In America, you are what you eat

By THE DATA TEAM

WHAT is soppressata? Google searches for the Italian meat surged last week, thanks to a column by David Brooks in the New York Times in which he recounted an awkward lunch at an upscale delicatessen with “a friend with only a high-school degree”. Upon suspecting that his less-credentialed companion might have felt alienated by a menu which listed ingredients such as “soppressata, capicollo and...striata baguette[s]”, Mr Brooks and his colleague then retreated to a Mexican restaurant, which he surmised would constitute a class-neutral haven.

Mr Brooks used this anecdote to highlight a broader argument: that cultural social barriers, such as differences in what people eat, contribute just as much to inequality as economic trends and government policies do. Snarky commenters on social media—many of them the sort of educated coastal liberals who might frequent gourmet sandwich shops—promptly tore into Mr Brooks, both for his theory and his anecdote. But whereas proving that there is a causal link between social mobility and salami savvy is difficult, it appears that Mr Brooks’ general observation that Americans’ culinary preferences vary by social class rings true.

To test this hypothesis, we asked YouGov, a pollster, to survey 1,500 American adults about how often they dine out, how often they eat sushi and how often they eat four particular ethnic foods: French, Mexican, Italian and Indian. Some cuisines did seem to be strong markers of educational attainment. For example, 57% of respondents with postgraduate degrees have eaten sushi within the past year, compared with just 26% of those who never attended college.

Similarly, despite the mockery that Mr Brooks’s tale sparked online, his intuitions turned out to be spot-on. For fear that soppressata itself would be too obscure to ask about in a medium-sized survey, we substituted the better-known term, prosciutto. Indeed, 25% of those with high-school degrees or less had never heard of the cured ham, compared with 11% of postgraduates. The data also support his assumption that Mexican food was likely to span the country’s class chasm: less-educated Americans were nearly as likely as college graduates were to have tucked into a taco.

Democrats who fancy their party to be the natural home of sophisticated cosmopolitans might be cheered to learn that their political fellow-travellers have slightly more exotic palates than Republicans do. But whatever differences exist mostly reflect Democrats’ higher average education relative to their level of income, and their habit of clustering in cities, rather than being a partisan characteristic. Rural Democrats who did not attend college are no more likely to scoff the likes of soppressata than their Republican counterparts are. However, Indian food appears to be an exception. A simple logistic regression on the survey data reveals that Republicans are around 40% less likely to consume it than are Democrats in areas of similar urban density with equivalent schooling and wealth.

Near the end of his piece, Mr Brooks asserts that it is cultural codes rather than prices that ensure that 80% of shoppers at Whole Foods, a posh grocery chain, are college graduates. It’s difficult to separate the effects of education and income on consumer tastes, since richer people also tend to be better-educated. However, the YouGov numbers do suggest that the degrees a person holds are a statistically significant predictor of culinary preferences, even after accounting for household earnings. The culinary chasm may not be the cause of America’s class cleavage, but at the very least it seems to be a highly visible symptom of it.