How we calculate ROI

• College Scorecard, Most Recent Cohorts. The U.S. Department of Education publishes program-level cost, debt and earnings data by institution and field of study. We use the most recent published cohorts.
• IRS earnings records. Scorecard earnings are measured from federal tax records for students who received federal financial aid, linked to the program they completed.
• State high-school earnings baseline. For each state we use the typical earnings of a worker with a high-school education, so a degree is measured against the realistic local alternative rather than a national average.

We build five years of cumulative earnings for each program from its year-one through year-five figures. Then we build the same five years for the state high-school baseline. The difference is the five-year premium: the earnings the degree plausibly added, not just the earnings the graduate has.

5-yr premium = (Yr1..Yr5 graduate earnings)
             − (5 years of state high-school baseline earnings)

A large premium bought with an enormous loan is not the same deal as a modest premium bought cheaply. So we express the premium as a multiple of both the sticker cost and the debt students actually take on.

premium / cost = 5-yr premium ÷ 4-year sticker cost
premium / debt = 5-yr premium ÷ median student debt

The composite score combines these ratios so that affordable programs with solid outcomes rank ahead of expensive programs with similar earnings. Ranks are assigned within each major, so you're comparing nursing programs to nursing programs.

• Federal-aid recipients only. Earnings cover students who received federal financial aid. Graduates who paid entirely out of pocket are not in the data, which can shift a program's picture.
• Small cohorts are noisy. Where only a few graduates are in the earnings cohort we flag the program as a small cohort. Read those rows as directional, not precise.
• Costs are sticker prices. Grants, scholarships and in-state discounts can reduce your real cost dramatically. Always run your own net-price number.
• Correlation, not causation. Programs that admit high-achieving students would show strong earnings regardless. The premium is a measurement, not a guarantee of what a degree will do for you.
• U.S. only, past cohorts. Coverage is U.S. institutions, and every figure describes people who already graduated — not the labour market you will enter.
• Money isn't everything. Some of the lowest-ROI programs on this site train people we badly need. ROI is one input to your decision, not the decision.

Around 300 program rows in the source data are missing a rank, score, state or cost figure. Rather than show blanks that look like zeros, we left them out of the explorer. Everything shown has a complete cost-and-earnings picture.

If a figure looks wrong for a program you know well, tell us in the discussions and we will check it against the source release.