Six weeks ago today, I gave birth to a baby girl. Like her older sister, she spent the first few days of life without a name.
You see, my husband and I wanted to get our children's names just right, and that meant taking some time to consider the options and get a feel for how well they fit each new baby. But we also happen to be cognitive scientists of an evidence-based persuasion so, for us, it also meant gathering and analyzing some data.
The first round of data collection was of the usual, informal sort, and it happened long before our first baby was due. We perused lists of baby names on websites, investigated name meanings and origins and looked up name frequencies over time. Then we generated a rough list of our favorites — the ones we could agree on, that is — and solicited limited feedback from our families. We vetoed input on how much our family members liked various names. But we did want to make sure we weren't overlooking any problems, such as cultural associations we didn't know about or possible meanings in other languages.
It was the next stage of data collection, which happened just before our first daughter's birth, that betrays our scientific training.
First, we browsed the research literature on how names affect their bearers. We learned, among other things, that an unusual name isn't such a bad thing (but it helps if it's easy to pronounce) and that naming your daughter Denise probably won't up the odds that she'll become a dentist, contra some previous claims.
Second, we used Amazon Mechanical Turk — a crowdsourcing platform that social scientists are increasingly employing to run experiments — to collect data from a couple-dozen strangers on our top baby names, plus some control names for comparison: names that we disliked, or that we did like but had ruled out for one reason or another, such as very high popularity.
We asked people whether the names had any strong associations for them, and what they would infer about the personality or religious or ethnic identity of someone with each name. And we received plenty of imaginative responses:
Katia: "I think of them as Russian (or a former Soviet block country), teenager, Christian. Makes me think that this person is a foreign exchange student or a super model."
Leila: "American or Western Asian. Possibly muslim, but perhaps not. It wouldn't surprise me to find that she comes from an educated family. I would expect her to be sophisticated."
Emily: "I imagine this person would be young, about elementary school age, white skinned, dark OR light eyes, and have brown hair in pigtails. The person would be Catholic, or Protestant, and they would be sweet natured."
Some names generated widely divergent responses. "Allegra" was assumed to be American, Latina, Middle-Eastern, Greek, Italian, English, European and "faux French," Muslim, Catholic, Protestant and Wiccan, young and old. "Kira" was assumed to be white, black, half-Asian, English, and Irish, but almost always young. Others were surprisingly consistent. Here were two of the responses for "Austen" as a girl's name:
contemporary little kid - probably caucasian/white, middle class to upper class kid - dresses cute, maybe a bit of a tomboy.
I would expect a WASP (white, anglo-saxon protestant) from a fairly affluent background.
Our final stage of data collection occurred in the hospital after the birth, and involved data from only two sources: my husband and myself. We wanted our daughters' names to open doors, not close them, and let's face it: some names that are cute for baby girls and poodles don't make for the most convincing judges or prime ministers. So we rated how well we thought each of our candidate names would work for a Nobel laureate in science, a novelist, a rock star, the secretary general of the United Nations and a CEO. Then we took our average ratings to generate a "career potential" index for each name.
By this point we'd been in the hospital for over 24 hours with our nameless "Baby Lombrozo," and the nurses were starting to give us funny looks. Also, our parents — preposterously enough — wanted to send out a timely birth announcement with a name. We had to pick soon, and to do that we had to figure out what to make of all the data we'd collected. At this point, we made a radical decision: ignore the data.
Now don't get me wrong — there's a time and a place for evidence-based decision-making. (In fact, it's probably most of the time and most of the place.) And we did take evidence into account when it was relevant and reliable. For example, we wanted a name that would work well in Spanish, which required consulting more proficient speakers. And we wanted a name that wasn't among the most common in the U.S., which made the Social Security dataset handy. These considerations helped with the shortlist, but they didn't give us a winner.
When it came down to the things that mattered most, our modicum of data just wasn't very useful. It mattered how much we liked each name, and how much we'd like it in the future. Polling others wouldn't tell us that. And it mattered that we launch each child into the world with the best prospects possible. But a name is only a name and the assorted prejudices of a handful of Mechanical Turk users (plus our own) wouldn't help much there, either.
It also mattered that we not regret our choice. And if our choice was based on highly indirect evidence that was supposed to tell us something about our child's future well being, then we'd constantly be tempted to reevaluate that choice as time went on and new data became available. So rather than basing our decision on our imperfect data, we decided to change the basis for our decision altogether: We went with our gut.
It's now been more than three years since Baby #1 was finally named, and more than a month since Baby #2 received her name. Between diaper changes and night-time wakings, we haven't had much time to reassess our final choices. But the evidence so far? No regrets.
You can keep up with more of what Tania Lombrozo is thinking on Twitter: @TaniaLombrozo