I [Wattenberg] recently came across a list I created back in 2009. Parents Magazine had challenged me to look a decade into the future and predict the top baby names of 2019. To make it more interesting, I restricted my choices to names that ranked outside the top 40 at that time. Any bullseyes would be genuine predictions of change, not just staying the course.
My prediction list:
As of 2019, only three of the names—Amelia, Harper and Lucas—actually made it all the way to the top 10. (Henry joined them a year later.) As a group, though, the names rose in popularity by a hearty 88% from the statistical baseline I was working with.
So far, that looks like a solid set of predictions. But wait. To predict future top 10 hits, I naturally chose names that were already on the rise. Did my so-called expertise add any predictive value, or did the names simply follow their expected trendlines?
As a test, I created a second control list. These names were statistical doppelgangers to the predicted names at the starting point: close in popularity, and rising at a similar rate.
It turns out that the doppelganger names rose less dramatically and were already in decline a decade later. None made it to the top 10. Score one for expertise!
What made my prediction list different from the control group? Looking name by name, the answer appears to be sounds—specifically, trendy ending sounds. My choices mostly avoided the hot suffixes of the 2000s like -den and -ton. My one concession to the trend, Peyton, turned out to be my single worst prediction. Meanwhile the biggest hits I failed to identify, Scarlett and Willow, have much more distinctive sounds.
The lesson is that not all popularity is created equal. Names that are part of creative sound trends share their style space with dozens of other similar names.
Good stuff. Also fun that she did a control group. We associate the control group with causal inference, but here she’s using it to evaluate a prediction, which mathematically is a similar problem–it’s a comparison!–but without the direct estimation of any causal effect.
4 thoughts on “How did Laura Wattenberg’s baby name predictions turn out, 15 years later?”
The name “Paul” has long since become unpopular and unnewsworthy in the English language. But, my “family name” is sort of unexpected in several senses. It was bestowed on Ellis Island around 1900 and despite the Bucharest old country, this is what pops up
“The name Alper originates from Turkish roots, combining two elements: “alp” and “er.” In Turkish, “alp” translates to “hero” or “brave,” and “er” means “man” or “soldier.” Hence, Alper can be roughly interpreted as “brave man” or “heroic soldier.” This etymology reflects the values of courage and strength, making Alper a name imbued with positive connotations.”
I love this! This is both treating the prediction seriously in the sense of genuinely trying to confirm or disconfirm it, and also taking seriously the communication/visualization piece.
Yes. As a society we need a lot more of this. Wattenberg and the whole baby name exploration is a great example of what statistical thinking and scientific thinking can contribute to topics outside regular “science”.
I have no idea of Wattenberg’s background, perhaps she has statistical training. But plenty of people without such training still have the right instincts. There’s a beekeeper/oysterman/farmer on North Haven, Maine, who I’ve found it interesting to speak with over the years. He tries different types of bee husbandry practices to see what works best…he reads about the different options, he’s not just making stuff up, but then he makes some informed choices and tests them. His dad is the same way, always trying different things. Both father and sun usually do simple comparisons and don’t do any statistical tests, but they have a good intuitive understanding of the role of variability and from examples I’ve seen they don’t indulge in noise-chasing.
Recently, the younger one has been cutting and mulching Glossy Buckthorn — an invasive species of small tree that takes over fields and creates monoculture forests that provide little habitat value for local birds etc — and at the same time he happened to be reading a book by an orchard-keeper. The book says that the best mulch for fruit trees is from small branches and small trees, because the ratio of cambium to xylem is higher and that creates a richer environment for fungi and soil bacteria. (Something like that). The buckthorn he was mulching was mostly small trees, perfect, so he started mulching around his fruit trees…but not all of them. Six months or so later, he took a few soil samples from under the mulched trees and some from under the trees he had mulched with conventional wood chips, and had them tested by the state agricultural service, and the guy doing the testing said the ones from the buckthorn mulch were “among the richest samples he had seen that year” or something like that. This is a useful scientific experiment in spite of its informality. I find this sort of thing — like Wattenberg’s baby names forecast and her post facto evaluation of it — to provide a heartening counterpoint to the more dispiriting stories on this blog: there are so many highly paid, highly educated researchers out there publishing bullshit, but at least there are also people quietly going about their business and doing good stuff, even if it never appears in a scientific journal.
The name “Paul” has long since become unpopular and unnewsworthy in the English language. But, my “family name” is sort of unexpected in several senses. It was bestowed on Ellis Island around 1900 and despite the Bucharest old country, this is what pops up
https://venere.it/en/the-meaning-and-history-of-the-name-alper/
“The name Alper originates from Turkish roots, combining two elements: “alp” and “er.” In Turkish, “alp” translates to “hero” or “brave,” and “er” means “man” or “soldier.” Hence, Alper can be roughly interpreted as “brave man” or “heroic soldier.” This etymology reflects the values of courage and strength, making Alper a name imbued with positive connotations.”
On the other hand, as my linguistic father-in-law enjoyed pointing out, in the Basque language,
https://projetbabel.org/basque/dictionary.php?q=alfer
“Lazy. From auher or alper.”
I love this! This is both treating the prediction seriously in the sense of genuinely trying to confirm or disconfirm it, and also taking seriously the communication/visualization piece.
Yes. As a society we need a lot more of this. Wattenberg and the whole baby name exploration is a great example of what statistical thinking and scientific thinking can contribute to topics outside regular “science”.
I have no idea of Wattenberg’s background, perhaps she has statistical training. But plenty of people without such training still have the right instincts. There’s a beekeeper/oysterman/farmer on North Haven, Maine, who I’ve found it interesting to speak with over the years. He tries different types of bee husbandry practices to see what works best…he reads about the different options, he’s not just making stuff up, but then he makes some informed choices and tests them. His dad is the same way, always trying different things. Both father and sun usually do simple comparisons and don’t do any statistical tests, but they have a good intuitive understanding of the role of variability and from examples I’ve seen they don’t indulge in noise-chasing.
Recently, the younger one has been cutting and mulching Glossy Buckthorn — an invasive species of small tree that takes over fields and creates monoculture forests that provide little habitat value for local birds etc — and at the same time he happened to be reading a book by an orchard-keeper. The book says that the best mulch for fruit trees is from small branches and small trees, because the ratio of cambium to xylem is higher and that creates a richer environment for fungi and soil bacteria. (Something like that). The buckthorn he was mulching was mostly small trees, perfect, so he started mulching around his fruit trees…but not all of them. Six months or so later, he took a few soil samples from under the mulched trees and some from under the trees he had mulched with conventional wood chips, and had them tested by the state agricultural service, and the guy doing the testing said the ones from the buckthorn mulch were “among the richest samples he had seen that year” or something like that. This is a useful scientific experiment in spite of its informality. I find this sort of thing — like Wattenberg’s baby names forecast and her post facto evaluation of it — to provide a heartening counterpoint to the more dispiriting stories on this blog: there are so many highly paid, highly educated researchers out there publishing bullshit, but at least there are also people quietly going about their business and doing good stuff, even if it never appears in a scientific journal.
Wattenberg has a masters degree in cognitive psychology from Stanford (hence some statistical training).