Version 2026.08-calibration-3
2026-08-28
Expanded the first-pass evidence set to the 50-institution release floor and replaced manual review booleans with record-level second-review and correction-window gates.
Public audit trail
This log makes changes inspectable instead of silently rewriting history. It also shows exactly which conditions still block the calibration data from becoming a publicly indexed edition.
Release history
2026-08-28
Expanded the first-pass evidence set to the 50-institution release floor and replaced manual review booleans with record-level second-review and correction-window gates.
2026-08-27
Expanded the calibration set to ten institutions and added automated release gates, explicit audit status, shareable institution graphics, and a public correction log.
2026-08-27
Created the complete 2025 Carnegie R1 sampling frame, six-dimension rubric, five calibration audits, evidence profiles, and open data contract.
Accepted corrections
Only accepted changes appear here. Each entry names the institution, version, evidence, and effect of the correction.
No corrections have been accepted in the calibration version.
The empty state is preserved explicitly so readers can distinguish “none yet” from an unavailable history.