Rick writes:
, basically, that this whole debate is wrongheaded. The problem is not Republican performance among the working class, but stupid journalists. Gelman noticed that incomes follow an odd pattern when you break them down by states: In poor states, income is heavily related to partisan affiliation, with the rich being almost monolithically Republican. In rich states, income is less strong correlated to voting preferences. Journalists live in rich states. They know a lot of rich Democrats. They tend to assume Democrats are strongly supported by the rich. Meanwhile, they know that states like Mississippi and Arkansas are poor, and are Republican, and they presume that means that poor voters are Republican. That's not true, but journalists are lazy, and don't notice that that's not true, and so we have an odd political discourse in which people are confused by the Democratic coalition even as its composition is not, in any real way, confusing.All that said, no one denies that Republicans win some working class voters, and the question that someone like Tom Frank is asking is, simply, "why?" There's no doubt that Democratic economic policies are, in the aggregate, better for the working class. And over time, there's little doubt that Democratic presidents produce more income growth for all income brackets, and do so more equitably. Bartels even has a relatively famous graph to that effect:
Someone needs to clear this up for me. The thesis of 'What's the Matter With Kansas' is that Republican economics don't benefit enough people to win elections, so they exploit hot-button culture issues to win. But Larry Bartels says "no" the working class never really left the Dems. And it is true that if only the bottom third of the income distribution voted, Dems would win big time. So then which is it? Has the right peeled away the working class with culture war issues? And if Bartels is right, how have they been winning elections?This is a complicated question, and I have nothing definitive to say on it, but the outcome of the 2008 election makes it interesting to dig into.First, you have to somehow handle the methodological dispute between Thomas Frank and Larry Bartels, which Matt Yglesias adeptly summarizes here. The essential question is how you define working class. If it's by income, then Democrats win the white working class. But the bottom third of the income distribution includes students, retirees, and non-profit magazine writers, which isn't exactly what folks are going for when they reference the white working class. If you're defining it through education -- say, those who lack a four-year college degree -- you've rolled two-thirds of the electorate into your definition, their average wage is above the median wage, and one of them founded Microsoft. Andrew Gelman, a statistician at Columbia, would argue
, basically, that this whole debate is wrongheaded. The problem is not Republican performance among the working class, but stupid journalists. Gelman noticed that incomes follow an odd pattern when you break them down by states: In poor states, income is heavily related to partisan affiliation, with the rich being almost monolithically Republican. In rich states, income is less strong correlated to voting preferences. Journalists live in rich states. They know a lot of rich Democrats. They tend to assume Democrats are strongly supported by the rich. Meanwhile, they know that states like Mississippi and Arkansas are poor, and are Republican, and they presume that means that poor voters are Republican. That's not true, but journalists are lazy, and don't notice that that's not true, and so we have an odd political discourse in which people are confused by the Democratic coalition even as its composition is not, in any real way, confusing.All that said, no one denies that Republicans win some working class voters, and the question that someone like Tom Frank is asking is, simply, "why?" There's no doubt that Democratic economic policies are, in the aggregate, better for the working class. And over time, there's little doubt that Democratic presidents produce more income growth for all income brackets, and do so more equitably. Bartels even has a relatively famous graph to that effect: 
