Why I Won’t Blog or Link to the SurveyUSA Poll with Obama Leading by 3

[This was supposed to go up yesterday but never made it.  Sorry]

Any reader of this blog knows I’ll blog polls whether they be good or bad for Mitt Romney with the noted exception of PPP who does advocacy and push polling rendering them not a credible organization.  But when a polls internals are so ludicrous, I won’t waste my time or your time.  This rant is actually quicker than blogging the poll (if you can believe it).

First Obama leads by 3 in Ohio.  It’s a close race so that’s more than possible. But look at these internals:

  • Mitt Romney gets 22% of the African-American vote
  • Barack Obama gets 70% of the African-American vote

Ohio is 80%+ White but come on.  With those %s there is a zero percent chance Obama is leading with support levels like that. Additionally the idea that Mitt Romney is getting 22% of the African-American vote is metaphysically impossible.

The poll says 26% of those surveyed already voted which compares to 8% who have actually voted.  Since Democrats tend to vote early and all voters who said they voted make it through the likely voter screen, you are guaranteed to over sample Democrats.

Party ID is D +7, almost equal to the reported  D +8 in 2008 — a turnout credibly shown to actually be D +5. And as any reader know, Obama’s 2012 turnout will be nothing like his 2008 advantage.

This is a waste of time to think this poll is remotely reflective of the electorate in Ohio today.


  1. novahockey
    Posted October 24, 2012 at 11:30 am | Permalink | Reply

    Would be interested in your comments on this, from Greg Sargent at Washington Post:

    “Keep an eye on early voting in Nevada: Nate Cohn explains the importance of it:Nevada is a state where the early voting numbers matter, since Romney’s chances depend, at least in part, on a poor Democratic turnout. But the early vote numbers suggest he’ll need to win plenty of Democrats, since they’re not staying home. More than 20 percent of Nevada’s eventual electorate has probably already cast ballots and the Obama campaign is running up the score in Las Vegas’ Clark County, home to 70 percent of the state’s population. Democrats now lead by 23,000 votes in Clark County, more than one-fifth of their 90,000 vote registration lead.”

    Are they running up the score? Your earlier posts seem to indicate otherwise.

    • Porchlight
      Posted October 24, 2012 at 11:42 am | Permalink | Reply

      “More than 20 percent of Nevada’s eventual electorate has probably already cast ballots”

      Probably? What does that mean? And how does Sargent know the size of the “eventual electorate”?

    • valleyforge
      Posted October 24, 2012 at 11:49 am | Permalink | Reply

      He’s ignoring the improvement over 2008, but he’s not wrong to say that the early voting trendline currently favors Obama. There is no evidence yet that Romney can overcome an early vote deficit with an election day turnout advantage in Nevada, as there is in Iowa.

    • Keith W
      Posted October 24, 2012 at 11:49 am | Permalink | Reply

      Not suprisingly, he fails to compare that 23k vote lead number to 2008….it sounds good until you add context. The context is of course that Obama’s margins are almost 40% of the pace they were at this time in ’08.

    • Posted October 24, 2012 at 11:58 am | Permalink | Reply

      As the other comments say, Sargent is dead wrong. Democrats are enjoying an advantage, only it is dramatically lower than their advantage in 2008 which is what Republicans are gunning for.

      • novahockey
        Posted October 24, 2012 at 12:06 pm | Permalink

        Thx all — they absolutely think this is in the bag for Obama over there.

    • Eric
      Posted October 24, 2012 at 12:09 pm | Permalink | Reply

      Nate Cohn is a Democratic hack. He views everything in rose-colored glasses for Democrats. Just check out his articles, and you’ll see that’s true.

      There’s nothing wrong with being a hack. Just be up-front about it. He’s not objective.

  2. Tom
    Posted October 24, 2012 at 11:32 am | Permalink | Reply

    Well said. The Mitt-mentum cannot be stopped by polling and pollsters trying to set a narrative that obama is winning. When the evidence is examined it is obvious that Romney is winning and will win this election in a mini-landslide.

    • No Tribe
      Posted October 24, 2012 at 11:34 am | Permalink | Reply

      I’ve never seen an election where a incumbent is polling at 47% and people like Charlie Cook are writing that he’s got the advantage over the challenger whom is tied or up. Just bizarre how these folks have lost their mind or historical sense.

      • allthingsgeography1
        Posted October 24, 2012 at 11:44 am | Permalink

        I’m a no-spin guy and smell spin a million miles away. Clearly Obama isn’t doing great, although 47% is still bordering on enough support to survive a close election. Obama still stands a chance, but any idea that he has a solid advantage and no matter what is happening…national polls showing Romney at 2-5pts, tied in multiple swing states or Romney slightly ahead…Obama has some kind of structural firewall in the Electoral College is…well, simply wrong.

      • Ron
        Posted October 24, 2012 at 1:47 pm | Permalink

        The magic number 47 seems to be where Obama’s at nationally. If past history is accurate, then OH will reflect the national consensus. Obama has yet to hit 50%. He got a small bump today at Gallup but that’s small consolation.

