Redaction analysis
576 FOIA markers countedRedaction is quantified. We count explicit FOIA exemption markers in the text and separately flag documents blacked out so heavily they OCR to near-empty.
Density is a normalized proxy: FOIA markers per page. It indexes how much of a document is withheld, not the sensitivity of what was removed.
- DOW-UAP-D3-Mission-Report-Arabian-Gulf-202091 marks
- DOW-UAP-D32-Mission-Report,-Syria-October-202476 marks
- DOW-UAP-D25-Mission-Report-Greece-January-202460 marks
- DOW-UAP-D27-Mission-Report-United-Arab-Emirates-October-202358 marks
- DOW-UAP-D28-Mission-Report-East-China-Sea-202453 marks
- DOW-UAP-D74-Mission-Report-Syria-November-202334 marks
- DOW-UAP-D75-Mission-Report-Gulf-of-Aden-July-202426 marks
- DOW-UAP-D12-Mission-Report-Iraq-May-202217 marks
Redaction tracks (non-)resolution
what gets withheld, by case outcomeSorting redaction by what the agency concluded shows the record is most opaque exactly where it is least explained. Cases left unresolved carry far heavier redaction than cases the agency explained; the handful flagged simply “redacted” are the most blacked-out of all.
Mean FOIA-marker density per page, by resolution status. Read it as a property of classification policy, not of the sightings: the most-withheld material is concentrated on the cases that were never closed. Confound — “unknown” cases are often thin scans where little is present to redact, so their low density is partly an artifact.