Sunday, June 21, 2009

Divergent Polls in Ohio

Four pollsters have released approval numbers on Ted Strickland since last November.

Two of them- Quinnipiac and the Ohio Poll- show Strickland in a strong position. The most recent Quinnipiac showed Strickland at 57/29 for a net of +28. The latest Ohio Poll showed it at 56/34, or +22.

SurveyUSA and PPP show a very different picture. SUSA's most recent numbers were 45/44, or +1 and our release tonight finds 43/42, also +1.

It's worth noting that Quinnipiac and PPP are showing the same trend. In January Quinnipiac had Strickland at +38, so his +28 now is a ten point drop. We had him at +13 in January and now at +1 for a similar 12 point drop. So while we show very different pictures on Strickland's overall popularity we do both find it declining.

Why the discrepancy? I'm not sure. Quinnipiac and the Ohio Poll use live interviewers while PPP and SUSA utilize automated polling, but that may or may not explain the difference.

Who's right? All four pollsters have done a good job in recent Ohio elections. Last fall Barack Obama won the state by four points. PPP, SurveyUSA, and the Ohio Poll all came within two points of that outcome, and Quinnipiac came within three.

Hillary Clinton won the state's primary by ten points. SurveyUSA picked that outcome right on the nose, PPP and the Ohio Poll each came within a point, and Quinnipiac came within five points.

It's odd for a quartet of outfits that have all shown a good track record of predicting Ohio election outcomes to show such divergent numbers, but that's probably at least somewhat attributable to the amount of time before the election, and it will doubtless become clearer as we move toward 2010 just how vulnerable Strickland is or is not.

4 comments:

researchrants said...

I haven't looked at any of these, so this is just a hunch: I'll bet it's much easier for wording/questionnaire differences to move an approval number around than it is for them to move an at-the-wire horse race number around.

Meaning, regardless of the exact screen and weightings they each used in the primary and general elections, they were all asking, at heart, a single question that most respondents had an answer for: who are you voting for? Here, with approvals, everyone is probably asking about a whole host of people and issues, of varying familiarity and interest to the respondents, and perhaps leading to some respondent drop-offs along the way.

ytownMetz said...

I believe you guys are dead on right with Ohio. Kasich will end up winning.

Here is why: http://mvred.com/2009/06/19/the-gop-comeback-in-ohio/

But can you please narrow the Party ID labels down to the 2008 Ohio Exit Polls of:
Dem: 39%
GOP: 31%
Ind: 30%

Not 50-35-15?!

Thanks!

Rasmus said...

This is snark, right? You think Kasich will end up winning because he won an online poll at a high school sports site? Srsly?

ytownMetz said...

Rasmus, he will win because Strickland has NO accomplishments as Governor. He is cutting library funding by 50% and is very close to raising taxes.

He stands no chance in 2010. This is coming from a Strickland supporter in 2006!

 
Web Statistics