        Another gauge, as Michael Barone points out, would be the 2010 election. Presidential elections usually closely repeat midterm elections in terms of percentages ascribed to each party, give or take a pt. or two. In 2010 the score was 52-45, in favor of the Republicans. Gallup and Ras have it close to that.

  3. No Tribe
    Posted October 24, 2012 at 11:33 am | Permalink | Reply

    On Nevada, Sandoval is really sounding weak:

  4. novahockey
    Posted October 24, 2012 at 11:33 am | Permalink | Reply

    Oh, an ancedotally, I haven’t heard a peep from my liberal friend from Akron (who’s part of the Dem machine out there) since that first debate. All summer the text messages where free flowing. Crickets since then. I’m sure he’s busy … but that silence is telling IMHO.

    • Tom
      Posted October 24, 2012 at 11:57 am | Permalink | Reply

      I have some democrat relatives in Ohio as well and I’d concur …. crickets.

    • Posted October 24, 2012 at 11:59 am | Permalink | Reply

      The trolls all disappeared from this site as well….

      • Utah Libertarian
        Posted October 24, 2012 at 2:06 pm | Permalink

        It’s not as fun to troll when you’re losing.

      • Posted October 24, 2012 at 3:24 pm | Permalink

        Actually this has become Baghdad Bob Central, so why bother posting alternative views? Every post is the same, claiming polls that show R losing are bad.

      • Posted October 24, 2012 at 3:44 pm | Permalink

        PETER!!!!! You’re back! I missed you. I hope you missed us too.

      • Posted October 24, 2012 at 4:14 pm | Permalink

        I did, I did. You guys are ..unique!

      • Posted October 24, 2012 at 4:20 pm | Permalink

        Not at all Peter…if you look there are plenty of posts of Obama leading here there or yonder. but when you have all the polls showing tied or Romney up and then all of a sudden 1 outlier showing Obama up 3 that is pure BS. Same reason by he really doesn’t blog the Gallup daily…it has been known to be inaccurate. This race is dead tied. This will come down to Ohio. Out of 50 states the residents of 1 will determine our President.

      • Posted October 24, 2012 at 4:25 pm | Permalink

        Shane, there is not a single recent OH poll with R leading. Time just came out with a O+5 OH poll (yes, I know, bad sample, conspiracy or something like that).

        At some point you need to face reality. Things can change but Obama is ahead in the EC count today.

      • Posted October 24, 2012 at 5:19 pm | Permalink

        And if you read what I am saying here and elsewhere, i AGREE. This is still, mathmatically Obama’s race to lose. The best Romney has in Ohio is a tie ballgame from a couple of polls. If he is going to take the state he has to pull ahead in some credible polls (or the RCP average) or at worst be at a tie and work a major turnout game. Without Ohio, Romney doesn’t win. Period (short of a miracle sweep of Iowa, nevada, NH and WI). It doesn’t matter if he wins Florida by 27 or Virginia by 3 and North Carolina by 14….if he doesn’t win Ohio then he loses the presidency by a close margin. Close only counts in horse shoes and hand grenades.

        thos of us pro-Romney need to be confident and upbeat but also not freaking delusional…no one wins this race with more than 285 electoral votes is my call. This will be razor close

  5. valleyforge
    Posted October 24, 2012 at 11:45 am | Permalink | Reply

    On the other hand, it shows Romney holding his base together better than Obama, and leading Independents by 8.

    But ignore the internals, SurveyUSA’s internals are usually godawful: the key is the incumbent is only at 47%. Romney is trailing only because they didn’t push leaners. They show 14% undecided in the Senate race, for instance (where Mandel leads Indies by 13).

    Just another bad poll for the president that the averages won’t reflect properly.

    Intrade doesn’t seem to be buying the toplines though. Romney’s now at 46% odds in Ohio and 44% nationally.

  6. Eric
    Posted October 24, 2012 at 12:04 pm | Permalink | Reply


    Rigged voting machines defaulting to Obama in North Carolina.

  7. Suzanne Morris
    Posted October 24, 2012 at 12:20 pm | Permalink | Reply

    I just wanted to THANK YOU for the Battleground Watch Blog. I am so happy to found it! It’s a “must check” site for me all day long. WELL DONE SIR. Thank you and keep up the great work!
    Suzanne, MN

    • Posted October 24, 2012 at 12:22 pm | Permalink | Reply

      you’re quite welcome. It’s a labor of love that quite honestly I’m shocked it’s as popular as it is.

    • Porchlight
      Posted October 24, 2012 at 12:43 pm | Permalink | Reply

      Ditto here. This blog is to me in 2012 what Jim Geraghty’s Kerry Spot (now the Campaign Spot) and Jay Cost’s Horserace Blog were in 2004. It was their reporting that convinced me in mid-October that Bush was going to win. I am continually hitting refresh and sending links to friends. Thanks for everything you’re doing.

