There is no consensus vocabulary to describe the subgroups, and, in part, as a result, many questions that might work elsewhere don't in the American context. Ask people what social class they belong to and 90% will tell you that they are middle class. Ask people what their ancestry is with an option to provide many responses, as the census does, and the results are too jumbled to make sense of. But, religious identification turns out to be a pretty convenient demographic marker, both because it has a strong inherited aspect, as most people end up in the religion of their parents or go through a significant personal identity changing process to get from point A to point B, and because the vocabulary is more clear than most approaches.
I personally don't care much about who does or does not read the DaVinci Code, but a Barna survey on the topic tantalizes us in the way that it implies findings of religious demographics in the United States.
American Catholics are more likely than Protestants to have read it (24% versus 15%, respectively). Among Protestants, those associated with a mainline church are almost three times more likely than those associated with non-mainline Protestant congregations to have read the book. Upscale individuals – i.e., those with a college degree and whose household income exceeds $60,000 – are nearly four times more likely to have read the book than are “downscale” people (i.e., those without a college degree and whose household income is $30,000 or less). . . . The people most likely to have altered their religious views in response to the book’s content were Hispanics (17% of those who read the book), women (three times more likely than male readers to do so), and liberals (twice as likely as conservatives). Upscale adults were also much more likely than downscale individuals to shift their thinking based on the novel.
The underlying data aren't available for free, but the corrolations between religion, politics and social class, likely motivated in large part by white subgroups, are intriguing. They seem to imply that there is more validity to the stereotypes of upper middle class, liberal mainline Protestant, and downscale, conservative conservative Christians, as well as strong leanings of women towards liberalism and men towards conservativism, than one might expect.
One also wonders if the fact that the survey was to some extent indirect, asking about a popular book, rather than directly probing into generally sensitive religion, class and politics, didn't produce a more honest response that the usual barrage of polling questions.