REPORT · №02 · 2026-07-15
The $1.22M blind spot.
A month ago we showed the major is a +$1.24M lever — measured across 19 broad categories. This report zooms in to 115 subfields, 25,597 bachelor’s programs, 2,855,432 graduates — and finds the categories were smoothing over the sharpest edges in the data. The highest-median-ROI degree in America is not an engineering degree. A health-category major loses money for 87.6% of its graduates. And inside a single category label sits a +$1.22M spread the label never mentions.
Data: FREOPP 2021 per-program lifetime ROI, aggregated to 115 CIP subfields (25,597 programs, 2,855,432 graduates, cohort-weighted) · full methodology + downloadable CSV linked at the end · a LE TEEN report.
1 · One label, +$1.22M apart
The biggest spread hiding inside a single major category is in Business and Management: the median construction management graduate clears +$1,458,418 over a lifetime, while the median sales & merchandising graduate clears +$242,736 — a +$1,215,682 gap between two majors filed under the same word. For scale: the headline finding of our first report was a +$1.24M swing across all 19 categories. Nearly that entire swing fits inside one category label.
Health and Nursing is worse, because its gap crosses zero: registered nursing at +$915,056 and communication sciences & disorders at −$179,917 sit +$1.09M apart in the same category. “I’m doing something in health” is not a plan. It’s a coin with a million-dollar spread between its faces.
2 · The best-paying major in America builds buildings
Rank all 115 subfields by median lifetime ROI and the top of the list is not computer science, not any engineering discipline, not anything with prestige attached. It’s construction management: median +$1,458,418, the highest of all 115 — ahead of every one of the 12 engineering subfields, including the best of them, computer engineering (+$1,283,447), and ahead of computer science (+$1,003,700). Zero percent of its graduates are in programs that never break even, and among its programs the median graduate is in the black by age 26.
It is 54 programs and 3,819 graduates — a small door, and nobody’s bragging about it at a college fair. The prestige hierarchy and the payoff hierarchy are different lists. One caveat the data itself insists on: dropping out of the №1 subfield still costs −$114,252. The ladder only pays if you finish it.
3 · 92.3% of dance graduates never break even
The worst subfield in the data is dance: median lifetime ROI −$362,573, and 92.3% of its 3,914 graduates are in programs whose ROI never crosses zero — across 107 programs, not a handful of outliers. Even the 75th-percentile dance graduate — doing better than three-quarters of the field — is −$240,564 underwater.
It is not alone. In drama & theatre arts, 87.6% of 20,828 graduates never break even (median −$182,326); in film, video & photography, 84.7% of 20,318 (median −$179,245); music and fine arts run 70% and 74.3%. None of this says don’t make art. It says: if the degree costs six figures, the data says the degree — not the art — is the part to renegotiate.
4 · A health-care degree that loses money
The third-worst median in all 115 subfields — behind only dance and drama — is not an arts degree. It’s communication sciences & disorders, the bachelor’s that feeds speech-language pathology and audiology: median −$179,917, with 87.6% of its 8,088 graduates in programs that never break even, across 154 programs.
The honest read: FREOPP measures the bachelor’s degree’s own payoff, and in fields where practice typically runs through graduate school, the bachelor’s-only number is the price of the on-ramp, not the destination. But that is the finding: this degree only works as a package deal with more school — and nobody prices the package at the campus tour. A 17-year-old choosing it is committing to six-plus years of tuition; the four-year version, on its own, has the third-worst payoff in America.
5 · The safest large bet in the data is nursing
Registered nursing is 274,445 graduates — roughly one in ten bachelor’s graduates in this entire dataset — across 820 programs. Share of graduates in programs that never break even: 0%. Zero. Median lifetime ROI +$915,056, median break-even by age 26. Only 17 of 115 subfields post a 0% never-breakeven share, and most of the others are small engineering disciplines; nursing does it at industrial scale.
That’s the same category — the same brochure page — as the subfield in finding 4. The category average splits the difference and tells you nothing.
6 · “A computer degree” is not one thing
Inside Computer and Information Sciences, the median computer science graduate clears +$1,003,700. The median software & media applications graduate clears +$28,787 — 35 times less, a +$975K gap between two majors that both get called “computers” at the dinner table. And 35.3% of software & media applications graduates are in programs that never break even, against 0.1% for computer science.
The tech premium is real, but it attaches to the transcript, not to the vibe. A degree that is about computers is not priced like a degree in computer science.
7 · When the mean and the median disagree, believe the median
In 6 of the 115 subfields, the mean and the median have opposite signs — the average says one thing about the field while the typical graduate lives the other. The sharpest case is graphic communications: mean +$39,153 — positive — while the median graduate is at −$105,964 and 60.8% never break even. A minority of strong programs drags the average over the line; most graduates never follow it across. It cuts the other way too: arts & entertainment management has a negative mean (−$12,533) but a positive median (+$11,320).
Any single field can hide this. Across all 115 subfields, 19 have a negative median — 218,376 graduates in this cohort are in fields where the typical outcome is a lifetime loss, from anthropology (−$13,872) to dance (−$362,573). When someone quotes you an average, ask for the median. When they quote the median, ask for your school’s number.
What we didn’t do
We named no school. We prescribed nothing — the data says the top quarter of most negative-median fields still lands positive, and a cheap program changes every number on this page. We reported FREOPP’s own published figures, aggregated to CIP subfields with at least 50 ROI-bearing programs, cohort-weighted, with the methodology and its limitations stated in full — sticker cost not net price, 2021 earnings vintage, fields not destinies.
↓ Download the data (CSV) · Per-major deep pages: all 115 subfields · Methodology
Cite this report:
LE TEEN (2026). “The $1.22M blind spot: what 115 college-major subfields reveal that 19 categories hide.” Report №02. Data: FREOPP 2021, via LE TEEN College ROI Data (doi:10.5281/zenodo.21351602). https://le-teen.com/reports/the-122m-blind-spot