How the numbers are computed.

The question the calculator answers

/worth-it computes a 30-year projected net value for a specific school and home state. It compounds career earnings above the high-school wage baseline forward to the terminal year, then subtracts compounded out-of-pocket attendance costs and loan payments. The legacy NPV field is an accumulated-value comparison, not discounted present value or an observed return.

Costs in this net-value field: tuition plus room and board before aid, net of the borrowed share, and standard amortized loan payments. Foregone wages during study are reported separately; their compounded total is an upper bound assuming every dollar is saved. Earnings input: the saved institution earnings proxy, with an optional same-school program estimate in the calculator. BEA state price parities adjust the alternative high-school wage; the institution earnings estimate is not adjusted by that factor. The estimate’s price year and suitability across a 30-year career remain unknown, so this is not a verified constant-dollar return.

Data sources

• FREOPP Higher Education ROI (2021) — per-program lifetime-ROI estimates and institution earnings for ~30K U.S. bachelor’s programs, published by the Foundation for Research on Equal Opportunity (freopp.org, CC-BY-4.0). We use the ages-29–31 cohort-weighted mean as the institution-earnings proxy of record. 1,661 of 3,392 institutions carry real FREOPP earnings; the remainder fall back to the BLS national bachelor’s median ($74K) and are excluded from all named rankings.
• IPEDS 2023–24 (Final) — the U.S. Department of Education’s universe of postsecondary institutions: identity, published tuition & fees (resident and non-resident), room & board (nces.ed.gov/ipeds, U.S. government work).
• BEA Regional Price Parities (2023) — cost-of-living deflators by state, used to adjust the alternative high-school wage for the selected home state; institution earnings are unchanged (bea.gov, U.S. government work).

Public data only; no institutional scraping. We are an analysis layer over these sources — when you cite a number, the underlying figure is theirs and we name them inline.

Default assumptions (all user-editable in the calculator)

• Program length: 4 years (bachelor’s default).
• Career horizon: 30 years post-graduation.
• Borrowed share of cost: 60% of attendance costs as a model default; use your actual borrowing share.
• Loan rate: 6.5% APR as a model default, with interest accrued during study and a 10-year repayment term; use the rate and terms in your offer.
• Alternative investment return: 7% as the model’s annual compounding convention. It is not a guaranteed market return; test other rates.
• High-school baseline wage: $34,000 as the model baseline before the BEA state price-parity adjustment; it is not a personal wage forecast.

How each ranking is built

• Worst-ROI majors — ranks FREOPP’s 19 major categories by FREOPP’s own published lifetime-ROI figure, cohort-weighted by graduate count. Their number, reported verbatim; names no institution.
• The out-of-state tax — runs the calculator twice per public four-year school: once at resident tuition, once at non-resident, with earnings and cost-of-living held identical. The gap is purely the residency pricing line. Only schools that stay NPV-positive both ways are listed.
• Best value by state — each state’s own four-year schools at resident pricing, ranked by 30-year projected net value. NPV-positive schools only.

Editorial gates (what keeps the rankings honest)

• No school is ever named with a negative number. Negative results exist in the calculator (your school, your inputs, your screen) — but our published rankings only name schools in positive or neutral cost-comparison context.
• Thin-coverage exclusion: institutions whose FREOPP earnings roll up from fewer than 3 reported programs are excluded from named rankings — one outlier program shouldn’t crown a school.
• No estimated earnings in rankings: schools without real FREOPP coverage are excluded rather than estimated with the national-median fallback.
• Hand-verified exclusions: schools confirmed closed since the IPEDS 2023–24 vintage are removed by hand.

National model coverage in the API

The colleges API reports aggregate coverage before the positive-value and top-500 named selection: 1,424 quality-eligible four-year institutions, with 358 negative, 1,066 positive, 0 zero and 0 unknown model results under the resident defaults above. The negative share is 25.14% of institutions with a finite model result. Each institution counts once; this is not enrollment weighting or a graduate’s loss probability. The JSON retains missing/invalid-data partitions and the full funnel.

The sign comes from the calculator’s rounded accumulated net value at year 34, not discounted present value, actual losses or FREOPP’s published lifetime ROI. It subtracts compounded out-of-pocket costs and loan payments from career earnings above the high-school path; foregone wages during study are a separate reported field. Costs are before aid. Institution earnings use saved FREOPP estimates at ages 29–31, weighted by positive cohort counts with weight 1 for missing or zero counts. That legacy earnings field is not a measured median ten years after entry. The estimate’s price year, applicability across this career horizon and completion probability remain unknown. Changing assumptions can change a sign; the aggregate identifies no negative-result school.

Known limitations (published on purpose)

• Sticker cost, not net price. We do not apply a financial-aid adjustment — a single aid factor cannot honestly represent the public/private/for-profit × in-state/out × Pell/no-Pell distribution. A school generous with aid will look worse here than your actual offer; run your real numbers.
• Gross-of-tax on both paths, so taxes approximately cancel; no real wage growth applied to either stream.
• No completion-risk weighting. The model assumes you finish. Roughly 40% of starters don’t finish in 6 years (federal six-year completion data, IPEDS), and non-completion is the worst financial outcome — costs without the premium.
• Category means hide spread. A major category’s mean ROI is not your program’s ROI; a specific program at a specific school can land far from its category.
• Earnings vintage: FREOPP 2021 (cohorts 2017–19). Field-level relative differences are stable; absolute levels drift with inflation.

Use the data

Every ranking is downloadable and free to cite with attribution (ranking tables: free to share with attribution; underlying data carries its upstream license — FREOPP CC-BY-4.0, IPEDS/BEA U.S. government works).

• worst-roi-majors.csv — 19 major categories, mean lifetime ROI, % of graduates below zero
• out-of-state-tax.csv — every qualifying public four-year (445 schools), resident/non-resident NPV + the penalty
• best-value-by-state.csv — all 51 states’ top-15 lists at resident pricing

The full set — including the 3,392-institution join table and the per-major AI-exposure scores — is published as a permanent, citable dataset with a DOI: doi.org/10.5281/zenodo.21351602 (canonical repository, CC BY 4.0).

Developers: every number is also a free JSON API — keyless, CORS-open, OpenAPI 3.1 spec included.

Cite this page:

Cite the dataset: