Poli-Sci Perspective is a weekly Wonkblog feature in which Georgetown University's Dan Hopkins and George Washington University's Danny Hayes and John Sides offer an empirical perspective on the issues dominating Washington. In this edition, Sides looks at the politics of...baby names. For past posts in the series, head here.

Republicans and Democrats don't seem to agree on very much these days. They're divided on the kinds of television shows they watch, cars they drive and beers they drink. And now new research by political scientists at the University of Chicago adds one more thing to that list: baby names.

Why study baby names at all? In order to understand all those other differences. The authors of this paper -- Eric Oliver, Thomas Wood and Alexandra Bass -- note that it's often unclear whether partisan consumer habits are really due to partisanship. It could be that companies successfully market products to specific demographics that happen to have a partisan leaning. Baby names are different. As Oliver and colleagues write, baby names "are highly related to taste and fashion but largely free from market effects."

To understand whether Democrats and Republicans choose different kinds of baby names, the researchers compiled an unusual set of data. They took all of the births in the state of California from 2004 -- about 500,000 in all. For each baby born, the data contained the child's first name, the mother's first name, the father's first name (where available) and the mother's education, race and address. Using these addresses, they then matched each mother to her Census tract and thereby determined whether she lived in an area that was predominantly Democratic, Republican or somewhere in between. The question is whether mothers who lived in red, blue and purple neighborhoods were systematically different. They were, in two respects.

The first difference has to do with whether the baby's name was unusual. Oliver and colleagues ascertained whether each baby's name was unique (such that no other child born in California in 2004 was given that name), uncommon (20 or fewer children born that year were given that name), or popular (one of the 100 most common names in California that year). Unique baby names were more common among blacks and Asian Americans than among whites and Latinos. Within any racial group, unique baby names were more common when the mothers had less formal education or lived in a lower-income neighborhood.

But among whites, partisanship and ideology mattered, too. Mothers who had at least some college education were more likely to give their child an uncommon name -- and less likely to give the child a popular name -- when they lived in relatively Democratic or liberal areas.  If neighborhood characteristics corresponded to the mother's own characteristics, better-educated Democrats or liberals were more likely to give their babies unusual names than better-educated Republicans or conservatives.

This leads to the second difference: the names they chose. Oliver and colleagues find that there were roughly two kinds of uncommon baby names: ones that are completely made up or just different spellings of common names (like "Jazzmyne" for Jasmine), and ones that are just esoteric. When racial minorities and the poor chose uncommon names, they were more likely to choose the former. When Democrats or liberals chose uncommon names, they were more likely to choose the latter.

Oliver and colleagues argue that liberals, consciously or unconsciously, signal cultural tastes and erudition when picking their child's name. In conversation with me, Oliver used himself as an example. He and his wife, a novelist, named their daughter Esme -- a name gleaned from a story by the writer J.D. Salinger.

On the other hand, conservatives, by being more likely than liberals to pick popular or traditional names (like John, Richard, or Katherine), signal economic capital. That is, they are choosing names traditional to the dominant economic group -- essentially, wealthy whites. Oliver noted to me that some immigrants also try to help their children assimilate and succeed by choosing names in this fashion. And, given research that shows that the ethnic connotations of a job applicant's name can affect the possibility of getting an interview, choosing names this way may make economic sense.

The names chosen by Democrats and Republicans differed in another respect: how they sound. Oliver and colleagues categorized each name by whether the sounds, or phonemes, it contained were more common in boy names or girl names. Boy names are more likely to contain "hard" sounds -- consonants like K, B, D, T -- while girl names are more likely to contain "soft" sounds -- like the L's in "Lola," the A in "Ella," and the Y in "Carly." Oliver and colleagues found that, for both boy and girl babies, "softer" sounds were more prominent among educated whites living in more Democratic or liberal neighborhoods. That is, a boy's name like "Julian" or "Liam" or a girl's name like "Malia" would be more common in Democratic neighborhoods. A boy's name like "Trig" or a girl's name like "Bristol" would be more common in Republican neighborhoods. (Oliver and colleagues cannot help but note that the Obamas and the Palins conform to their findings.)

Oliver, Wood and Bass do not have a definitive explanation for why the sounds of names might depend on partisanship or ideology. Oliver speculated to me that it could be tied to economic capital. Traditionally, masculinity has been associated with economic success. Thus, he suggested, just as conservatives signaled economic capital by choosing more popular names, they also did so by choosing names that "sound like" the kinds of names you might fight among the economic elite.

It is important not to exaggerate the tendencies that this research uncovers. Many names --- Oliver mentioned "Joshua" to me -- cross political boundaries. And partisanship or ideology is just one ingredient in naming decisions, alongside ethnic, religious and familial traditions and general fads. At the same time, given that these other ingredients are likely more salient to families than partisanship, Oliver and colleagues were surprised to find any differences based on partisanship or ideology.

Oliver and colleagues also emphasize that these partisan or ideological differences were largely confined to better-educated whites. As other political science research shows, partisanship and ideology often operate most strongly within this group. Thus, it is a mistake simply to divide America into red and blue. This leads to the paper's provocative conclusion:

As we see in patterns of baby names, liberal elites use esoteric cultural references to demonstrate their elevated social position just as conservatives invoke traditional signals of wealth and affluence. Instead of divides between “Red and Blue states,” it is more accurate to say that America is divided not just by “Red and Blue elites,” but also in the ways these elites seek to differentiate themselves from the largely “purple” masses.