The Fix, one of the Washington Post’s political blogs, has an excellent post today looking at new polling results to see if the divisions from the primary have carried over much to the general election. It seems like they haven’t too much.
Inside the Post Poll: What Obama-Clinton Divide? – The Fix:
Among women, the strongest pillar of Clinton’s support in the primary, Obama holds a wide 54 percent to 39 percent lead over McCain. And, even among white women, who were one of Obama’s weakest constituencies in the primary season, he fights McCain to a statistical draw — 47 percent to 46 percent. Compare that to the 2004 presidential race in which Sen. John Kerry (Mass.) lost white women by 11 points to President George W. Bush and won women overall by just three points.
As for the white, rural, blue-collar voters of whom much was made following large Clinton victories in places like Ohio, Kentucky, Pennsylvania and West Virginia, there also seems to be no entrenched resistance to Obama’s candidacy.
While Obama is losing among white voters (50 percent McCain, 42 percent Obama), there is no gap between how he performs among whites with college degrees and those without a college education; McCain lead 51 percent to 44 percent among white college graduates and 50 percent to 41 percent among white non-college graduates.
I did not realize that Kerry lost white women by quite so much. Also, it’s very interesting to see that although Obama did far better among highly-educated whites in the primaries, the same does not hold true in the general. It’s almost as if primaries are an entirely different contest with a different set of choices.
…
Oh wait, they are. I wonder if anyone will remember this four years from now.
The Fix uses this data, reasonably enough, to point out the seeming lack of a divide leftover from the Obama-Clinton fight. I would also argue that it speaks more broadly to the problem of trying to infer the preferences of voters in a general election from the smaller set in a primary, or more generally, the problem of trying to infer the preferences of people choosing between A and C by looking at how they chose between A and B.
Here are some earlier posts where I talked about this problem.
Electability by the Numbers
Wherein I Express Displeasure With Lou Dobbs
A (Budding) Sociologist's Commonplace Book 
Carl
/ July 18, 2008I think at least some of this just boils down to the security issue. As Lakoff points out, conservatives are just plain stronger on the patriarchal stuff. If that’s your issue, you’re not likely a Democrat. So it doesn’t really kick in until the main showdown.
I had a young woman remark in class last semester that she could never vote for a woman because she wouldn’t feel safe. About half the class nodded as if this was the most obvious thing in the world. Once you get to the general election, the same discourse applies to the Democrats, who are in this respect ‘feminized’. Bush took full advantage of that, of course, and I still have no idea why Kerry played right into it by trying to stand on his war record. Dude, if that’s the issue you lose. Try something else.