Small programs are hidden. Federal data suppresses tiny cohorts to protect privacy.
Formats are pooled. Evening, executive, online, and full-time versions of a degree blend into one row.
Elite no-loan schools are nearly invisible. The data sees the borrowing public best.
Long horizons are modeled. Nobody can observe a 40-year career in advance — past year 10, we model and label it.
Medians describe groups, not you. The people who chose a program may differ from the person deciding now — treat degree-value estimates as leaning optimistic; we disclose that rather than model it. And medians inherit the market’s inequities — pay gaps in the records become pay gaps in our projections; we report the market, we don’t endorse it.
The rules keep changing. Repayment and forgiveness terms have been rewritten roughly every three years — every projection that depends on a federal program is a disclosed risk, never a promise.
Advice changes markets. If many students chase today’s best-paying fields, tomorrow’s pay in those fields likely falls — it happened to 1970s engineers. Our numbers describe the past; crowds change the future.
The data sees borrowers best. Programs whose students rarely take federal aid barely appear in federal earnings records — the most elite schools are often the least visible. Where school-level data is missing, we say so and use national figures for the major.