Last time we looked at the big number: across almost the entire OECD, young women hold more degrees than young men, and have for a generation. The overall figure is settled. But an overall figure is a blanket thrown over a lot of variety, and the interesting part is what it covers up.
So here is the same question asked field by field. Of everyone who graduates in a given subject, what share are women?
Women are the majority in six of these eight fields. Education and health, the two most female, are roughly three-quarters women. Social sciences, the arts, business and law, even natural sciences and mathematics all sit above the halfway line. One field after another down the list, women graduate in greater numbers than men.
Then the pattern runs out. Engineering is 29% women. Computing is 24%. These are the two hold-outs, the only broad fields where men still graduate in clearly greater numbers.
There is a wrinkle worth catching, because it hides in plain sight. Natural sciences, maths and statistics come in at 54% women, a small majority. Fold that field together with computing and engineering into the single word everyone reaches for, "STEM," and the combined figure drops to about a third women, which makes the whole cluster sound uniformly male. It is nothing of the kind. The maths and the lab sciences are close to even. The shortfall lives in two specific rooms, engineering and computing.
The gap that won't close
You might expect those two to be slowly catching up, the way so much else in education has. Here is the last decade for each STEM field.
Natural sciences never had the gap to close, sitting above half for the whole decade. Engineering and computing are creeping upward, but the word to sit with is creeping. Engineering went from 25% to 29% in ten years. Computing went from 20% to 24%. That is roughly a third of a percentage point a year, the kind of movement you need a decade of data to even see.
Play that pace forward and the numbers get quietly startling. If engineering keeps closing its gap at the rate of the last ten years, it reaches an even split around 2080. Computing gets there around 2095. That is not a forecast, straight lines rarely hold for seventy years, but it is a fair measure of how slow the current pace really is. The girls who would finally balance a computing class at that rate have not been born.
This is the part that matters beyond the chart. Engineering and computing are among the highest-earning fields a degree can lead to, so a gap that stays open there feeds directly into the pay gap in a way a gap in the arts would not. Women winning almost every field is the headline. The two they have not won are two of the most lucrative, and those are the two barely moving.
The data can show the stall. It cannot tell you why these two fields resist a tide that swept through everything around them, or what would finally shift them. Those questions stay open. What it settles is that "women are more educated now" is true almost everywhere, and the exceptions are neither vague nor closing quickly. The next time someone says "women avoid STEM," it is worth remembering that half the scientists are women. It is the engineers and the programmers who are missing, and they have been missing at almost exactly the same rate for ten years.
Data: OECD, share of tertiary graduates in each field who are women, from DF_UOE_NF_SHARE_FIELD, via the SDMX API (unit: percentage of graduates in the same field). Field spectrum is 2022; the trend covers 2013 to 2023, the full span of this series. The chart shows the eight most common broad fields; agriculture, services and generic programmes are left off for space and sit in the accompanying data. "STEM" is the OECD's combined science, technology, engineering and mathematics grouping (F05T07), about 34% women. The parity-year figures assume the last decade's pace continues, which is an illustration, not a projection. Pulled with oecd-mcp. Companion to the previous issue on when women overtook men in degrees overall.