The Denominator
Chapter 01·the premise
What was asked
- 01If a ratio is stable worldwide but volatile country by country, what is the world median actually describing?
- 02Which half of the fraction does the variation live in — and which half does every discussion of it talk about?
- 03How many of the deaths that could have been called suicide were called something else, and does that differ by sex?
- 04At what ages is the gender gap larger than the uncertainty in how those deaths were classified?
What it found
The male:female suicide ratio is quoted like a property of human beings. Across 183 countries its world median has not moved in two decades — but 52 individual countries have moved by more than a point, the variation between them comes 104.1% from the female side and −4.1% from the male side, and at the oldest ages two defensible ways of counting the same deaths disagree by more than the entire age effect.
There is a sentence you have almost certainly heard. Men die by suicide about three or four times as often as women.
It gets said the way people say that water boils at a hundred degrees. Not as a finding somebody had to go and measure, but as a fact about how human beings are built — solid enough to stand on while you argue about something else. It shows up in policy papers, in newspaper columns, in the opening line of charity campaigns.
I want to take that sentence apart. Not to argue it is false, because it is not false. Men really do die more, in every country I could find data for, at every age, under every way of counting I tried. That much survives everything in this piece.
But a sentence can be true and still be doing something strange. And this one is doing three strange things at once.
Start with what the number actually is. It is a division. You take how many men out of every hundred thousand died by suicide in some country, you take the same figure for women, and you divide the first by the second. If it comes out at 3.5, men died three and a half times as often. That is the whole calculation. Two numbers, one division.
Now here is the thing that made me start pulling on the thread. Across 183 countries, that ratio in 2000 had a middle value of . Two decades later, in 2021, it was . And if you look at every single year in between, the whole thing wanders by — less than a quarter of one point, across twenty-two years and most of the planet.
Social statistics do not usually behave like that. Almost nothing about human beings holds that still for that long. It really does look like a constant of nature.
The rest of this piece is about what that stillness is hiding.
Komentar