Noam Scheiber over at TNR &C. has a
post about the demographic breakdown of Kerry and Edwards supporters in WI that shows how they tend to be richer and better educated than Kerry suporters, the opposite of what might have been expected.
Consider some of the numbers. Kerry led Edwards by large margins among people who make under $15,000 per year and under $30,000 per year (rolling up 28- and 13-point leads against Edwards in these categories, respectively). This much you'd expect, since poor voters tend to be overwhelmingly liberal. The interesting thing is that Kerry opened up a 5-point lead among voters making between $30,000 and $50,000 per year--a category you'd probably think of as mostly blue-collar voters--while only beating Edwards by one point in the $50,000-$75,000 category, and actually losing to Edwards (and by 7 points!) in the $75,000-$100,000 category. And Edwards and Kerry were tied among voters making over $100,000. Other than the result among the poor, this is the exact opposite of what the conventional wisdom would predict, which is that Edwards does better among downscale-but-not-poor voters (who tend to be more culturally conservative), and Kerry does better among upscale voters (who tend to be more culturally liberal).
And it's not just an anomaly of the income data. You get pretty much the same result in just about every category that correlates with blue-collar-ness and white-collar-ness. Kerry led Edwards by 13 points among those with only a high school diploma, but the two candidates were tied among those with a college degree, and Edwards led Kerry by 4 points among those with some post-graduate training. Ditto union membership: Kerry beat Edwards by 8 points among households with a union member. He beat Edwards by only 3 points in households without a union member. What about the urban/suburban/rural divide? Same story. Kerry, as expected, carried big cities (with their large concentrations of poor people) by a whopping margin (some 21 points). But Kerry also won by a surprising 11 points in rural areas, and by a seven-point margin in small towns--both places more likely to be culturally conservative. Edwards, by contrast, won in relatively affluent enclaves, like suburbs, which he carried by four points.
He suggeststhat the reason might be that wealthy Republicans are switching parties but explains that he doubts that this would have as large an effect as is evident.
He argues that the real reason is that poorer people are less educated about the candidates and are voting for Kerry because they know he is the frontrunner and a veteran.
I think Scheiber is dead wrong. I think that the people who Edwards supposedly appeals to are in fact unimpressed with his message. I think that rich voters imagine that Edward's populist anti-trade rhetoric about keeping jobs in America resonates with poorer voters when in fact it doesn't. That would seem to be the simples explanation. It also suggests that the recent "edwards is more electable than Kerry" pushed by, among others, William Saletan isn't true. Any candidate that can't appeal to poorer workers will have a hard time in key battleground states, particularly Ohio.