I have had a lot of focus over the course of my blogging career, on an American group that is largely cultural, perhaps even ideological: the Arts & Humanities tribe. Let me give a quick summary of that before I try and shoot this film from a different angle. The Arts & Humanities tribe is largely white, liberal, educated word-people. They might be miseducated, but not uneducated. They are something close to the first of Snow's Two Cultures (now Three, with the social sciences). They heart NPR. They weren't all literature or anthropology majors, are not exclusively white, nor are they always liberals. The writers at Maggie's Farm or National Review, for example are certainly A&H, but not liberal. Yet on the whole, they tend that way. Boundaries are fluid depending on how intense an identification, or how much actual scholarship rather than posing one requires for membership, so estimates are suspect. But let us say they constitute 7-17% of the populace. They are disproportionately represented in running things, especially government. Lawyers (some), librarians, writers, psychologists, academics of many stripes. Very Canadian border, as they are concentrated across the top of the country: New England, Upper Midwest, Pacific Northwest.
My people. That's who I come from. I'm rather suspicious of them and in rebellion against them in these later decades, but culturally, these are who I am most comfortable talking to.
When we look at all those "Romney Won White Voters" graphs and articles, keep in mind that those numbers were in spite of this sizable group of A&H tribe. Subtract them out, and ponder.
hbd* chick linked to a set of graphs by The Audacious Epigone which shows the electoral vote breakdown if we counted variously by income, sex, or race. For example, if only white men vote, Romney wins 490-41 (7 undecided).
Coupling that with similar data that Steve Sailer extracted from the Reuters-IPSOS poll, where we learn that married Jewish women voted for Obama 66-34, we have an entirely different way of looking at the culture, at least in terms of voting.
Let us assume, arguendo, that this second method is much better than mine; that a far more simple tribalism of race, religion, gender, etc, is true. Who we will tend to vote for, and what appeals by candidates we will respond to, are largely a result of accidents of birth, with all the worrisome implications that we can be manipulated with increasing accuracy by campaign teams with each passing year.
If that is true, then the demographics are going to go so consistently against the Republicans that they will not ever win more than an occasional narrow majority in House, Senate, or Electoral College. They will never be able to bend their message enough to peel off enough stragglers from all the other groups.
The Democrats are not a unified party, but a coalition. Have been for a few decades now. Black voters are not the natural cultural allies of union members or Jews, but they can agree to defeat Republicans and then each get something of what they like in the next year's legislation.
Therefore, what the Republicans need not a broadened message, but a coalition partner. A group that can sign on despite differences in exchange for cabinet positions, or the introduction of particularly dear legislation. There is something of this already. Social conservatives, economic conservatives, and libertarians overlap, but are not the same thing. They are a proto-coalition.
What coalition partner can they hope to attract with consistency, whether ethnic, ideological, or cultural?