President Obama is at 49 percent and Mitt Romney at 46 percent among likely voters in a Washington Post/ABC poll out today.
The partisan breakdown is 35 percent Democrat, 26 percent Republican, and 33 percent independent. That’s nearly a double-digit advantage for Democrats . . . which is baffling. (In 2008, the Democratic advantage was points.) I’m sympathetic to the arguments that a pollster shouldn’t try to control the partisan ID in his sample, just like he shouldn’t try to control what percentage of support each candidate gets. (In other words, partisan ID is fluid and something the pollster is trying to measure, not determine.)
But at the same time, if you’re showing that high a discrepancy, which has no precedent in modern history (Obama’s plus-7 advantage was unheard of; prior to that, Democrats had not had in recent times more than a plus-4 advantage), is it really possible that everything is right? Is the pollster interviewing the right mix of cell phone and landlines, of different income levels, of different races, of different educational backgrounds? With GOP enthusiasm higher than Democrat, it’s perplexing how a sample could find this many more Democrats than Republicans — yet have no underlying problem.
In the Politico/George Washington University poll out today, Obama is at 46 percent and Romney at 45 percent among likely voters. The partisan breakdown is 43 percent Democrat, 40 percent Republican, and 15 percent independent. In battleground states (Colorado, Florida, Iowa, Michigan, Nevada, New Hampshire, North Carolina, Ohio, Virginia, and Wisconsin), Romney has a two-point lead, 50 percent to Obama’s 48 percent.