Last month, I linked to a paper (PDF) by Princeton’s Larry Bartels taking issue with Thomas Frank’s bestselling What’s the Matter with Kansas?

I focus on four specific questions inspired by [author Thomas] Frank’s account: Has the white working class abandoned the Democratic party? Has the white working class become more conservative? Do working class “moral values” trump economics? Are religious voters distracted from economic issues? My answer to each of these questions is “no.”

Via Eric Alterman, I see that Frank has issued a rejoinder (PDF). Frank makes an especially strong case against Bartels’ definition of the white working class:

Arming
himself with a fact that is well-known to poll-readers everywhere—that society’s very
poorest members tend to vote Democratic— [Bartels] simply switches the labels, claims that
those poorest Americans are “the working class,” and—hey presto!—declares the
problem solved. Nothing to worry about, go back to sleep.

Bartels performs this spectacular feat of reassurance by defining “working class” as
“people with family incomes in the bottom third of the income distribution,” (11)3 which
is to say, people with household incomes below $35,000. To justify this definition, he
brushes off other ways of determining working-classness. First, subjective self-
identification: You can’t trust people to give their own class accurately, Bartels writes,
because that might be an effect as well as a cause of their political views. (page 11.
Remember this: it will resurface later on.) Second, Bartels argues that educational levels
can’t be used to determine “socio-economic status” because more people get college
degrees now than in the old days.

And that’s it. Bartels’ own definition—household income below $35,000—is simply
supposed to be beyond controversy, as he implies in a double-negative passage so
windingly indirect that I had to read it over three times to get it:

Nevertheless, as a general matter, it does not seem implausible to suppose that
people’s relative positions in the current income distribution provide a
meaningful, historically consistent indication of their socio-economic status.

Well, I can think of half a dozen reasons why it is implausible, and I’m not even a
political scientist. Maybe those people are students or recent college graduates or just
starting out in the workforce or not married yet: They will promptly cease to be “working
class” by Bartels’ estimation as soon as they get a promotion or get married. Maybe those
people are professionals who just don’t make very much money, like, say, a newly
minted poli-sci PhD who is an instructor or an adjunct somewhere: They may be
struggling but they are not “working class” by anyone’s definition. Or maybe those
people are retirees, living on pensions or Social Security—a large and growing segment
of the population, but not necessarily “working class.”

In fact, according to a pair of professors who have also analyzed the NES data for 2004,
over a third of Bartels’ “working class” demographic are, in fact, retirees. Eight percent
are disabled. Only a third of his chosen cohort are actually employed, and only half of
these are over the age of thirty. This is not a profile of “the working class” as anyone
uses the term. This is “the poor,” “the young,” and “the retired.”