The current Republican Party is based upon a tribal appeal to the "white vote" as the only true Americans as they seek to have the white voter identify with them. When we Democrats talk about the national "white voter" as a single identity group, we are cooperating with a Republican meme. Starting with Republican apologists like Micheal Barone, Republicans have succeeded in planting this mode of analysis into the media. Additionally, a great deal of the certitude and morale of Republicans comes from being the largest part of the largest group in America.
It is time to stop cooperating with this Republican meme, which encourages Republicans and puts Democrats at a disadvantage in the Midwest.
The recent Pew Poll is a robust poll of 3,151 voters. This allows meaningful crosstabs on a variety of issues, which Pew thankfully makes generally available.
Earlier this year, Larry Bartels addressed the continuing fascination with the "decline" of the Democratic Party among the white working class. In “The Party of the American Working Man and Woman” he wrote:
(3) Overall, there has been no discernible trend in presidential voting behavior among the “working white working class.”With that in mind, let's look at the Pew data on the "white vote." Among likely white voters (2107), Obama does lose 39-54%. But what happens if we look at it regionally? The first thing that leaps out, is that white southerners are 35% of the white vote. And among southern whites, Obama does get creamed: only 27% support him. As a white southerner myself, I can confirm across broad regions of the South, support for Democrats has collapsed. That 27% is comparable to the support for Obama in Utah, or the support for Romney in the Hispanic community. Sure there are islands of support for Dems among whites, such as NOVA, various communities with universities, such as Austin or the triangle area of NC, just as Republicans have areas of Hispanic support, such as among Cubans or in Texas. But broadly just as there has been a group identification among Hispanics with Democrats, there has been a similar ID with group pressure among white southerners (which political scientists might refer to as "tribal").
(4) While southern working whites without college degrees have become more Republican in their presidential voting behavior (by 4.5% per decade), non-southern working whites without college degrees have become more Democratic (by 1.6% per decade).
So what happens if you take southerners out of the mix. Obama's support among likely voters varies from 42-51 % among the other 3 regions and Romney's varies from 41-51 with Obama having 22 fewer net votes out of 1370, a difference of 1.6%. Since Obama has a higher lead among RV's it's likely he is tied among all non-southern white voters.
Looked at this way, Democrats have their lead among several minorities, Hispanics, blacks, and Asians, and Republicans have their lead among one minority - white southerners - amounting to 27% of the LV population.
A final note: there are a lot of ways to slice and dice the "white vote" from the standpoint of voter analysis. Others likely have more explanatory power of how Democrats win and lose elections. See Tom Edsall's article in the NYTimes: White Working Chaos for a start to exploring the subject. But to counter the Republican meme of a threatened group of whites who are defending "real Americans" against the others, the southern/non-southern dichotomy is the broadest and most inclusive way of changing the frame.
PS for gauging Democratic prospects in states like VA and FL, the national white vote is probably a useful metric, since the white vote in those states probably reflect a blend of the southern outlook and the rest of the country.