The Romney campaign is counting on independent voters to carry their man across the finish line. "The race is going to come down to independents," Romney pollster Neil Newhouse told reporters on a conference call Wednesday. "It's simply cold, hard math. In 23 of the 25 national polls released in the last month … Mitt has led among independents … by 7 points." And he's right—Romney has had a consistent lead among self described independents, but he's wrong when he suggests that independents will swing the election.
The problem with using support from "independents" as a metric is that it assumes independents are (a) a static group of people who are (b) not functional partisans. Most political science majors (and, I would think, all political science professors) can tell you that (b) is false—very few self-described independents are actually politically independent. Most are partisans—they consistently vote Republican or consistently vote Democratic—who, for whatever reason at that moment see themselves or want to be seen as not being tied to a political party. Independents, in other words, are distinct from swing voters.
Then there's the question of the composition of "independents," which varies. To understand what I mean, check out this chart from Talking Points Memo's PollTracker.
Note the sharp decline in GOP identification starting in the fall of 2010 and the simultaneous rise of independents. Republican identification goes from a high of 31.7 percent in late August, 2010 to a low of 22.5 in early April, 2011, a fall of a bit more than 9 percentage points. Where did these Republicans go? Well it's probably safe to say that they didn't become Democrats, so you would expect to see a corresponding rise in "independents." And sure enough, independents go from 26.3 percent identification in late August, 2010 to 36 percent in April 2011. Or to put it more simply, the GOP lost 9 percentage points at a time when "independents" gained 10. Since then the two groups have remained relatively stable (this being a period of time, as Steven Taylor points out at Outside the Beltway, where the GOP remains an extraordinarily unpopular brand) with independents dropping a couple of percentage points while the GOP picks them up. Is it any surprise, then, that Romney has an advantage with "independents," when that group has grown disproportionately more right-leaning?
Incidentally, as TPM's Josh Marshall points out, this also punctures one of the conservative critiques of many of the polls they don't like—that Democrats are oversampled in the polls showing Obama ahead. But if Democratic identification is +7 nationwide, it's not out of line that people should respond that way in individual surveys.