It’s hard to measure something like the effect of a name. But Freakonomics co-author Steve Levitt has spent his academic career trying to come up with clever ways to measure things. And he’s thought quite a bit about the names we give our kids.
“I think it really is about the parents,” Levitt says. “As I’ve studied naming, what I’ve come to believe is that the primary purpose, when a parent gives a name, is to impress their friends — that they are whatever kind of person that they want to be.”
On this week’s episode of Freakonomics Radio, we brought back a classic. It’s called “How Much Does Your Name Matter?” and it’s about whether who you are, and who you turn out to be, may be related to what you’re called when you’re born.
Beginning in the 1970s, at the height of the Black Power movement, African Americans started giving their children very different names than white Americans. (In the decades immediately prior to that, black and white Americans had typically given their children similar names.) Some years back, Levitt started to wonder if these distinctively black names mattered — that is, whether they affected, for better or worse, the life of a kid with such a name.
So, Levitt launched a research project with Roland Fryer, an economist at Harvard who is devoted to explaining the gap between blacks and whites in education, income, and culture. (We should note Fryer has recently been placed on a two-year administrative leave for “unwelcome conduct of a sexual nature.”)
“The ultimate question we wanted to answer is, does your name matter for the economic life that you end up leading,” Levitt says. “Are people who are ‘saddled’ with distinctively black names facing a burden when they enter the labor market?”
Levitt and Fryer found a dataset from the state of California that included: the first and last name of every person born in the state between 1960 and 2000; the first, last, and maiden name of every baby’s mother; and information about the hospital where the baby was born and the health care the mother received. The dataset allowed Levitt and Fryer to compare outcomes (for example) for two African American girls born in the same year, at the same hospital, one with a black-sounding name (like Shaniqua), the other with a more traditional name (like Anne or Elizabeth).
“We were able to see something quite remarkable, which is that the name that you were given at birth seemed not to matter at all to your economic life,” says Levitt.
What Levitt and Fryer found, in other words, is that it’s not the name your parents give you; it’s the kind of parents you have in the first place. And different kinds of parents of course choose different kinds of names. So let’s say two similar families, both African American, each have a baby girl. One is called Molly, which is one of the whitest girls’ names in America. The other is called Latanya, which is a distinctively black name. Now, if, decades later, Molly becomes a professor at Harvard, and Latanya is just barely scraping by, the reason won’t be because Latanya’s parents named her Latanya.
Not long ago, Latanya Sweeney, a professor of government and technology at Harvard, was working with a colleague named Adam Tanner in her office. The researchers needed to find an old paper Sweeney had authored, so Tanner went to Sweeney’s computer and Googled her name. Alongside links to Sweeney’s papers and her bio, an ad popped up that read: “Latanya Sweeney, Arrested?”
Sweeney (who has never been arrested) was shocked. The two researchers next Googled a typical white, male name: Adam Tanner. A neutral ad popped up from the same company, but the word “arrested” did not appear.
The ads were for a company called Instant Checkmate, which sells public records. The ads appear when you do a Google search for the first and last name of a real person. But a given name search might generate different versions of the ad. Some of them are neutral, like “Looking for Molly Sweeney?” Others, like the one Latanya Sweeney found, seem to offer up arrest records.
Sweeney and Tanner started doing lots of name searches to see if they could find a pattern to the ads. They started by comparing “Latanya” to “Tanya,” and found that the former produced an ad suggesting the person had an arrest record, while the latter did not. Tanner suggested that the arrest ads were linked to black-sounding names. Sweeney thought the idea was crazy and set out to prove him wrong.
The first step for Sweeney was to simply define what a black name is and what a white name is, so she assembled some data on the whitest and blackest names among baby boys and girls. In order to prompt the Google ads, Sweeney needed to find real first and last names, some black and some white. So she typed in searches like “Shanice Ph.D.” or “Molly MBA” to find real people — some of whom were, like herself, professionals — and then she fed those real names back into Google to see what ads they’d prompt. She found that the black names were 25% more likely to produce an ad suggesting a person had an arrest record. The ads appeared regardless of whether the name was associated with a criminal record in the company’s database.
Google makes most of its money with a program called Google Ads, which used to be known as AdWords. It serves ads that are linked to the content that you search for. Advertisers, like Instant Checkmate, agree to pay a certain amount each time their ad is clicked on. They provide Google with several versions of ad text, and they can specify which keywords — or in this case, which key names — will prompt each version of the ad. It is, of course, in the best interest of both Google and the advertiser to serve the ads that will get the most clicks.
