Tuesday, October 2, 2012

Quinnipaic Poll - October 2, 2012

Likely Voter, 1912 sample size, Obama leads 49-45, 4% undecided.

Quinnipiac released a poll today that has a very large sample size.  The Partisan ID of this poll is pretty odd with 500 Republicans, 593 Democrats, 706 Independents, and 133 "Other".  What exactly "other" is difficult to figure out.  It is also hard to figure out if their 2% Ind preference for Romney includes the "other" or not.  I am lumping these 113 in with the 706 for a D/R/I of 31/26/43.  This is a really odd sample, with Independents considerably over sampled compared to both parties.

The turnout models give the following results:

O+4 - Current result
O+5.1 - 2008 turnout
R+0.9 - 2010 turnout
R+1.0 - 2004 turnout
R+2.6 - Rasmussen Party ID

Because of the huge Ind oversample, adjusting to historic turnout models forces a wider distribution in both directions.


  1. Really like this blog, it keeps me sane :) thanks again for the great analysis.

    My own "internal" battle has been that the argument that most pollsters are wildly oversampling Democrats is a VERY valid argument of why the polls look the way they do. My pessimistic side, however, keeps telling me that if a side has to make excuses all the time for why they're losing in the polls, they are probably going to lose.

    Curious what Dave in Fla's guesstimate about what the 2012 partisan makeup will be?

  2. I think 2012 will be even, just like 2010. Everyone talks about 2010 being a big year for Republicans, but it wasn't. Democrats matched the Republican turnout. Both parties have shown they can turn out their base.

    In 2008, Democrats had a 2% increase in their turnout, while the Republicans had a 5% decrease. Both of these were erased in 2010. Given the loss of Democrats being reported and the enthusiasm numbers that are showing up in some of the polls, I think we are looking at a D/R/I of 35/35/30.

    1. I think that's a fair assessment, and along the lines of what I was thinking.

      Michael Barone has an excellent article about how historically, Presidential elections tend to mirror the preceding off-year election in terms of partisan make up.

      I do think, however, that the polling data DOES unfortunately depress turnout on our side. So many people I know think "it's over" and I know that has the effect of drying up donations, volunteers, voter outreach, etc.

      If the partisan makeup is like 2010 or 2004 though, I think Romney will win, maybe even comfortably. Most of the polls I've seen have Romney winning independent voters handily. Also, these things usually swing pretty hard against the incumbent in the closing weeks if the economy is in rough shape. I just don't see America surging towards Obama in final days, not many people I know are eager for 4 more years of this.

  3. Zombie over a PJMedia had a post a few days ago questioning the validity of the "depress the vote" strategy. Bottom line, there is no evidence to prove that it works.