ABC/Washington Post Poll - October 15, 2012
Likely Voter, 923 sample size, Obama Leads 49-46, 2% undecided.ABC/WaPo continues it tradition of releasing polls that severely undersample Republicans and oversample Independents. The D/R/I for this poll is 35/26/39, which is a terrible sample. D+9 and over counting Independents by about 10 points. Note that they report the sample as 35/26/33, with no explanation of what the other 6% are, if they aren't Independents. Just a weird sampling methodology. It's making me start to respect Zogby. Independents favor Romney by 6 points. Running their numbers through the reweight model we get:
O+3 - Current result
O+1.8 - 2008 turnout
R+1.5 - D+3 turnout
R+4.0 - 2010 turnout
R+4.1 - 2004 turnout
R+5.9 - Rasmussen Party ID
This is a good poll for Romney, and shows him holding his lead comfortably.
Every year I hope it's the case, and every year I'm disappointed.
ReplyDeleteBut maybe *this year* will be the year that people start realizing that political polls are crap for 11/12ths of the year.
I agree to a certain extent, but I think it would be a mistake for conservatives to just completely ignore polling because we distrust the MSM. We can sometimes get so much in a bubble that we end up being our own worst enemy. Nominating candidates like Sharon Angle or Christine O'Donnell is what happens when you think everyone is on board with a pure Tea Party message and don't care about the actual candidates.
DeleteI think one of two things are going to happen this election, either all of the polls are magically going to show some huge movement towards Romney in the final week in order to save some reputations (you could make that case right now) or the pollsters just keep going along with D+9 samples and watch their reputations get shredded to tatters as the big story is "how all the pollsters got it wrong".
Remember though, Gallup's "final" poll before the 2008 election showed Obama beating McCain by 11 points, yet the MSM has never really "called them" on having such a terrible prediction.
http://www.gallup.com/poll/111703/final-presidential-estimate-obama-55-mccain-44.aspx
I don't think we will ever get away from this, Allen. This year has been one of the most egregious, but that is because the Obama campaign and the media pollsters (BIRM) figure out that they can game the RCP average. We were warned back in December that OFA was going to game the RCP average as a ploy.
DeleteWhat gives me comfort is that a lot of folks now are legitimately questioning the methodologies and sample results. We now have records of what they have done, and will be calling them on it in the future. I've noted that no one bothered to do this in the past, finding analysis of historical poll data has been difficult to find.