REPRODUCIBLE METHOD
How the ledger is built
The June 10, 2026 College Scorecard featured institution and field-of-study downloads. Institution pages include currently operating main-campus records. Program pages require at least 75 working graduates plus non-suppressed four-year earnings and debt medians.
Why main-campus records
Federal outcomes can be calculated at the six-digit OPEID level and repeated across branch UNITIDs. Filtering to MAIN = 1 reduces duplicated outcome pages while preserving a stable institution identifier.
Why the 75-graduate rule
College Scorecard already suppresses sensitive small cells. Our additional threshold avoids thin program pages and makes peer percentiles less sensitive to tiny cohorts. It is a publication rule, not an official quality standard.
Peer context
Program percentiles compare only records with the same four-digit CIP code and credential level that pass the publication rule. Earnings percentile ranks higher medians upward. Debt context reports the share of published peers with higher median debt. Neither is a recommendation.
Cohorts cannot be blended casually
Net price, completion, retention, debt, repayment, and earnings cover different years and student groups. The site displays them together for research but never combines them into a composite score, ROI, or payback period.
Privacy and suppression
The source uses privacy suppression and, for earnings, differential-privacy techniques. Suppressed values stay missing. We do not infer them from nearby fields.
Update policy
Each edition is rebuilt from a named featured release, reruns all thresholds and peer groups, and updates every canonical page together.