For the record, a Google spokesperson told us that “AdWords does not conduct any racial profiling… It is up to individual advertisers to decide which keywords they want to choose to trigger their ads.” Instant Checkmate didn’t respond to our query, but an official statement from the company about Latanya Sweeney’s study says, “Instant Checkmate would like to state unequivocally that it has never engaged in racial profiling in Google AdWords, and that we have absolutely no technology in place to even connect a name with a race.”
So whoever the villain is here — and it may be us, the people who click — the point is that in this case, your name matters. It’s easy to imagine a circumstance wherein Latanya Sweeney, before she got hired at Harvard, might have suffered the consequences of her name if an HR person was Googling her and saw that “arrested” ad — even if the HR person didn’t bother to click on the ad, and even though Latanya Sweeney, herself, hadn’t been arrested.
“When I got really moved… was… looking at the faces of the names of these young Ph.D. students and people who are just launching their careers,” Sweeney says. “There was one… it was a young woman, she was so proud, she had just published her first paper, she was a graduate student in a Ph.D. program. And there’s her name, and there’s this ad, ‘arrested.’”
There’s other research which finds that a name may matter, at least on some dimensions. Boys with feminine names, it’s been argued, act up more in school. A girl with a masculine name, meanwhile, is more likely to have a successful legal career.
Another study, by Marianne Bertrand and Sendhil Mullainathan, was called “Are Emily and Greg More Employable Than Lakisha and Jamal?” This study found that if you send out a resume with a white-sounding name, it’s about 50% more likely to get a callback than an identical resume with a black-sounding name.
So which argument is right? Does a name matter? Or does it not matter?
Steve Levitt thinks there are a couple of ways to reconcile the results of the resume study with his research finding that names don’t matter. Callbacks, for example, don’t necessarily equate to job offers. Levitt also suggests that people with black names might struggle in the formal labor market but fare better in the informal labor market.
“You could certainly imagine that, within the black community, having a distinctively black name would help you get along better with people, signal that you’re part of the community, and might work in your favor in all sorts of informal networks that aren’t captured in these data,” Levitt says.
Even if your name doesn’t affect your life in any significant way, it can tell people a little something about who your parents are. There are patterns to be gleaned from names data about your parents’ values, and their social standing. Levitt and Fryer, for example, found a clear trend in their data from California: almost all popular names start as “high-class names,” then migrate their way down the income distribution, at which point more-educated parents stop using them.
Eric Oliver, a professor of political science at the University of Chicago, is interested in ideological polarization, whether liberals and conservatives have different tastes, and if they signal those tastes and values differently. Oliver and his coauthors Thomas Wood and Alexandra Bass thought that baby names would be a great way to study these questions.
Like Steve Levitt, Oliver and his coauthors used the very rich database from the state of California. In addition to listing every baby born, it contained information about the mother, including age, race, education level, and ZIP code. It’s easy to figure out from this information whether the family lives in a predominantly liberal or conservative neighborhood. Oliver found that, in general, more-educated mothers were more likely to give their children popular names, and less likely to give their children uncommon names. 36% of African American mothers with less than a high-school degree gave their daughters unique names.
For the most-educated mothers, however, ideological differences emerge. Conservative, educated mothers tend to give their children common names, while liberal, educated mothers are more likely to use uncommon names. These names are different from the kinds of uncommon or unique names given by less-educated, non-ideological mothers in Oliver’s studies, who tended to change the standard spelling of typical names, or made up entirely new names.
“Whereas our educated-liberal mothers tend to be choosing names that are obscure cultural references,” Oliver says. “And so these are the Esmés, and the Unas, and the Archimedes, and the Emersons. And we think this is a way that liberals sort of signal — for lack of a better word, their sense of cultural superiority. It’s a way of signaling great cultural capital.” (For the record, Oliver’s daughter is named Esmé, in reference to a short story by J.D. Salinger.)
Oliver and his colleagues also found that the names that liberals and conservatives choose for their children even sound different. Conservatives choose more masculine-sounding names (monosyllabic, with hard consonants), while liberals lean toward feminine-sounding names (ending with an “a” or “e” and often including the letter “l.”).
“If you really want to know the most quintessentially-ideological sounding names, let’s compare the Obama girls and the Palin kids,” Oliver says. “So the Obama girls are Sasha and Malia, very nice, feminine, soft-sounding names. And then think about the Palin kids. We have Trig, Track, Bristol, and Piper. There’s Willow there too, and I think that was an ideological hiccup on Sarah Palin’s part.”
One thing that most of us can probably agree on: Just about every parent thinks that his or her kid is special, on some level. And part of what makes each of our kids special is the names we give them. But from what we can tell, your name is not your destiny — even if your name is Destiny. Or Esmé. Or Archimedes, or Track.
About this PODCAST