How Azimuth measures college value

The default overall score combines standardized pillar scores with these weights. Standardization puts measures with different units on a common scale before they are combined.

The Return score uses 40% cost-of-living-adjusted median earnings, 30% value added and 30% program benchmark comparison. These are weights within Return. Loan repayment is no longer a component of this pillar. The score adjusts earnings for state cost of living; the dollar figures in a brief are reported, unadjusted earnings.

National institution ranks cover eligible public and private nonprofit colleges that predominantly award bachelor’s degrees, with enough data to score at least three of the four pillars and at least one bachelor’s program. Distance-only institutions are excluded; branch campuses need an explicit decision to rank separately. A school’s presence in the research directory does not by itself establish ranking eligibility.

Hidden-gem lists

A hidden-gem list includes only colleges whose overall score is at or above the 50th percentile nationally and whose admission rate is above 50%. Colleges without a reported admission rate are left out. The colleges that qualify are then ordered by overall score.

Other ranked lists

Most-affordable lists order colleges by average net price for families earning under $30,000, lowest first. Colleges without that figure are listed last. This differs from the Affordability pillar, which uses the average across all family incomes.

Program lists rank bachelor’s programs at ranked colleges by the program’s median earnings four years after completion, or two years when the four-year figure is not published. Each list needs at least five colleges.

Social-mobility program lists include colleges at or above the 60th percentile on the Mobility pillar whose program has a published earnings figure. They are ordered by the college’s median earnings ten years after entry, not by program earnings. Each list needs at least five colleges.

Rank #1 is the highest position. Higher pillar percentiles indicate a stronger relative position in the comparison population. A percentile is not a student’s probability of success, and small rank differences say little about any one student’s result.

The headline earnings measure describes federal-aid recipients who completed a credential, were working and were not enrolled during the measurement year, four years after completion. It does not describe every student who entered the college or predict a starting salary. Read completion rates alongside graduate earnings.

Older entry-cohort earnings cover a different population and period. The Mobility pillar still uses historical low-income earnings measured ten years after entry; that measure has not become a four-year completer measure. Comparing different horizons does not trace one group’s wage growth.

On outcomes pages, downside and upside earnings come from the college’s mix of programs and their four-year earnings. They are not percentiles of individual graduates’ earnings or confidence bounds. Cost-page Parent PLUS risk scenarios instead vary assumed household income.

Read about the earnings-cohort change

The overall six-year completion rate follows first-time, full-time students seeking a bachelor’s degree. The Pell six-year rate follows Pell Grant recipients in a separate federal-aid cohort. Because the groups differ, subtracting these rates does not measure a like-for-like Pell completion gap. First-year retention describes first-time, full-time students returning to the same institution for their second year.

Understand who graduation rates count

Value added is the difference between reported four-year earnings and the model’s expected earnings. The expectation accounts for student background, selectivity, expected completion and economic context. It also applies an earnings floor, using the state benchmark where available and a fallback otherwise.

A positive value means reported earnings exceed that expectation; a negative value means they fall below it. This is an association after statistical adjustment. It does not isolate the causal effect of attending a college or measure what a particular student would gain.

A separate explanation examines how program mix, institutional scale and labor-market context relate to value added. It does not change the headline score. The unexplained remainder is not proof of teaching quality or institutional impact.

How expected earnings are estimated

Program comparisons depend on the field of study, credential level, cohort and reporting coverage. The Scorecard program data used here generally groups fields at the four-digit CIP level. A more detailed display label does not establish a separately measured specialization.

The program ranks on school pages cover eligible bachelor’s programs at institutions in the top 500 of Azimuth’s overall ranking. Programs need reported four-year earnings, at least ten reported awards and a loan-payment burden measure. These rankings do not cover every program nationally.

Program award counts describe credentials conferred. They are not counts of unique students or the sample used to calculate earnings; one person can receive more than one credential.

National program benchmarks describe the same field and credential across reporting institutions. Their lower and upper reference values are not the earnings spread of students at one college. Small or suppressed cohorts limit what can be compared; an institution’s overall rank does not substitute for a program result.

How to evaluate program rankings · Read the labor-market analysis

Net price is an average after grants and scholarships for the reported student group. Pages show two of these averages: one across all family incomes and one for families earning under $30,000. Income-band averages show how the price changes with family income; your aid offer and housing choices may produce a different bill. Borrower debt figures do not describe every student or every parent.

The Affordability ranking compares institutional cost and borrowing measures. Financial GPS models loan payments against earnings or household resources under stated assumptions. A ranking percentile, a GPS tier and a family’s payment scenario answer different questions. Read the earnings basis, interest rates, repayment term and income assumptions with any scenario.

Read the Financial GPS framework

College Scorecard (June 2026 release) and IPEDS supply federal education measures. BEA regional price parities adjust earnings for the cost of living in the Return pillar. Measures may cover different years and student populations, even when released together.

The Common Data Set, which colleges publish about themselves, supplies reported admissions, test-score and financial-aid details for schools that post one, from the 2019–20 to the 2025–26 editions. These appear in school briefs as cited facts. They do not feed any ranking score, and a missing section is shown as missing, never as zero.

The U.S. Bureau of Labor Statistics supplies occupational wages (Occupational Employment and Wage Statistics, May 2025, by state and metro area) and 2024–34 employment projections. Azimuth maps programs to occupations to describe the job market on ranking pages. These figures do not feed any ranking score either.

Missing or suppressed figures remain unavailable in the school briefs. When a college is missing a component, the pillar re-weights the components it has or treats the missing one as average. When a whole pillar is missing, the overall score uses the median of that pillar among ranked colleges. A fallback used to calculate a score is not presented as an observed student outcome or a reported price. Missing Parent PLUS data does not mean parents borrowed nothing.

The date a page was retrieved or refreshed is different from the year its data describes. Read the source dates attached to each brief. Every figure in the written explanations is checked against those sources, and a sentence that disagrees with them is withheld.

College Scorecard data and documentation · BLS Occupational Employment and Wage Statistics · Coverage

• A value-added approach to college outcomes: how expected earnings are estimated
• Four-year versus ten-year earnings: why the earnings cohorts differ
• What the graduation rate doesn’t count: who completion rates follow
• How to evaluate program rankings
• The net price illusion: sticker price versus what families pay
• The Financial GPS framework: payments against earnings
• The supply and demand map for college degrees