Nov 05 2012

Polling In An Alternate Universe Part Deux: The Acid Trip

Published by at 8:32 am under 2012 Elections

I plan to come out with my  predictions later today. Needless to say I have been waiting because I want to see Gallup’s last hurrah today before jumping in. But to be honest, after seeing the CNN poll today I am about ready to make Michael Barone look cautious:

The poll, released earlier tonight, shows a 49-49 tie among likely voters. But to get that result CNN had to use one of the most skewed samples we’ve seen this campaign (see page 29):

 Among those likely voters, 41% described themselves as Democrats, 29% described themselves as Independents, and 30% described themselves as Republicans.

A D+11 sample! By comparison, the electorate in 2008, when Obama-mania was at its peak, was merely D+7, according to exit polls.

You know why you have to put out such a laughable poll to show anything other than a Romney blow out? Reverse engineering this puppy is easy.

If Obama is losing independents well into the double digits, and/or he is is losing left-of-center Dems in such numbers he cannot hold more than 80% of his party, then you have to ramp the Dem turn out numbers up to “absurd” to compensate. Another data item noted:

CNN conducts a poll where Romney leads independents by 22 points and it’s TIED.

Folks, if Obama is losing indies by 20%, he is losing a lot of dems in crossover as well. More here, and more later.

Update: Ed Morrissey chimes in:

So we are expected to believe that since 2008, (a) Obama has lost thirty points in the gap with independents, (b) Obama has lost fifteen points in the gender gap, and (c) Obama is still just four points below his 2008 share of the electorate?  Only in a world where 41% of the voters will be Democrats and only 30% Republicans, and that world won’t be what we see tomorrow.

Watch out – You are entering the Twilight Zone!!

5 responses so far

5 Responses to “Polling In An Alternate Universe Part Deux: The Acid Trip”

  1. oneal lane says:

    At this point there is so much poll data, one can easily cherry pick, throw in some wishful thinking, and a helping internet echo and get any result you want. Not unlike the 2000 and 2004 elections the race has tightened toward the end. Obamas “Sandy” exposure helped his status. It gong to be close. But I hope I’m wrong.

  2. DJStrata says:

    Its ok. The Redskins lost yesterday. Time for the incumbent to lose! 😉

  3. oneal lane says:


    Now thats the kind of solid evidence I can hang my hopes on! 🙂


  4. clintsf says:

    I’m fairly confident this year — but my big caveat is that pollsters are only getting 9% response rates.

    Yes, the CNN poll is utter crap. But it’s not just the top line that’s crap — the whole thing is. They selected 10,000 phone numbers at random — and they’re reporting on the 900 people who were willing to talk to them. And there’s absolutely no reason to believe that the 900 who answered the poll questions are in any way representative of the 9,100 who didn’t. There’s nothing remotely scientific about that.

    (And don’t get me started on the problem of cell phones — does anyone have an area code that actually reflects where they live anymore?)

    There are all kinds of reasons for confidence — newspaper endorsements, GOP enthusiasm, Obama fatigue, overflowing stadiums and half empty middle school gymnasiums, gas prices, the terrible economy, the first debate, GOP senators campaigning with Romney and Dem senators running away from Obama, unemployment, Catholics, evangelicals, coal, and on and on.

    But I’d be so very much happier if we actually had some data…

  5. Frogg1 says:

    I think Dem strategist, Joe Trippi, said the truth in an analysis he gave on FoxNews last night. He said there are definately two very different poll models being used. There is the “Obama model” and the “Romney model” as far as demographics, turnout, early voting models, etc. Trippi said that he thinks one of those models will be correct….and, he also said he thinks it will be a landslide in one direction or the other depending on which model is correct. We will know soon enough, won’t we?