Crossposted on One Million Strong
Jerome Armstrong cited an interesting but ultimately unconvincing analysis from National Journal's Michael Brownstein about a potential Obama-McCain general election:
Obama's strong support from affluent and college-educated voters in the primaries demonstrates his opportunity to convert Republican-leaning upper-income voters (especially men) now disaffected from President Bush.
But Obama's struggle during the primaries with working-class white women suggests an opening for McCain to court those downscale "waitress moms" with the same security issues that drew many of them to Bush in 2004. The first trend should boost Obama in Virginia and Colorado (two affluent states atop Democratic target lists); the second should help McCain defend Ohio and besiege Pennsylvania.
Brownstein is clearly conflating primary results with general election competitiveness. Why does this matter? Brownstein slips. Obama's struggle has not been with working-class white women voters, it's been with working-class white women Democrats, indeed many of them have been called core Democratic voters, who are likely to go for the Democratic nominee no matter who it is. And what does Obama's appeal with creative-class, birkenstock-wearing, granola-chomping liberal upper-middle-class Democrats necessarily have to do with winning over Republican "upper-income" voters?
My question: so what evidence is there that Obama wouldn't appeal to women as the Democratic nominee, or indeed, that McCain would have any special appeal to them in 2008?
Indeed, recent Rasmussen polling of Pennsylvania testing a potential match-up between Obama and McCain suggests the exact opposite conclusion:
In Pennsylvania, Clinton leads McCain by six among women but trails by nine among men. Obama leads McCain among both men and women, but has a larger lead among women. McCain wins 69% of the conservative vote when matched against Clinton, 60% against Obama.
Recent polls in other states have shown the same phenomenon --- not only does Obama perform better in head-to-head general election matchups, he's also performng better among women. Here's Rasmussen's Oregon poll:
The current poll finds Clinton leading McCain by four points among men but trailing the Republican by eleven points among women. Obama leads McCain by fourteen percentage points among women and by five among men.
SurveyUSA found the same Obama advantage among women in Oregon, leading McCain by seven points instead of Clinton's four. And here's Rasmussen's latest New Hampshire poll:
In New Hampshire, McCain currently trails Clinton by twenty-five points among women and Obama by twenty-six.
According to SurveyUSA, Obama even matches Clinton's near thirty-point advantage among women in New York over McCain, while performing significantly better among men --- giving him a 21-point lead over McCain compared to Clinton's 11-point lead... in Clinton's own home state.
More SurveyUSA results: Obama does five points better among women in Iowa, four points better among women in Wisconsin, and in four points better among women in Virginia.
But going back to Brownstein's point, Obama seems like the candidate who can equal Clinton's advantages among women while doing significantly better among men --- the reason that he leads in all of these general election head-to-heads while Clinton struggles. In other words, Obama appeals to Clinton's coalition, but Clinton doesn't hold the same appeal for Obama's coalition.
Second, McCain's total inability to address bread-and-butter economics issues, indeed his lack of attention to economics generally, doesn't speak well to his appeal to lower-income voters.
And lastly, the American public views the Iraq war much differently than it did in 2004. If McCain runs the campaign Bush ran in 2004 of "stay the course," he will lose.
None of these states have substantial Hispanic populations, leaving us without a lot of current polling to measure Obama's potential Latino support. But in Virginia, with its 4% Latino population, Obama lead McCain 72%-23%, while Clinton trailed 63%-35% --- admittedly a small sample size but not a good sign. Rasmussen does not provide crosstabs for Hispanic voters --- but they may be an interesting part of his 7-point lead over McCain in Colorado in a recent poll.
Now, don't get me wrong. These are general election polls reflecting the race as it stands today, not how it will stand in eight months, and the election could play out as Bernstein suggests --- there just isn't any evidence for it right now.
Second, polls also don't always register changes in turnout. Could Clinton turn out more women than would otherwise make it to the polls? Just as Obama would most likely turn out more young people? That's difficult to say. What we can say is that in these contested states, the evidence we have from most recent polling suggests that Obama pretty much universally performs as well if not better than Clinton.