David Brooks wrote the other day (as a prelude to one of the Strange New Respect Santorum columns that conservatives have been writing these past few days with what strikes me as a deficit of sincerity; Charles Krauthammer has another) that the Republicans are "the party of the white working class." When I saw that I scratched my head, because I had a dim memory of looking into this once and finding that the political affiliation of the white working class was a question complicated by disagreements as to how you define "white working class" and the fact that it's shrinking. Brooks defined "white working class" as "whites with high school degrees and maybe some college," and he wrote that they "overwhelmingly favor Republicans."

The Berkeley sociologist Michael Hout has a useful blog post answering Brooks's claim. Maybe it's true, he wrote, that in Iowa the white working class (defined as high school graduates who went no further in school) is overwhelmingly Republican, but nationwide the GOP can't claim even a narrow majority. The white working class is pretty evenly split between Republicans (29 percent) and Democrats (26 percent). The rest are mostly independents. Conceivably when you count independent "leaners" the narrow Republican plurality becomes a majority, but no way would it be an "overwhelming" majority because independents don't lean "overwhelmingly" Republican. Brooks fudges his calculation with "maybe some college," which I guess means people who began but didn't complete college--of whom there are a lot, in part because college is so outrageously expensive nowadays. But are there really so many of them, and so many who vote Republican, as to make this group "overwhelmingly" Republican? And are people who have "some college" properly categorized as members of the working class? Many of them have been trained, at community colleges or elsewhere, to perform relatively skilled labor. I think it would be more accurate to call such workers "middle class."

Among white college graduates, Hout points out, the Republican plurality is more meaningful--38 percent to the Democrats' 27 percent. But the GOP edge among white college grads has been declining since the late 1980s, while the Democratic share of white college grads has been growing since 2004.

Another point worth making is that the working class isn't just white. It is also, increasingly, nonwhite. And among the nonwhite working class Democrats hold a 40-point advantage over Republicans.

None of this is to say that the working class and the Democrats are peas in a pod. They used to be, but starting in the 1960s the Democrats started losing the white working class. Starting in the 1970s they also lost some of their edge with the nonwhite working class. The initial catalyst for the change was the passage of civil rights legislation; sociologists and political scientists argue about how much the subsequent change was driven by hostility toward or discomfort with resultant changes in American society. I think the decline of labor played a big role, too. And yes, liberals made a few mistakes along the way, particularly in their attitudes toward crime, that didn't help matters.

The point is, beware sweeping generalizations about the political affiliation of the white working class.