  8. Jan
    Posted October 24, 2012 at 12:36 pm | Permalink | Reply

    I´m also very happy to have found this site. Being a political junkie from the Netherlands, it is very difficult to get your hands on some decent Dutch news coverage regarding this election. Reports form major networks are even worse than what MSNBC tends to air. It felt like me and a friend of mine were about the only ones here in the Netherlands who mourned over the victory of Obama., Now i´m very glad I can read about assuring numbers and analyses several times a day on this site. Keep up the good work Keith!

  9. Kevin
    Posted October 24, 2012 at 12:46 pm | Permalink | Reply

    Thanks for not blogging about SurveyUSA polls. Also, thank you for not blogging about Behavior Research Center polling. After seeing how they have Obama up by 2 in Arizona, they’re the most biased, and void of reality polling center there is.

    • Tom
      Posted October 24, 2012 at 4:11 pm | Permalink | Reply

      I live in AZ and I can attest to the fact the Behavior Research Center is run by hard core leftists. I encountered several of them when attending Arizona State University many years ago. There polls are a complete joke. I’m not sure they actually make calls, I’m fairly convinced they drink beer and write a fake press release.

  10. Utah Libertarian
    Posted October 24, 2012 at 2:09 pm | Permalink | Reply

    The news from my neighbor state is that the NV Republican Party (leadership/structure) is a mess and has been for a while. The biggest reason Romney may pull the state out in the end is because of his own ground game.

  11. AndyN
    Posted October 24, 2012 at 2:34 pm | Permalink | Reply

    Keith, What the hell is going on with Gallup and it’s Obama job approval poll? Please give your avid readers some color?

    • Posted October 24, 2012 at 2:45 pm | Permalink | Reply

      A couple weeks ago Gallup changed the weighting of its polls to include non-White at just over 30% of its “adults” survey. They then skinny it down to come to a “likely voter” model. This is a ludicrously large amount of non-whites surveyed.

      From a voting standing point, non-Whites made up 19.3% of the electorate in 2004 and 23.7% in 2008. The big jump in 2008 is based off a) enthusiasm among non-Whites for Obama and b) White voters staying home: http://battlegroundwatch.com/2012/09/26/the-reality-of-2012-voter-turnout-the-white-voter/

      Gallup decided late in this cycle to give into Democrat demands for a far higher non-White weighting than can reasonably be assumed to appease Team Obama. This is why Obama’s job approval spiked about two weeks ago. The reality is his approval is a few points lower.

      Team Obama uses these over generous racial demographics to run their polls showing Obama winning:

      But the reality is the #s they are using are all wrong. They are using notoriously unreliable exit polling data that says White voters were only 74% of the electorate in 2008 when the Census bureau said it was really 76.3%. Some would argue there is no way Team Obama would knowingly use wrong #s, but they are actively trying to make Obama seem like the inevitable winner so the White voters don’t show up again. It happened last time, so it is very possible.

      Gallup is simply carrying water for the President with the racial makeup of its polls which despite their sop to Axelrod & Co still show Romney leading.

      • AndyN
        Posted October 24, 2012 at 2:49 pm | Permalink

        Keith, thank you, crystal clear now.

    • zang
      Posted October 24, 2012 at 2:46 pm | Permalink | Reply

      For job approval, Gallup polls adults. Not likely voters, not even registered voters. Completely worthless this close to an election.

      People should have been wary of jumping on the Gallup bandwagon. Gallup and Pew polls are among the most volatile of any public pollster.

  12. JGS
    Posted October 24, 2012 at 4:04 pm | Permalink | Reply

    Isn’t the geographical weighting of this SUSA poll way off as well, under-inclusive on conservative Cincinnati relative to Columbus and Cleveland?

  13. JGS
    Posted October 24, 2012 at 5:00 pm | Permalink | Reply

    New Time poll shows Obama +5 in Ohio, but if I understand this correctly, the sample includes 9% more Democrats than Republicans. (See p. 3.)



    Only 206 out of 742 respondents included in the survey were Republicans, which is 27.76%, whereas 273 of the 742 respondents were Democrats, which is 36.79%, a 9.03% difference.

    • spottedreptile
      Posted October 24, 2012 at 8:58 pm | Permalink | Reply

      The key here is ‘included in the survey.’ If you say you’re Republican, some pollsters just hang up on you there and then.

  14. JGS
    Posted October 24, 2012 at 5:03 pm | Permalink | Reply

    Among likely voters, however, the difference may be “only” 5.35% (see p. 5). I’m not an expert at reading these things so would appreciate commentary from those more familiar with reading these polls.

  15. JGS
    Posted October 24, 2012 at 5:06 pm | Permalink | Reply

    I actually think my first post was correct, my second post only looked at those likely voters who had not yet voted, but if you include both likely voters who have not yet voted and those who report they have already voted, the difference in party ID is indeed 9.03%.

  16. spottedreptile
    Posted October 24, 2012 at 5:36 pm | Permalink | Reply

    Keith, whatever happened to Unskewed Polls? A few weeks ago they were on everybody’s lips, now you never see a mention of them. Did they go over to the dark side or something?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: