That's interesting, but why are you using such a narrow definition of working class?
(Though I agree the whites without college degrees as a proxy for white working class has its limitations).
When I think of the populist right constituency, I think it's largely made up of white men without degrees but they're usually not poor and they're often petty bourgeois rather than working class.
It's simultaneously too narrow and too broad. A lot of the "poorest" white people, measured purely by income, are going to be college students (or similarly un/underemployed people) from comfortable backgrounds*. And the working class (measured by culture and/or financial stability) extends a fair amount beyond just the bottom third. You could even argue that "working class" should cut off the bottom 20% or so, with those in full-on poverty being an entirely different class.
Obviously labels like "working class" is something we can argue endlessly about, and obviously income (like most variables used to measure concepts we're interested in in social sciences) is only an approximate measure.
Still, if you're interested in how the poorest White Americans vote, that's the best you've got (which, let's face it, is what a lot of people have in mind when they talk about this). And I will argue strongly against stretching the term "working class" too far up the income ladder.
Many of the "poorest white Americans" are in fact privileged college students. You think people have the campus of Georgetown in mind when they talk about the white working class?
"Many" of them, really? How many out of the maybe around 40-50 million people that would fall in this category would be College students? I doubt it's enough to skew the results too strongly.
Also, I'm not entirely sure so don't take my word for it, but I think the ANES measures family income, which should filter out college kids with wealthy parents.
According to
this source there were over 12 million under-25 (i.e. traditional age) college students in 2013. Presumably a majority of them were white, in excess of the percentage of white people overall... so almost certainly north of 10 percent of that 40-50 mil sample. That's enough to introduce quite a bit of skew!
And even if it is "family" income I'm not sure how you measure that besides counting households, and students living away from home in a dorm don't get to count as part of the household anymore. (Of course not all college students do that, so that would lower the percentage some.)
Perhaps there could be some sort of income cap in addition to the degree/salaried requirements, but "lowest third" is obviously too stringent (come on, a plumber who makes in the 45th percentile obviously belongs in this category), and the
vast majority of higher-earning individuals are filtered out by the salary requirement anyway.
