Skip to main content

View Diary: Daily Kos/SEIU State of the Nation poll: Barack Obama takes six-point lead over Mitt Romney (63 comments)

Comment Preferences

  •  You mean (2+ / 0-)
    Recommended by:
    yellowdog, eps62

    How do the likely voters screens work? As I say, it varies from pollster to pollster, but this is how PPP does it for its in-house polling (and how they'll likely do it when they switch for us after the holiday):

    We're going to switch to likely voters for our polls this week. Here's what that entails. Right now when we call people for a poll the introduction is 'this is a short survey about some important Florida issues. We appreciate your participation. If you're not a registered voter, please hang up now.' Now the introduction will be 'This is a short survey about the Presidential election in Florida this fall. If you don't plan to vote in the Presidential election, please hang up now.' That's the entire shift. We will continue to use people who voted in at least 1 of the last 3 general elections (2006, 2008, or 2010) as the basis for who we call, as we have been doing throughout the cycle. As you can see the difference between our likely voter polls and what we've been doing over the course of the year is pretty subtle so if you're expecting huge shifts toward the GOP in our polls once we make this change you will probably be disappointed.

    Political Director, Daily Kos

    by David Nir on Tue Aug 28, 2012 at 09:02:40 AM PDT

    [ Parent ]

    •  Fallacy here. (4+ / 0-)
      Recommended by:
      cocinero, Jerry J, pademocrat, tofumagoo

      If you had to have voted in either 2006, '08, or '10 to pass the screen as a likely voter, this would automatically exclude any new registrant as a likely voter.  This would specifically exclude 18-20 year olds who wanted to vote for the first time.

    •  How do they know they are calling (0+ / 0-)

      someone who voted in the last 3 elections?

      And here's a separate question:  (from the demographics section in the link)

      When have you ever in your life seen a breakdown like this?

      Women                 500     50
      Men                   500     50
      Democrat              350     35
      Republican            350     35
      Independent/Other     300     30

      This is, quite frankly, hard to believe.

      •  Because the voter files from Aristotle (2+ / 0-)
        Recommended by:
        eps62, askew

        Inform them of that.

        And what's the issue with that breakdown? If you're concerned about the roundness of the numbers, that's just an artifact of the weighting process, as explained in our methodology statement:

        Using a statistical formula to weight the results can result in a round number of respondents in each demographic category because the Daily Kos/SEIU State of the Nation Poll generally has a base of 1,000 respondents. However, an examination of the raw data (which is also provided each week) demonstrates that this pattern does not exist at the individual respondent level, but rather is merely a by-product of weighting.

        Political Director, Daily Kos

        by David Nir on Tue Aug 28, 2012 at 09:52:09 AM PDT

        [ Parent ]

        •  Note: PPP does not weight for party identification (0+ / 0-)
          (Note: PPP does not weight for party identification.)
          That is from your methodology statement.

          The raw data says :

                                Raw #    %
          Democrat              377     37.7
          Republican            356     35.6
          Independent/Other     267     26.7

          Yet, the party identification numbers presented are clearly weighted.  Or, only data conforming to the desired equal breakdown was thrown out for other reasons. It is rather hard to believe randomness produced a 350/350/300 party breakdown.

          •  If you prefer to believe (3+ / 0-)
            Recommended by:
            Delilah, askew, Vote4Obamain2012

            That we're lying, rather than that PPP's normal procedures (which do not involve party weighting), are responsible for this result, I'm afraid there's little I can say.

            Political Director, Daily Kos

            by David Nir on Tue Aug 28, 2012 at 10:11:25 AM PDT

            [ Parent ]

            •  well, no.. I think some mistake may have been (0+ / 0-)

              made somewhere along the line..  Or, more likely,I don't understand something about how all this is put together.

              Like why the raw numbers don't jibe with your statement that a high percentage identified themselves as Dems (46 percent vs raw percent of 37.7).  

              But, as my time for having fun with polls numbers is up for today, I'll check back later..


        •  Women vote in a higher proportion than men (0+ / 0-)

          I think women were 52% of the vote in 2008.  This time it may be even higher.  I'm curious if gender gap got bigger.

          The scientific uncertainty doesn't mean that climate change isn't actually happening.

          by Mimikatz on Tue Aug 28, 2012 at 11:26:21 AM PDT

          [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site