Charles Murray posts this interesting graph:
I’ll give his explanation and then some discussion of my own. First, Murray:
The General Social Survey, a mother lode of information for social scientists that has been collected annually or biannually since 1972, has asked people in every survey to say whether they are extremely conservative, conservative, slightly conservative, moderate, slightly liberal, liberal, or extremely liberal. A really simple question.
The graph represents the percentage of people who answered “extremely liberal” or “liberal” minus the percentage of people who answered “extremely conservative” or “conservative” in any given survey. I [Murray] won’t go into the statistical details (for that, buy the book in a couple of years), but think of the classes this way:
Traditional Upper: Someone at the 95th percentile of income, with a graduate degree, who is a business executive, physician, engineer, etc.
Intellectual Upper: Also at the 95th percentile of income and with a graduate degree, but a lawyer, academic, scientist (hard or soft) outside academia, writer, in the news media, or a creator of entertainment programming (film and television).
Traditional Middle: Same occupations as the Traditional Uppers, but with just a bachelor’s degree and at the 75th percentile of income.
Technical Middle: Someone working in the many technical specialties that have proliferated in health, information technology, and industrial technology, with an associate’s degree and at the 50th percentile of income.
Working: Someone working in a skilled blue-collar job, with just a high school diploma and at the 25th percentile of income.
Lower: Someone working at a low-skill job who didn’t finish high school, at the 5th percentile of income.
The graph is based exclusively on non-Latino whites (because that’s who the book is about).
And now, my thoughts:
1. Perhaps it;s a difference between journalism and academic writing that Murray will give the statistical details in his book but not in a blog entry, whereas for me, it’s the opposite, I wanted to keep Red State, Blue State clean and focused on the substance, but on the blog I can give as many statistical details as I want. In retrospect, though, maybe I should’ve included more statistical discussion throughout the book, and not just in the Notes section at the end. Adding statistical discussion would give the reader more of a sense of the process of discovery and maybe even make the book more interesting and fun to read.
2. The divergence of the “intellectual upper class” in Murray’s graph is dramatic, but isn’t it partly so because he’s excluding other categories? Beyond the problem with excluding ethnic minorities (whose fraction in the population, and in the electorate, has been growing), he’s also excluding lots of white people. For example, where are “intellectual non-uppers”–that is, people in intellectual jobs such as teachers, writers, scientists, grad students, etc., who are below the 95th percentile of income? Grad students–even white grad students–are really really liberal, I think, and even at Columbia, they’re not at the 95th percentile of income. And there are a few million teachers out there, not to mention lots and lots of professors making less than the 95th percentile, etc. Also, where do nurses with college or graduate degrees fit into the picture? Or college-educated journalists? I’m not saying that Murray is wrong here and I’m not trying to “debunk” him; I just worry that including only a subset of categories can give a distorted view of the trends. (I also don’t quite see the logic of putting scientists and MTV directors in a common category–but only if they’re rich.)
3. As we know, Republicans tend to be richer than Demoocrats, but liberals and conservatives, on average, hav e pretty much the same income profiles as the general population. We can redo our graphs with just whites alone and compare to what Murray got. It just looks funny that a single high-income group stands alone in his plot.
4. Murray’s graph is pretty damn good, but let me offer a couple of suggested improvements that he can make before putting it into a book: (a) Label the y-axis at -10%, 0, +10%, +20%. The current excessive labeling is just clutter and makes it harder to follow the pattern. (b) Start the graph at 1972 (when the data start); it’s a little misleading to start at 1960. (c) Most importantly, get rid of the different dots and symbols. Just label the lines directly. (See Red State, Blue State for some examples of how to do this).
5. Murray writes, “every white socioeconomic class in America has become more conservative in the last four decades.” But what about the various groups he’s excluded from the graph? Again, I’m not saying he’s wrong, I just would like the full story. Also, in the following sentence of his article, it would be more amusing for Murray to say “careered” rather than “careened.”
6. In linking to Murray’s graph, Matthew Yglesias points to Murray as writing, “The late New Yorker film critic Pauline Kael famously said after Nixon’s landslide reelection, ‘How can he have won? Nobody I know voted for him.’” I’d just like to point out that “famously” != “actually.” I looked into this a bit and could never find any evidence that Kael said this. Beyond this, I’m sure she read the newspapers, which back in 1972 didn’t hide from people that Nixon was ahead in the polls, expected to win the election, etc.
What’s funny to me is how many people believe that Kael could really have thought this. Kael wasn’t an idiot. There are certain cognitive errors that nobody would make themsevles, but it somehow is easy to believe that others could make the error. See here for discussion of related points.