Understanding the Confusion About Class and Voting

John Sides and Larry Bartels have recently spent some space explaining the “political science” view of social class and voting in American politics, in contrast to the claims of journalists Thomas Frank, George Packer, and Jonathan Chait, that working-class whites vote Republican. Frank, Packer, and Chait are outspoken liberals, but their views of class and politics align well, at least in this point, with arguments by conservatives such as David Brooks, Michael Barone, and Charles Murray about the prominence of upper-class liberals.

John and Larry presented the data clearly. What I’d like to add here is a brief discussion from Red State Blue State about how all this confusion can have arisen. First take a look at this graph of trends in voting by occupation class. We created the graph using data from the National Election Study. (Sociologists Clem Brooks and Jeff Manza published a similar graph before we did; our improvements were to include more years of data and to simply plot raw differences. Brooks and Manza fit some sort of statistical model, a step I would often favor, but in this case the secret weapon seemed like a better choice.)

The key message from this graph is that there have indeed been changes over time in the social composition of the Democratic and Republican vote. So, even while I agree with Larry and John about the voting patterns of working-class whites, there is something real that Frank, Packer, Chait, Brooks, Murray, Barone, etc. are trying to get at. It’s important to point out that they got the details wrong (for example, when Barone, in his eagerness to slam the hated liberal “trustfunders,” mangled vote reports from Colorado) but it could also help in this discussion to find points of contact in data trends that are consistent with the journalists’ impressions.

My second point is the now-familiar idea that rich and poor vote quite differently in faraway (to me) “red states” in the middle of the country but not so differently in the richer, more urbanized “blue states” where many of the people in this discussion (actually, all of them, I think, before Larry moved to Tennessee a couple years ago) live.

I’ve shown some variant of this graph many times:

This time let me also share my graphs of Maryland and Texas. In Maryland, some of the richer areas, including David Brooks’s own Montgomery County, counties are among the most Democratic:

In Texas, not so much. The richest county is a highly-Republican suburb of Dallas:

In addition to their direct relevance to voting, I believe these red-state blue-state differences are important to the discussion in that it relates to the evidence journalists get from their own perceptions. If, like Packer, Chait, etc etc., you live in a blue state, you’ll see lots of rich liberals, and you can naturally mistake the patterns you see as reflecting something in the country as a whole.

All this isn’t helped by the fact that rich journalists, according to an Indiana University study of a few years ago, are more likely than lower-income journalists to vote for Democrats. For most groups of Americans, the richer you are, the more likely you are to vote Republican—-but not for journalists.

In short, in answer to Marx’s classic question, “Who you gonna believe, me or your lying eyes?”, I say: Believe the statistics. But then again I would say that, I’m a statistician.

[Cross-posted at The Monkey Cage]

Andrew Gelman

Andrew Gelman is a professor of statistics and political science and director of the Applied Statistics Center at Columbia University.