X writes:
Sur France Inter ce matin, le 23 septembre est le jour en France avec le plus de naissances (+5%). Have you done the same analysis on births for France than you did for the US?
I replied that we (Aki, really) have not analyzed any French data, but if you have the numbers, we could take a look. From the U.S. data and the cover of BDA, I recall that late Sept is the peak of births. It’s not a day-of-year effect, it’s a seasonal effect, I assume it reflects when people get pregnant, 9 months earlier. Or it could reflect differential rates of miscarriages, but I’m guessing it’s variation during the year of when people get pregnant.
X replied:
I found this rather fascinating paper by Gérard Calot who founded INED, and a more recent paper [by Arnaud Régnier-Loilier and Jean-Marc Rohrbasser] about the historical moves for most frequent periods, but no daily data so far.
This is great stuff. I love birthday data. I guess I should collaborate with some demographers! In the meantime, above and below I’ve reproduced some of the fun graphs from the papers that X pointed us to.
Again, I have no idea how much of this comes from what time of year people do that thing that makes babies, and how much comes from babies conceived at different times of year not making it all the way to birth. But I guess the experts know.
People who think they see patterns in birthdays over the year are obviously succumbing to the hot hands fallacy. Yet another example of human irrationality and the human tendency to see patterns where non exist.
Haven’t you heard about the hot hand fallacy fallacy?
Ether way, people are irrational dummies that I can feel superior to, so I’m good either way.
Well, something’s hot, but I don’t think it’s the hands.
The decline in births on Sunday or holiday (“Dimanche ou fête”) in “Figure 1. – France. Evolution des coefficients journaliers des naissances de 1946 à 1978” (Evolution of daily coefficients of births from 1946 to 1978) might reflect a progressive increase in births by C-section, as those would tend not to be scheduled on a Sunday or holiday.
Inducing labor would also be unlikely to be scheduled for Sunday or a holiday.
Right, that would also contribute to that trend; maybe even more than C-sections do…
France (2015): 22.6% Induced deliveries
France (2017): 19.7% C-section
https://obgyn.onlinelibrary.wiley.com/doi/pdf/10.1111/1471-0528.13805
https://www.statista.com/statistics/283123/cesarean-sections-in-oecd-countries/
I haven’t been able to find trends for those; but the technique for modern induced labor seems to have developed over the 20th century, perhaps also making it more widespread.
Oddly enough, posted just yesterday:
https://kieranhealy.org/blog/archives/2020/02/26/a-new-baby-boom-poster/
Do birthdates in the temperate zone of the southern hemisphere show a similar seasonal pattern? If so, they should show a peak in late March, eh?
In NZ, there is a seasonal peak but the peak is also September.
https://www.stats.govt.nz/tools/most-common-birthday-in-new-zealand
In NZ the peak is the last week in September.
https://www.stats.govt.nz/tools/most-common-birthday-in-new-zealand
Interesting. :) Sounds calendrical but not seasonal. It’s not those cold winter nights in New Zealand.
What’s up with the big peak in April/May back in the 1970s and ’80s? Did lots of women get pregnant during August vacations back then? If so, has that become less of a thing over time?
August weddings sir! :)
An easy way to test this is to look at the birth years of first-born children. April/May births should be much more common amongst first-borns.
Its hard to believe someone hasn’t done this already.
Correction: “birth months”
August honeymoons?
I’m sticking with August weddings until Terry comes up with data to shoot it down :)
Makes perfect sense: declines in ’85 and more ’95 as sexual norms change and more people engage in premarital sex.
We all seem to be forgetting that the graph is about births, and pregnancies do not all result in births.
Also,I’m not following the logic in Jim’s last sentence.