Monday, October 8, 2012

October 8, 2012 Analysis - Update


Polls included:
Even - Rasmussen Daily Track
O+3 - CNN/ORG
Even - National Journal
O+3 - NBC/WSJ
O+1 - Battleground
R+4 - Pew

O+0.50% - Current RCP Average
O+1.04% - Average using the 2008 turnout model
R+3.76% - Average using the 2010 turnout model
R+3.81% - Average using the 2004 turnout model
R+5.45% - Average using the Rasmussen Party ID turnout model

I wanted to do an update to include the new Pew poll out today.  Because of the partisan split of the Pew poll (the first poll I've seen with an R+anything sample), this poll moves all of the polling averages marginally, EXCEPT for the published RCP average.  In my opinion, this confirms the accuracy of the reweighting models.  If a poll with a slant toward Republicans doesn't move the normalized results, then we are seeing a great deal of accuracy out of the results.

3 comments:

  1. I wonder what you think of the Sept 22 post here:

    http://publicpundit.blogspot.ca

    Is it a valid analysis in your opinion? Just wondering.

    ReplyDelete
    Replies
    1. It tracks with some other analysis out there. I think Techguyblog came up with something similar. The GOP tends to out perform the Rasmussen Party ID.

      I'm hesitant to go there though. People who get paid for this stuff continue to claim D+3 is likely. Also, Romney is winning this with an Even turnout, so no reason to get too wild eyed.

      I still think in the end the Dems will get their turnout machine in gear and will match the GOP turnout. But only match it, not exceed it.

      Delete
    2. Fair enough. My gut feel is between a D=R to R+1 turnout. Dems are better at ground game but they are hindered by apathy in their major voting blocs. There's huge enthusiasm on the R side and the indies are fleeing our way. Tipping point was the first debate. I don't think we'll see a swing back but the game's not over by a longshot.

      Delete