Thursday, October 18, 2012

October 18, 2012 Analysis

Polls included:
R+1 - Rasmussen Daily Track
R+1 - Fox News
O+2 - IBD/Tipp
O+1 -Battleground
O+3 - ABC/WaPo
R+1 - Survey USA

O+0.42% - Current RCP Average
O+0.81% - Average using the 2008 turnout model
R+1.94% - Average using the D+3 turnout model
R+4.14% - Average using the 2010 turnout model
R+4.20% - Average using the 2004 turnout model
R+5.88% - Average using the Rasmussen Party ID turnout model

The big news yesterday was the Gallup poll showing Romney with a 6 point lead, hitting 51%.  Of course, I don't include that poll in my average, because I can't see their internals.  However, if I was able to, all of these models would reflect higher levels of support to Romney.

Both Rasmussen and IBD/Tipp moved one point in Obama's direction.  This is not surprising.  Rasmussen does not include any polling from after the debate yet, but IBD/Tipp does.  They will be capturing a larger number of Democrats in their likely voter screen due to increased enthusiasm.  Keep in mind though, IBD/Tipp is a D+7 sample size, yet is only showing an Obama lead of 2.


  1. Dave,

    Do you feel that the polling people adjust their sample distribution right at the end to make sure they can brag about accuracy the next time? I was looking at alot of the states/national results from the past few cycles and RCP is really close on most of them. I'm not sure why they would roll along 2-5 pts incorrect when everyday people like us can see the party breakdown isn't historically accurate.

    I guess I just keep drifting between frustration with Ohio's stubbornness and the analysis like yours that shows a comfortable Romney lead...

    Are you submitting your analysis as evidence that Romney is doing that well, or simply as a factual balance?

  2. I'm doing this because I became frustrated by getting obvious spin. The polling results were clearly off somewhere, since both R+2.6 and D+19 couldn't be both true.

    I just wanted a clear picture, even if it wasn't a good picture.

    As for why the samples are so off, I can't tell you. Pew is saying they only get a 9% response rate now. My theory is that Democrats are much more motivated to answer polls during this cycle, and it is resulting in over weights.

    However, some polling firms are obviously adjusting their polls to get a specific turnout. IBD/Tipp always has a D/R/I of 39/32/29 and always has a 1000 sample size. That requires manipulation to achieve.