012 @ 8:24 am

    — Don’t say “Obama” and “dogs sampled” in the same post!

    Icy (ad8dd3)

Powered by WordPress.

Page loaded in: 0.4767 secs.