Dueling polls: Obama lead narrowing or expanding in Ohio?
posted at 9:21 am on June 27, 2012 by Ed Morrissey
In one corner — Quinnipiac, by way of National Journal:
Brown leads Mandel 50 percent to 34 percent in the poll. In the previous survey, released on May 10, Brown led 46 percent to 40 percent over Mandel, who is also an Iraq War veteran.
Mandel has attracted a lot of negative headlines in the local press but he’s been the beneficiary of over $8 million in outside spending aimed at attacking Brown, which has kept him in the race. The Democratic Senatorial Campaign Committee is keeping a close eye on the race, and has already reserved fall ad time in the state.
President Obama leads Mitt Romney 47 percent to 38 percent in the Ohio survey.
In the other corner — PPP, by way of Politico:
Obama leads Mitt Romney 47 percent to 44 percent in the Buckeye State, according to the latest poll by the Democratic-leaning firm PPP. But that’s down from a seven point lead that he maintained in the last two PPP surveys of the state.
Obama is clinging to his lead based on his strength with three demographic groups: black voters, young voters and women. Obama leads among African-Americans 93 percent to 6 percent, and he beats Romney among young voters 54 percent to 36 percent. Women support Obama over Romney 52 percent to 41 percent.
So which one’s right? National Journal takes a peek under the hood of the Q-poll:
Quinnipiac’s sample is slightly more Democratic-leaning than the previous poll. In the latest poll, 34 percent of voters said they generally consider themselves Democrats, compared to 26 percent who say they are Republicans. In the early May survey, 31 percent of voters were Democrats, and 29 percent were Republicans.
In the 2010 midterms, exit polls showed the D/R/I in Ohio as slightly Republican at 36/37/28. The sample in the Q-poll now looks more like the 2008 model, where Democrats had an eight-point advantage, 39/31/30. I doubt that we’ll see either model exactly as it was, but I’m inclined to think that Republicans are going to turn out more enthusiastically to oppose Obama than Democrats in Ohio will to support Obama, especially if we see a couple of more jobs reports like we did for May.
PPP’s sample looks a lot more reasonable, at 39/37/24. It might be oversampling both Democrats and Republicans, but it gets the relationship between the two correct. Undersampling independents helps Obama a bit, though, as his job approval rating among indies is a disastrous 37/53, although Obama and Romney tie at 42/42 among them. However, that looks like a high-water mark for Obama as long as his job approval rating is that poor, while Romney would have more upside. Let’s not forget that an incumbent should be scoring at 50% or higher at this stage of a race, and the close race with Obama at 47% in both polls is an indicator that he’s going to have a lot of trouble making the sale.
In this case, I’d take PPP over the Quinnipiac poll, although I generally consider the Q-poll more reliable overall.