Here’s a link to a detail piece on the current ABC/Washington Post poll from yesterday.
The question is “Q: (AMONG LIKELY VOTERS) If the presidential election were being held today and the candidates were (Obama and Biden, the Democrats) and (Romney and Ryan, the Republicans), for whom would you vote?”
The poll finds Romney with a 49% to 48% lead in a survey of 1,382 likely voters. But doesn’t something look wrong?
If both Obama and Romney are attracting the same split of their parties voters, what’s the deal with Romney’s 9% lead among Independents? Wouldn’t that account for more that an 1% lead?
It would if the results weren’t skewed toward Independents, and away from Republicans.
So this poll is assuming Democrats hold a +5 advantage over Republicans. The problem is that while the 34% number for Democrats sounds about right, the number for Republican’s is laughable. Look at Rasmussen’s most recent party affiliation survey:
To get to their 49% Romney, 48% Obama total they have Independents 6% higher (assuming they lumped Other into that total) than Rasmussen and Republicans over 7% lower.
If you “deskew” the poll with applying Rasmussen’s party affiliation splits to the raw data (which you have to back out from the total LV’s) you come up with a Romney 51%, Obama 46% split overall. That’s a pretty serious difference from what is presented.
It also makes a lot more sense given the first graphic…
Update: The new daily ABC/Washington Post poll is Romney 50%, Obama 47%.