Amigos en Acción
Eighty-Five Point Five
We Had It in 1984 — Eighty-Five Point Five

Women in computing earn 85.5 cents for every dollar men in computing earn.

$310 a week. About $16,120 a year.

A woman working at dual monitors with code on the screen — licensed illustration, not a person in the statistic. We Had It in 1984 — Eighty-Five Point Five.
A woman working at dual monitors with code on the screen. Licensed illustration — not a person in the statistic. Photo: Pexels
HomeWe Had It in 1984 › Eighty-Five Point Five
$1,828median weekly earnings, women in computing
$2,138median weekly earnings, men in computing

Full-time workers in computer and mathematical occupations, 2025 annual averages, from the Current Population Survey. Women's median is 85.5% of the men's — a difference of $310 a week, or roughly $16,120 over a year.

A woman typing at a desktop workstation — licensed illustration, not a person in the statistic. We Had It in 1984 — Eighty-Five Point Five.
A woman typing at a desktop workstation. Licensed illustration — not a person in the statistic. Photo: #WOCinTech Chat (CC BY)

Across all occupations the same comparison is 82.1%, so computing is slightly better than the national average on this measure. Slightly.

What this number is not

It is an unadjusted occupational-group median. It compares every woman in the group to every man in it — not the same job, at the same level, with the same years of experience, at the same employer. A controlled comparison produces a smaller number, and anyone who tells you otherwise is selling something.

It is still worth publishing, for two reasons. It is what the federal statistical agency actually measures and publishes, so it can be checked. And the composition it reflects is itself the finding: if women are concentrated in the lower-paid and lower-titled parts of the group, the unadjusted gap is measuring exactly that, which is not a measurement error — it is the thing.

The honest version, without inflating anything to make a point: women in computing earn 85.5 cents on the dollar unadjusted, that a like-for-like comparison narrows it substantially, and that the narrowing does not make the unadjusted figure meaningless — because where people sit in a distribution is not a nuisance variable, it is the subject.

The arc, in one table

A woman teaching at the front of a classroom — licensed illustration, not a person in the statistic. We Had It in 1984 — Eighty-Five Point Five.
A woman teaching at the front of a classroom. Licensed illustration — not a person in the statistic. Photo: Unsplash
Academic yearWomen's share of US computer science bachelor's degrees
1983–84 — the peak37.1%
1984–8536.8%
1985–8635.7%
2007–08 — the floor17.6%
2008–0917.8%
2021–22 — most recent22.6%

Computed from the degree counts published in NCES Digest of Education Statistics Table 325.35, which covers computer and information sciences. The table prints counts, not shares, so these percentages are our arithmetic on its numbers and can be checked against it directly.

What is already paid for

Programme names, amounts and deadlines change. These are the doors; verify current terms with the source before counting on a number.

What to do this week

Before your next salary conversation, look up your occupation and state on the BLS Occupational Employment and Wage Statistics page. Walk in with the federal number rather than a range from a jobs site.

Eight routes in, and what each one asks of you

Ordered by what has to be true before the route exists — a school that offers the course, a public programme with money behind it, an open door, or nobody at all.

The high school course

Where the share is 33%, and where the decision to continue is mostly already made.

Depends on your school

Dual enrollment

A community college course taken in high school, funded by most states, and it counts twice.

Publicly or employer funded

Community college

Open enrollment in every state, and the least expensive route into the field.

Open, but nobody has to teach you

Registered apprenticeship

A wage from the first day. No tuition, and no gap in earnings.

Publicly or employer funded

Returning after time out

A career break is a gap on a résumé and not a gap in ability. Several employers now recruit for exactly this.

Open, but nobody has to teach you

An industry certificate

Months rather than years, and it dates your skills rather than your gap.

Publicly or employer funded

Moving across where you already work

The most common route into technical work, and the one that requires somebody to know you want it.

Open, but nobody has to teach you

Self-taught with a portfolio

Free, unlimited, and the route where the drop-off is highest and least discussed.

No disclosure duty

By state

Training money, apprenticeships and colleges are administered state by state and delivered county by county. Find yours.

AlabamaAlaskaArizonaArkansasCaliforniaColoradoConnecticutDelawareDistrict of ColumbiaFloridaGeorgiaHawaiiIdahoIllinoisIndianaIowaKansasKentuckyLouisianaMaineMarylandMassachusettsMichiganMinnesotaMississippiMissouriMontanaNebraskaNevadaNew HampshireNew JerseyNew MexicoNew YorkNorth CarolinaNorth DakotaOhioOklahomaOregonPennsylvaniaRhode IslandSouth CarolinaSouth DakotaTennesseeTexasUtahVermontVirginiaWashingtonWest VirginiaWisconsinWyoming
Sources. Degree shares: computed from the counts in NCES Digest of Education Statistics Table 325.35 (computer and information sciences) and Table 318.10 for all fields; the table publishes counts rather than percentages, so the shares here are our arithmetic on its numbers and can be checked against it. Occupation shares and earnings: US Bureau of Labor Statistics, Current Population Survey Tables 11 and 39, 2025 annual averages — the earnings figures are unadjusted occupational-group medians, not a controlled like-for-like comparison, and we say so. High school enrolment: Code.org Advocacy Coalition, State of Computer Science Education, 2023–24. AP exam composition: computed from the male and female counts in the College Board AP Program Summary Report 2025. Attrition: Accenture and Girls Who Code, Resetting Tech Culture, 2020 — the comparison there is women in technology against women in other work, not women against men, and we state it that way. Earnings effect of one course: Liu, Conrad & Blazar, EdWorkingPaper 24-904. County classifications: USDA Economic Research Service County Typology Codes; labour markets: 2020 Commuting Zones. County counts are computed from the committed county file at build time, not asserted.

Photographs on this campaign are licensed stock used as illustration — not portraits of the people inside the federal counts. Full credits: CREDITS.md.