Monday, October 8, 2012

Battleground Poll - October 8, 2012

Update: The article about the poll included information that was not included in the internals.  Namely that Independents prefer Romney by 16 points.  I was using the same 4 point that had been in the last version of the poll.  I also found a small math error in my model that incorrectly shifts most of the polls a couple 100ths of a percent toward Obama.  That error is corrected and will not occur in future posts.  I'm not going to bother fixing the ones below.

Likely Voter, 1000 sample size, Obama leads 49-48, 2% undecided.

This poll is showing gradual movement toward Romney.  It was Obama +3, last time it was Obama +2, now it is Obama +1.  In all cases, Obama's level of support has not changed.  All movement is undecideds deciding to vote for Romney.  The partisan ID didn't change much, going from D/R/I of 43/40/17 to 44/40/16.  They sampled an additional percentage of Democrats.  Independents now favor Romney by 16% in this poll.  The turnout models give the following results:

O+1 - Current result
O+1.1 - 2008 turnout
R+3.0 - 2010 turnout
R+3.0 - 2004 turnout
R+4.5 - Rasmussen Party ID

This poll shows significant support for Romney in all turnout models, and makes reelection very dicey for Obama even if turn out matched 2008.


  1. Hi Dave, a few questions...

    How many days of polling results are used in the rolling averages? Are they the same across all organizations?

    I lurk over at Ace of Spades quite a bit and was wondering if you had done another of those excellent winning the Senate analysis recently.

    Tim in Chandler, AZ

    1. Rasmussen uses a 3 day average in his national poll and a 7 day average in his battleground poll. Gallup uses a 7 day average in their national poll.

      You are right, I need to update the Senate percentages and post something about it.